Welcome to IDM malaria modeling¶
The Institute for Disease Modeling (IDM) develops complex software for disease modeling. The primary software, Epidemiological MODeling software (EMOD), helps determine the combination of health policies and intervention strategies that can lead to disease eradication. We share this modeling software with the research community to advance the understanding of disease dynamics.
Documentation structure¶
We recently separated the documentation into disease-specific documentation. For example, this documentation set includes general installation and usage instructions in addition to content specific to modeling malaria and other vector-borne diseases.
“Using the model” contains information for researchers who want to create simulations of disease dynamics using EMOD as developed by IDM.
“Advancing the model” contains information for researchers and developers who want to modify the EMOD source code to add more functionality to the model.
What’s new¶
This topic describes new functionality and breaking changes for recently released versions of Epidemiological MODeling software (EMOD).
EMOD v2.10¶
The EMOD source v2.10 release includes new and updated malaria configuration, demographic, and intervention parameters. With this release, EMOD now uses Visual Studio 2015, Boost 1.61.0, and SCons 2.5.0.
Malaria¶
Several habitat parameters in the malaria model have been upgraded, creating more flexibility in the model and enabling the user to have more control over habitat customization. These changes allow the model to more accurately capture real-world habitat availability and how it affects different mosquito species.
- LINEAR_SPLINE habitat type
This new option under Larval_Habitat_Types increases model customization by allowing the user to specify an arbitrary functional form (derived from data) for larval habitat availability throughout the year, instead of relying on climatological parameters such as rainfall or temperature. For more information, see Vector species parameters.
- LarvalHabitatMultiplier by species
This new feature in demographics allows larval habitat availability to be modified independently for each species sharing a particular habitat type. Prior versions of EMOD applied the same modifier to all species within a shared habitat; this upgrade enables you to apply modifiers to individual species within a habitat. For more information, see NodeAttributes parameters.
- ScaleLarvalHabitat by species
This new intervention enables species-specific effects of habitat interventions within shared habitat types, such that habitat availability is modified on a per-species level. For more information, see ScaleLarvalHabitat parameters.
- Breaking changes
PIECEWISE_MONTHLY has been changed to LINEAR_SPLINE.
EMOD software upgrades¶
- Microsoft Visual Studio
EMOD now uses Visual Studio 2015, and Visual Studio 2012 is no longer supported. The Visual Studio solution file in the EMOD source, EradicationKernel, has been updated for Visual Studio 2015. If you have custom reporter EMODules (DLLS) that were built using Visual Studio 2012, you will need to rebuild them with Visual Studio 2015; otherwise, your simulation will crash when it attempts to load the DLLs built by Visual Studio 2012.
- Boost
EMOD now supports using Boost 1.61.0. If you continue to use Boost 1.51.0, you will get the following warning, “Unknown compiler version - please run the configure tests and report the results.”
- Environment variables
To make it easier to use Boost and Python with Visual Studio, IDM paths have been created. These two paths, IDM_BOOST_PATH and IDM_PYTHON_PATH, need to be added to your environment variables by using either the setx command from a command line or the Windows System Properties panel.
- SCons
EMOD was tested using SCons 2.5.0, as it supports Visual Studio 2015. If you do not add the new IDM environment variables for Boost and Python, you will need to modify the Boost and Python paths in the SConstruct file in the EMOD root directory.
- Python
EMOD was tested with Python 2.7.11 and 2.7.12. If you are building the EMOD executable (Eradication.exe) and have an earlier version of Python (for example, 2.7.2), you will see the following warning message on some files, “c:python27includepymath.h(22): warning C4273: ‘round’: inconsistent dll linkage.” Upgrade to Python 2.7.11 or 2.7.12 to get rid of the warning message.
For more information, see Prerequisites for working with EMOD source code.
EMOD installation¶
Epidemiological MODeling software (EMOD) can be installed on computers running Windows 10, Windows Server 12, and Windows HPC Server 12 (64-bit) or CentOS 7.1 on Azure. For both operating systems, there are relatively few prerequisites required and the EMOD executable (Eradication.exe) or Eradication binary is pre- built from the latest EMOD source code release. After completing installation, you can run simulations locally or on an HPC cluster using real-world data.
If you are not familiar with disease modeling, we recommend using the training documentation that uses the Quick Start with Excel Front End to walk through several disease modeling tutorials before installing and working directly with EMOD. You are also free to walk through those tutorials running simulations directly with EMOD as described here.
Note
The Quick Start can only be run locally on a 64-bit Windows computer and the Excel Front End is only supported on 32-bit Microsoft Excel, though there is a workaround if you do not want to reinstall your version of Excel. See Troubleshooting EMOD simulations.
The Quick Start is not updated as frequently as the EMOD source, so may not contain all parameters available in the latest version of EMOD. In addition, while you can modify the parameter values, you cannot add or subtract parameters from JSON files within the Excel Front End.
If you want to download and modify the EMOD source code and build the Eradication.exe or Eradication binary yourself, see EMOD source code installation.
Install EMOD on Windows¶
Follow the steps below to install EMOD on computers running Windows 10, Windows Server 12, and Windows HPC Server 12 (64-bit).
The EMOD executable (Eradication.exe) is tested using Windows 10, Windows Server 12, and Windows HPC Server 12 (64-bit). Windows HPC Server is used for testing remote simulations on a high-performance computing (HPC) cluster and the other Windows operating systems are used to test local simulations.
Install the Microsoft HPC Pack 2012 Client Utilities Redistributable Package (64-bit). See http://www.microsoft.com/en-us/download/details.aspx?id=36044 for instructions.
Install the Microsoft MPI v8. See https://www.microsoft.com/en-us/download/details.aspx?id=54607 for instructions, being sure to run the MSMpiSetup.exe file.
Download the EMOD executable (Eradication.exe). See EMOD releases on GitHub.
(Optional) Install plotting software¶
None of the following plotting software is required to run simulations with EMOD, but they are useful for creating graphs from and evaluating the model output. In addition, EMOD provides many Python scripts for analyzing data.
Note
IDM does not provide support or guarantees for any third-party software, even software that we recommend you install. Send feedback if you encounter any issues, but any support must come from the makers of those software packages and their user communities.
Python and Python packages¶
Python is required to install many of the software packages described below.
We recommended that you download some of the Python packages from http://www.lfd.uci.edu/~gohlke/pythonlibs, a page compiled by Christoph Gohlke, University of California, Irvine. The libraries there are compiled Windows binaries, including the 64-bit versions required by EMOD.
Install Python 2.7.11 or 2.7.12. See https://www.python.org/downloads/ for instructions.
In the Customize Python dialog box, verify that Add python.exe to PATH is selected to add Python to the PATH environment variable on your computer.
From Control Panel, select Advanced system settings, and then click Environment Variables.
Add a new variable called IDM_PYTHON_PATH and set it to the directory where you installed Python, and then click OK.
Open a Command Prompt window and type the following to verify installation:
python --version
The Python package manager, pip, is installed as part of Python 2.7.11 or 2.7.12 and is used to install other software packages.
The Python utilities dateutil, six, and pyparsing provide text parsing and datetime functionality.
Open a Command Prompt window.
Enter the following commands:
pip install python-dateutil pip install pyparsing
The NumPy package adds support for large, multi-dimensional arrays and matrices to Python.
Go to http://www.lfd.uci.edu/~gohlke/pythonlibs/#numpy and select the WHL file for NumPy 1.11.3 (64-bit) that is compatible with Python 2.7.11 or 2.7.12.
Save the file to your Python installation directory.
Open a Command Prompt window and navigate to the Python installation directory, then enter the following, substituting the name of the specific NumPy WHL file you downloaded:
pip install numpy-1.x.x+mkl-cp27m-win_amd64.whl
Matplotlib is a Python library for making publication quality plots using syntax familiar to MATLAB users. Matplotlib also uses NumpPy for numeric manipulation. Output formats include PDF, Postscript, SVG, and PNG, as well as a screen display.
Go to http://www.lfd.uci.edu/~gohlke/pythonlibs/#matplotlib and select the WHL file for Matplotlib 1.5.3 (64-bit) that is compatible with Python 2.7.11 or 2.7.12.
Save the file to your Python installation directory.
Open a Command Prompt window and navigate to the Python installation directory, then enter the following, substituting the name of the specific matplotlib WHL file you downloaded:
pip install matplotlib-1.x.x+mkl-cp27m-win_amd64.whl
R¶
R is a free software environment for statistical computing and graphics.
Go to https://www.r-project.org/ and install R 3.2.0 (64-bit).
MATLAB¶
MATLAB is a high-level language and interactive environment for numerical computation, visualization, and programming. The MATLAB Statistics and Machine Learning Toolbox™ provides functions and applications to describe, analyze and model data using statistics and machine learning algorithms.
Go to http://www.mathworks.com/products/matlab/ and install MATLAB R2015a.
If desired, go to https://www.mathworks.com/products/statistics.html and install the MATLAB Statistics and Machine Learning Toolbox™ R2015a.
Install EMOD on CentOS on Azure¶
Follow the steps below to install EMOD on CentOS 7.1 on Azure environments. The setup script installs most prerequisite software, including Python and the Python packages dateutil, six, pyparsing, NumPy, and matplotlib. Other prerequisites, such as Boost 1.61.0 and Microsoft MPI v8, are declared by the script as required. Because the installation instructions will vary based on the particular distribution you are running, instruction information is not included here.
The Eradication binary is tested and supported on a CentOS 7.1 on Azure virtual machine. It has also been successfully built and run on Ubuntu, SUSE, and Arch, but has not been tested and is not supported on those Linux distributions.
Note
IDM does not provide support or guarantees for any third-party software, even software that we recommend you install. Send feedback if you encounter any issues, but any support must come from the makers of those software packages and their user communities.
The script provides the option of installing the EMOD source code and input data files provided by IDM, but these steps will install a pre-built version of the Eradication binary. For information on building the Eradication binary from source code, see EMOD source code installation.
Before you begin, you must have the following:
sudo privileges to install packages
15 GB free in your home directory (if you install the EMOD source code and input data files)
An Internet connection
Download and run the PrepareLinuxEnvironment.sh script on EMOD releases on GitHub.
Respond to the prompts for information while the script is running. If you choose not to download the EMOD source and input data files, do the following. This example assumes that a directory named IDM is in your home directory and contains the subdirectories EMOD, containing the EMOD source code, and EMOD_InputData, containing the input data files directory.
Set the EMOD_ROOT environment variable to the path to the EMOD source path:
EMOD_ROOT=~/IDM/EMOD
Put Scripts and . in the path:
export PATH=$PATH:.:$EMOD_ROOT/Scripts
Create a symlink from the EMOD directory to InputDataFiles:
ln -s /home/general/IDM/EMOD-InputData $EMOD_ROOT/InputData
If you run simulations using in the same session that you updated EMOD_ROOT and the Scripts path, reload the .bashrc file using
source .bashrc.
Download the Eradication binary for CentOS 7.1 on Azure. See on EMOD releases on GitHub.
(Optional) Install plotting software¶
None of the following plotting software is required to run simulations with EMOD, but they are useful for creating graphs from and evaluating the model output. In addition, EMOD provides many Python scripts for analyzing data.
Note
IDM does not provide support or guarantees for any third-party software, even software that we recommend you install. Send feedback if you encounter any issues, but any support must come from the makers of those software packages and their user communities.
R¶
R is a free software environment for statistical computing and graphics.
Go to https://www.r-project.org/ and install R 3.2.0 (64-bit).
MATLAB¶
MATLAB is a high-level language and interactive environment for numerical computation, visualization, and programming. The MATLAB Statistics and Machine Learning Toolbox™ provides functions and applications to describe, analyze and model data using statistics and machine learning algorithms.
Go to http://www.mathworks.com/products/matlab/ and install MATLAB R2015a.
If desired, go to https://www.mathworks.com/products/statistics.html and install the MATLAB Statistics and Machine Learning Toolbox™ R2015a.
Introduction to the software¶
The Institute for Disease Modeling (IDM) develops detailed simulations of disease transmission through the use of extensive and complex software modeling. The primary software, Epidemiological MODeling software (EMOD), helps determine the combination of health policies and intervention strategies that can lead to disease eradication. EMOD calculates how diseases may spread in particular areas and is used to analyze the effects of current and future health policies and intervention strategies. It supports infectious disease campaign planning, data gathering, new product development, and policy decisions. We share this modeling software with the research community to advance the understanding of disease dynamics.
EMOD supports the following simulation types for modeling a variety of diseases:
Generic disease (GENERIC_SIM)
Vector-borne diseases (VECTOR_SIM)
Malaria (MALARIA_SIM)
Tuberculosis (TB_SIM)
Sexually transmitted infections (STI_SIM)
HIV (HIV_SIM)
The illustration below shows how the simulation types are built upon one another. All parameters available in the generic simulation type are inherited by the vector simulation type. The vector simulation type adds additional parameters specific to the biology of vector-borne diseases, which in turn are inherited by the malaria simulation type and so on. Therefore, depending on the simulation type you select, different parameters are available for you to use.
In addition, simulation types for broader classes of disease can be extended to build your own disease-specific model.
Simulation type inheritance¶
EMOD is a stochastic, agent-based model that simulates the actions and interactions of individuals within geographic areas known as nodes to understand the disease dynamics in a population over time. EMOD can produce statistically significant results over a broad set of parameters and scenarios. Quantitative analysis of the simulated output enables disease eradication efforts to make more data-driven decisions. The IDM research team has published many articles related to modeling, as well as the modeling concepts underpinning EMOD. For a list of published articles, see IDM Publications.
This section provides the conceptual overview of EMOD. The architecture diagram below shows, at a high level, how the system functions. If you run simulations in parallel on a multi-node cluster, there is also a Message Passing Interface (MPI) component used to pass data between multiple instances of EMOD.
High-level EMOD system architecture¶
Input data files¶
Not all files accepted as input for running a simulation are considered input data files. Rather, input data files contain the relatively fixed information about a population to model. For example, climate, geography, demographics, and migration data.
Configuration file¶
The configuration file contains parameters that control many different aspects of the simulation. EMOD provides hundreds of parameters for you to configure your simulation. For example, configuration parameters can control the following:
The disease or disease class to simulate (simulation type)
The name and location for the input data files and output files
Whether to include births, deaths, and migration
Disease attributes, such as infectivity, transmission, immunity, and mortality
The computing resources to use
Campaign file¶
The campaign file contains parameters that distribute outbreaks and the interventions used to control the spread of disease. For example, campaign parameters can control the following:
Target demographic (age, location, gender, etc.) for interventions
Diagnostic tests to use
The cost and timing of interventions
Running a simulation¶
The EMOD executable (Eradication.exe) accepts the input data, configuration, and campaign files and then simulates the susceptibility and infection of individuals within each geographic node. The simulation type controls the transmission mechanism of the disease. After the simulation is complete, the Eradication.exe produces output reports that describe various aspects of the disease dynamics within the modeled population.
You can run simulations locally or on a multi-core cluster. Because the model is stochastic, you must run simulations multiple times to produce scientifically valid results.
Depending on how you run the simulation, other programs or environments external to EMOD may also output error and logging files.
Output files¶
After the simulation finishes, a reporter extracts simulation data, aggregates it, and outputs it to a file (known as an output report). Most of the reports are also JSON files, the most important of which is InsetChart.json. The InsetChart.json file provides simulation-wide averages of disease prevalence at each time step.
See the topics listed below for a more detailed explanation of each of these system components.
Input data files¶
EMOD accepts the following categories of input data files that contain the relatively fixed information about the population within each geographic node. For example, the number of individuals, geographic data, climate, demographics, and migration data. This is in contrast to the demographic, geographic, and migration parameters in the configuration file that control simulation-wide qualities, such as enabling air migration across all nodes in the simulation.
Although a demographics file is the only required input data file, additional files are generally needed for a realistic simulation. The demographics files use JavaScript Object Notation (JSON). The other input data files use both a JSON file for metadata and an associated binary file that contains the actual data.
The Institute for Disease Modeling (IDM) provides collections of input data files for many different locations in the world for download on GitHub. Except for the demographics file, you will typically use these input data files in their default state. See Use input data files for more information.
Demographics¶
Demographics files are JSON formatted files containing information on the demographics of the population in a geographical region to simulate. For example, the number of individuals and the distribution for age, gender, immunity, risk, and mortality.
In addition, demographics files are useful for creating heterogeneous groups within a population. You can define values for accessibility, age, geography, risk, and other properties and assign individuals to groups based on those property values. For example, you might want to divide a population up into different bins based on age so you can target interventions to individuals in a particular age range. Another common use is to configure treatment coverage to be higher for individuals in easily accessible regions and lower for individuals in areas that are difficult to access.
EMOD assumes homogeneous mixing and disease transmission for the generic simulation type. You can use the HINT feature to add heterogeneous transmission to your generic model. You cannot manually configure heterogeneous transmission using HINT with other simulation types because the heterogeneity in transmission for specific diseases and disease classes is already configured by the simulation type.
You can specify multiple demographics files, which function as a “base layer” file and one or more “overlay” files that override the base layer configuration. Overlay files can change the value of parameters already specified in the base layer or add new parameters. Support for multiple demographics layers allows for the following scenarios:
Separating different sets of parameters and values into individual layers (for example, to separate those that are useful for specific diseases into different layers)
Adding new parameters for a simulation into a new layer for easier prototyping
Overriding certain parameters of interest in a new layer
Overriding certain parameters for a particular sub-region
Simulating subsets of a larger region for which input data files have been constructed
Migration¶
Migration files describe the rate of migration of individuals in and out of a geographic node. There are four types of migration files that can be used by EMOD: local migration, regional migration, air migration, and sea migration.
For all types, migration data is contained in a set of two files, a JSON metadata file with header information and a binary data file with the actual migration data. Both files are required. The basic file structure is identical for all types, with the only exception being the number of columns per row allowed to each type.
Local migration describes the foot travel of people into and out of adjacent nodes. A local migration file is required for simulations that support more than one node.
Regional migration describes migration via a road or rail network. If a node is not part of the network, the regional migration of individuals to and from that node considers the closest road hub city. When you create the migration file, you must create a Voronoi tiling based on road hubs of the region, with each non-hub connected to the hub of its tile.
Air migration describes migration via airplane travel. It is usually required for simulations of an entire country or larger geographies.
Sea migration describes migration via ship.
Climate¶
There are two general types of climate files usable by EMOD: climate files generated through actual data, referred to as “climate by data,” and climate files generated from the Koppen classification system, referred to as “climate by Koppen.”
For both types, climate data is contained in a set of two files, a JSON metadata file with header information and a binary file that contains the actual climate data. Both files are required.
Climate by data files contain real data gathered from weather stations in the region to be simulated. This includes rainfall, temperature, relative humidity, and so on.
Climate by Koppen files contain the Koppen classifier for the region. The Koppen classification system is one of the most widely used climate classification systems. It makes the assumption that native vegetation is the best reflection of climate.
Load balancing¶
When running a simulation on a multi-core HPC cluster, you can include a load balancing file to control how the computing load is distributed across the cluster. The load balancing file is a binary file that allocates the simulation of each geographic node to cores in the cluster. If no file is submitted, EMOD allocates nodes to cores according to a checkerboard pattern.
Configuration file¶
The primary means of configuring the disease simulation is the configuration file. This required file is a JavaScript Object Notation (JSON) formatted file that is typically named config.json. The configuration file controls many different aspects of the simulation. For example,
The names of the campaign file and input data files to use
Simulation-wide demographics, climate, and migration data
General disease attributes such as infectivity, immunity, mortality, and so on
Attributes specific to the disease type being modeled, including treatment efficacy
The reports to output from the simulation
IDM provides complete simulation scenarios in the Regression directory on GitHub. Within each of the simulation subdirectories, there is a config.json file. The Scenarios subdirectory contains the files used in the Quick Start training tutorials.
The following is an example of configuration file.
{
"parameters": {
"Acquisition_Blocking_Immunity_Decay_Rate": 0.1,
"Acquisition_Blocking_Immunity_Duration_Before_Decay": 60,
"Age_Initialization_Distribution_Type": "DISTRIBUTION_SIMPLE",
"Animal_Reservoir_Type": "NO_ZOONOSIS",
"Base_Incubation_Period": 0,
"Base_Individual_Sample_Rate": 1,
"Base_Infectious_Period": 4,
"Base_Infectivity": 3.5,
"Base_Mortality": 0,
"Base_Population_Scale_Factor": 1,
"Birth_Rate_Dependence": "POPULATION_DEP_RATE",
"Birth_Rate_Time_Dependence": "NONE",
"Burnin_Cache_Mode": "none",
"Burnin_Cache_Period": 0,
"Burnin_Name": "",
"Campaign_Filename": "campaign.json",
"Climate_Model": "CLIMATE_OFF",
"Config_Name": "00_DEFAULT",
"Custom_Reports_Filename": "NoCustomReports",
"Death_Rate_Dependence": "NONDISEASE_MORTALITY_OFF",
"Default_Geography_Initial_Node_Population": 1000,
"Default_Geography_Torus_Size": 10,
"Demographics_Filenames": [
"../00_Default/demographics.json"
],
"Enable_Absolute_Time": "NO",
"Enable_Aging": 1,
"Enable_Birth": 1,
"Enable_Default_Reporting": 1,
"Enable_Default_Shedding_Function": 1,
"Enable_Demographics_Birth": 0,
"Enable_Demographics_Builtin": 0,
"Enable_Demographics_Gender": 1,
"Enable_Demographics_Other": 0,
"Enable_Demographics_Reporting": 1,
"Enable_Disease_Mortality": 0,
"Enable_Heterogeneous_Intranode_Transmission": 0,
"Enable_Immune_Decay": 0,
"Enable_Immunity": 1,
"Enable_Interventions": 1,
"Enable_Maternal_Transmission": 0,
"Enable_Property_Output": 0,
"Enable_Spatial_Output": 0,
"Enable_Superinfection": 0,
"Enable_Vital_Dynamics": 0,
"Geography": "",
"Immunity_Acquisition_Factor": 0,
"Immunity_Initialization_Distribution_Type": "DISTRIBUTION_OFF",
"Immunity_Mortality_Factor": 0,
"Immunity_Transmission_Factor": 0,
"Incubation_Period_Distribution": "FIXED_DURATION",
"Individual_Sampling_Type": "TRACK_ALL",
"Infection_Updates_Per_Timestep": 1,
"Infectious_Period_Distribution": "EXPONENTIAL_DURATION",
"Infectivity_Scale_Type": "CONSTANT_INFECTIVITY",
"Job_Node_Groups": "Chassis08",
"Job_Priority": "BELOWNORMAL",
"Listed_Events": [],
"Load_Balance_Filename": "",
"Local_Simulation": 0,
"Maternal_Transmission_Probability": 0,
"Max_Individual_Infections": 1,
"Max_Node_Population_Samples": 40,
"Migration_Model": "NO_MIGRATION",
"Minimum_Adult_Age_Years": 15,
"Mortality_Blocking_Immunity_Decay_Rate": 0.001,
"Mortality_Blocking_Immunity_Duration_Before_Decay": 60,
"Mortality_Time_Course": "DAILY_MORTALITY",
"Node_Grid_Size": 0.042,
"Num_Cores": 1,
"Number_Basestrains": 1,
"Number_Substrains": 1,
"PKPD_Model": "FIXED_DURATION_CONSTANT_EFFECT",
"Population_Density_C50": 30,
"Population_Density_Infectivity_Correction": "CONSTANT_INFECTIVITY",
"Population_Scale_Type": "USE_INPUT_FILE",
"Report_Event_Recorder": 0,
"Run_Number": 1,
"Sample_Rate_0_18mo": 1,
"Sample_Rate_10_14": 1,
"Sample_Rate_15_19": 1,
"Sample_Rate_18mo_4yr": 1,
"Sample_Rate_20_Plus": 1,
"Sample_Rate_5_9": 1,
"Sample_Rate_Birth": 1,
"Serialization_Test_Cycles": 0,
"Simulation_Duration": 90,
"Simulation_Timestep": 1,
"Simulation_Type": "GENERIC_SIM",
"Start_Time": 0,
"Susceptibility_Scale_Type": "CONSTANT_SUSCEPTIBILITY",
"Transmission_Blocking_Immunity_Decay_Rate": 0.1,
"Transmission_Blocking_Immunity_Duration_Before_Decay": 60,
"x_Air_Migration": 1,
"x_Birth": 1,
"x_Local_Migration": 1,
"x_Other_Mortality": 1,
"x_Population_Immunity": 1,
"x_Regional_Migration": 1,
"x_Sea_Migration": 1,
"x_Temporary_Larval_Habitat": 1
}
}
Overlay file¶
As you work more with EMOD, you may want to change the values of a view parameters of interest while keeping the rest constant, for example, when running simulation experiments or testing source code changes.
You have the option of creating an overlay file to keep configuration parameters of interest in a separate file from the complete configuration file that contains default values. These files can be flattened into a single file and the values in the overlay file will override those in the default file.
See Configuration parameters for a comprehensive list and description of all parameters available to use in the configuration file for this simulation type.
Campaign file¶
The campaign file is an optional file that distributes outbreaks and contains all parameters that define the collection of interventions that make up a disease eradication campaign. For example,
When, how, and who to test for a disease
When, how, and who to distribute treatments to
The vaccines or other preventative interventions to use
The medications to use for treatment
Like the configuration file, the campaign file is a JavaScript Object Notation (JSON) formatted file. It is hierarchically organized into logical groups of parameters that can have arbitrary levels of nesting. For some interventions, there can be a very complex hierarchical structure, including recursion. Typically, the file is named campaign.json. The relative path to this file is specified by Campaign_Filename in the configuration file.
To distribute an intervention, you must configure the following:
- campaign event
A JSON object that determines when and where an intervention is distributed during a campaign.
- event coordinator
A JSON object that determines who will receive a particular intervention during a campaign.
- intervention
A JSON object that determines what will be distributed to reduce the spread of a disease. An intervention can be distributed either to an individual (such as a vaccine, drug, or bednet) or to a node (such as a larvicide). Sometimes this can be an intermediate intervention that schedules another intervention.
IDM provides complete simulation scenarios in the Regression directory on GitHub. Within each of the simulation subdirectories, there is a campaign.json file. The Scenarios subdirectory contains the files used in the Quick Start tutorials.
Overlay file¶
As you work more with EMOD, you may want to change the values of a view parameters of interest while keeping the rest constant, for example, when running simulation experiments or testing source code changes.
You have the option of creating an overlay file to keep campaign parameters of interest in a separate file from the complete campaign file that contains default values. These files can be flattened into a single file and the values in the overlay file will override those in the default file.
See Campaign parameters for a comprehensive list and description of all parameters available to use in the campaign file for this simulation type.
Running a simulation¶
The EMOD executable (Eradication.exe) consumes the input data files, configuration file, and, optionally, campaign file to run simulations that model disease dynamics and campaign efficacy. You have a few different options for running simulations. The option you choose will depend upon whether you want to run one or more simulations, to run simulations locally or on a remote cluster (for large simulations or multiple simulations), or to run simulations for debugging the source code. This topic briefly describes the different options you have for running simulations.
Run a single simulation¶
You will generally want to run simulations using the EMOD command-line options. This will run a single simulation and put the output files in a local directory.
You must either download the latest version of Eradication.exe from GitHub or clone the EMOD source from GitHub and build Eradication.exe yourself. This gives you access to the latest features and parameters for EMOD.
Run multiple simulations¶
Because the EMOD model is stochastic, simulations must be run multiple times to return scientifically valid results. Therefore, you have the following options to run multiple simulations at a time, either locally or remotely on a high-performance computing (HPC) cluster. Generally, only small simulations should be run locally.
Many of these options are scripting languages that you can also use to modify the files consumed by EMOD, simplifying your workflow when running many simulations.
Run simulations for debugging¶
If you are helping advance the EMOD model by contributing to source code, there are other options for running simulations that provide debugging support. These options for running simulations are not recommended if you are not modifying the source code.
You can run a simulation locally from Visual Studio using the built-in debugger. You will be able to put in breakpoints and step through the code while inspecting the values of different state variables throughout the simulation.
You can use regression_test.py in the GitHub Regression directory to run multiple simulations on a cluster, including running the suite of regression tests run by the IDM testing team.
Output files¶
After running a simulation, the simulation data is extracted, aggregated, and saved as an output report. Depending on your configuration, one or more output reports will be created, each of which summarize different data from the simulation. Output reports can be in JSON, CSV, or binary file formats.
The EMOD functionality that produces an output report is known as a reporter. EMOD provides several built-in reporters for outputting data from simulations. By default, EMOD will always generate the report InsetChart.json, which contains the simulation-wide average disease prevalence by time step. If none of the provided reports generates the output report that you require, you can create a custom reporter.
If you want to visualize the data output from an EMOD simulation, you must use graphing software to plot the output reports. In addition to output reports, EMOD will generate error and logging files to help troubleshoot any issues you may encounter.
Introduction to disease modeling¶
To understand the complex dynamics underlying disease transmission, epidemiologists utilize a set of models called compartmental models. Developed in the early 20th century, these models stratify a population into groups, generally based on their risk or infection status. Underlying these models is a system of differential equations, so the model is able to track the population as a function of time. Further, the models track the number of people in each category, are able to predict the various properties of pathogen spread, can estimate the duration of epidemics, and can be used to understand how different situations or interventions can impact the outcome of pathogen spread. To see a typical plot of a population in SIR conditions, see the plot in Types of compartmental models. If you would like a more in-depth introduction to epidemiology and disease modeling, you may want to take the Epidemics course from The Pennsylvania State University through Coursera.
An agent-based model (ABM) is another powerful tool used to help understand the complexity inherent in disease transmission systems. These models form a type of “microscale model,” where they simulate the simultaneous interactions of agents in an effort to re-crate complex phenomena. Each agent (such as a human, or vector) can be assigned a variety of “properties” (for example, age, gender, etc), and their behavior and interactions with one another are determined by using decision rules. These models have strong predictive power, and are able to leverage spatial and temporal dynamics. Further, complex environments can be developed in which the agents act, and agents may “learn” from interactions or “adapt” to their environment. As a result, ABMs are excellent for identifying “emerging properties” of the system: patterns that are not explicitly modeled, but instead occur as a consequence of the rules that govern the agents.
Agent-based models are extensively used in epidemiology due to their predictive power in modeling the spread (or conversely, control) of epidemics. A popular type of ABM for this is one in which each agent’s rules follow the dynamics specified in the compartmental models, where each agent flows through the compartments as a function of both “within-host” rules (such as duration of infection) and interactions between agents (such as becoming infected when coming into contact with an infectious agent). By combining the epidemiological basis of compartmental models with the flexibility of an agent-based model, this type of ABM is quite powerful due to their ability to simultaneously address the ecology, epidemiology, and pathology of complex systems.
Why use disease modeling¶
Disease models play an important role in understanding and managing the transmission dynamics of various pathogens. We can use them to describe the spatial and temporal patterns of disease prevalence, as well as to explore or better understand the factors that influence infection incidence. Modeling is a key step in understanding what treatments and interventions can be most effective, how cost-effective these approaches may be, and what specific factors need to be considered when trying to eradicate disease.
Types of compartmental models¶
The following diagrams illustrate common compartmental models. See the Generic simulation tutorials in the Training documentation set for more information, as well as exercises on running EMOD simulations utilizing these models.
SIR Plot, showing typical categorization of a population into Susceptible, Infectious, and Recovered states.¶
Definitions¶
- Susceptible
Individual is able to become infected.
- Exposed
Individual has been infected with a pathogen, but due to the pathogen’s incubation period, is not yet infectious.
- Infectious
Individual is infected with a pathogen and is capable of transmitting the pathogen to others.
- Recovered
Individual is either no longer infectious, or “removed” from the population.
SI (Susceptible - Infectious) model¶
In this situation, people never leave the infectious state and have life-long infections. For example, herpes is a disease with life-long infectiousness. The dashed line shows how the model becomes an SIS (Susceptible - Infectious - Susceptible) model, where infection does not confer immunity (or there is waning immunity). Individuals have repeat or reoccurring infections, and infected individuals return to the susceptible state. For example, sexually transmitted diseases such as gonorrhea or chlamydia fall into this group.
SI - SIS model¶
SIR (Susceptible - Infectious - Recovered) model¶
In this category, individuals in the recovered state gain immunity to the pathogen. For example, measles, mumps, rubella, and pertussis may be modeled using the SIR framework. The dashed line shows how the model becomes an SIRS (Susceptible - Infectious - Recovered - Susceptible) model, where recovery does not confer life-long immunity, and individuals may become susceptible again.
SIR - SIRS model¶
SEIR (Susceptible - Exposed - Infectious - Recovered) model¶
In this category, individuals experience a long incubation duration (the “exposed” category), such that the individual is infected but not yet infectious. For example chicken pox, and even vector-borne diseases such as Dengue Hemorrhagic fever have a long incubation duration where the individual cannot yet transmit the pathogen to others. The dashed line shows how the model becomes an SEIRS (Susceptible - Exposed - Infectious - Recovered - Susceptible) model, where recovered people may become susceptible again (recovery does not confer life-long immunity). For example, rotovirus and malaria are diseases with long incubation durations, and where infection only confers temporary immunity.
SEIR - SEIRS model¶
Fundamental concepts in epidemiology and disease modeling¶
- Basic reproductive number (R0)
The average number of secondary infections generated by the first infectious individual in a population of completely susceptible individuals. R0=S*L* \(\beta\) (where S = the number of susceptible hosts, L = length of infection, and \(\beta\) = transmissibility). When R0> 1, disease will spread. It is essentially a measure of the expected or average outcome of transmission. This is the threshold parameter used to determine whether or not an epidemic will occur, and determines:
The initial rate of increase of an epidemic (the exponential growth phase).
The final size of an epidemic (what fraction of susceptibles will be infected).
The endemic equilibrium fraction of susceptibles in a population (=1/ R0).
The critical vaccination threshold, which is equal to 1-(1/ R0), and determines the number of people that must be vaccinated to prevent the spread of a pathogen.
- Disability-adjusted life years (DALY)
The number of years of life lost due to premature mortality plus the years lost due to disability while infected. Used to quantify the burden of disease.
- Epidemic
An outbreak of an infectious disease, such that a greater number of individuals than normal has the disease. Epidemics have very high R0 (Recall R0>1 for a disease to spread) and are often associated with acute, highly transmissible pathogens that can be directly transmitted. Further, pathogens with lower infectious periods create more explosive epidemics.
To control epidemics, it is necessary to reduce R0. This can be done by:
Reducing transmissibility.
Decreasing the number of susceptibles (by vaccination, for example).
Decreasing the mean number of contacts or the transmissibility, such as by improving sanitation, or limiting the number of interactions sick people have with healthy people.
Reducing length of infectious period.
- Force of infection (FoI)
A measure of the degree to which an infected individual can spread infection; the per-capita rate at which susceptibles contract infection. Typically increases with transmissibility and prevalence of infection.
- Herd immunity
Vaccines protect individuals, but also provide indirect protection to anyone those vaccinated people may have infected. The portion of the population that needs to be immunized in order to achieve herd immunity is P > 1 – (1/ R0), where P = proportion vaccinated * vaccine efficacy.
- Incidence (of a pathogen)
The number of new cases or infections in a given time period.
- Immune
Unable to become infected/infectious
- Prevalence (of a pathogen)
The proportion of a population that is infectious at any given time.
- Transmissibility (\(\beta\))
Also known as the effective contact rate, is the product of the contact rate and the probability of transmission per contact.
- Virulence
The capacity of a pathogen to produce disease. It is proportional to parasitemia, or the number of circulating copies of the pathogen in the host. The higher the virulence (given contact between S and I individuals), the more likely transmission is to occur. However, higher virulence means contact may be less likely as infected hosts show more symptoms of the disease. There is a trade-off that occurs between high transmissibility and disease-induced mortality.
How to use the software¶
This section contains detailed procedural information about how to use Epidemiological MODeling software (EMOD). For example, how to download input data files, how to set up groups in a population, how to create a configuration file and campaign file, how to run simulations both locally and on a remote HPC cluster, and more.
The architecture diagram below shows, at a high level, how the system functions.
High-level EMOD system architecture¶
For more information about how each component of the system fits together, see Introduction to the software. That section provides a conceptual overview of EMOD and describes each required or optional component for running a simulation.
Use input data files¶
The input data files contain the relatively fixed information about the population within each geographic node. For example, the number of individuals, geography, climate, demographics, and migration data. This topic describes how to download and use the input data files. These are in contrast to the demographic, geography, and migration parameters in the configuration file that control simulation-wide qualities, such as enabling air migration across all nodes in the simulation.
Except for the demographics file, you will generally use input data files without modifying them in any way. Only the demographics file is required, though migration files may be required for multi-node simulations. See Input data files for an overview of each of the different input files, including which are required for different simulations. See Input data file structure for reference information about the structure of each of these files.
Download input files¶
The EMOD-InputData repository uses large file storage (LFS) to manage the binaries and large JavaScript Object Notation (JSON) files. A standard Git clone of the repository will only retrieve the metadata for these files managed with LFS. To retrieve the actual data, follow the steps below.
Install the Git LFS plugin, if it is not already installed.
For Windows users, download the plugin from https://git-lfs.github.com.
For CentOS on Azure users, the plugin is included with the PrepareLinuxEnvironment.sh script.
Using a Git client or Command Prompt window, clone the input data repository to retrieve the metadata:
git clone https://github.com/InstituteforDiseaseModeling/EMOD-InputData.git
Navigate to the directory where you downloaded the metadata for the input data files.
Cache the actual data on your local machine:
git lfs fetch
Replace the metadata in the files with the actual contents:
git lfs checkout
Specify input files in the configuration file¶
Follow the steps below to specify which input files to use in a simulation. Only the demographics file is required, though additional files are generally needed for a realistic simulation.
Place all input files for a simulation in the same directory. You will specify this directory when you run a simulation. See Run simulations for more information.
In your configuration file, specify the path to each of these files, relative to the directory above, in the appropriate parameter. Generally, these parameters are appended with “_Filename” or “_Filenames”.
For example, the example below shows the relevant portion of a configuration file. See :doc :parameter-configuration for a complete list of the parameters.
{
"parameters": {
"Air_Temperature_Filename": "Namawala_single_node_air_temperature_daily.bin",
"Air_Temperature_Offset": 0,
"Air_Temperature_Variance": 2,
"Base_Rainfall": 100,
"Campaign_Filename": "campaign.json",
"Climate_Model": "CLIMATE_BY_DATA",
"Climate_Update_Resolution": "CLIMATE_UPDATE_DAY",
"Config_Name": "VectorAndMalaria_5_Namawala_Vector_ITNs",
"Demographics_Filenames": [
"Namawala_single_node_demographics.json"
],
"Geography": "Namawala",
"Land_Temperature_Filename": "Namawala_single_node_land_temperature_daily.bin",
"Land_Temperature_Offset": 0,
"Land_Temperature_Variance": 2,
"Load_Balance_Filename": "",
"Rainfall_Filename": "Namawala_single_node_rainfall_daily.bin",
"Rainfall_In_mm_To_Fill_Swamp": 1000.0,
"Rainfall_Scale_Factor": 1,
"Relative_Humidity_Filename": "Namawala_single_node_relative_humidity_daily.bin",
"Relative_Humidity_Scale_Factor": 1,
"Relative_Humidity_Variance": 0.05
}
}
Modify demographics files¶
The demographics files provided by IDM generally contain information about prevalence, immunity, risk, population size, and more for a geographic region. However, you will almost certainly want to modify the file to provide more detail or to set up groups within a population to more accurately model heterogeneous populations in terms of transmission, group transitions, or targeted interventions.
The demographics file is the only required input data file, with one exception. You have the option to run a simulation without a demographics file if you set Enable_Demographics_Builtin to 1 in the configuration file. However, this option is primarily used for software testing. It will not provide meaningful simulation data as it does not represent the population of a real geographic location.
Use demographics overlay files¶
You can use multiple demographic files when setting up a simulation. The “base layer” file contains default parameter settings and one or more overlay files contains additional parameters or different parameter values that override the values in the default file. This topic describes how to set up the base layer and overlay files.
Demographic overlay files allow you to do the following:
Separate different sets of parameters and values into individual overlays (for example, to separate those that are useful for specific diseases into different overlay files)
Add new parameters for a simulation into an overlay for easier prototyping
Override certain parameters of interest in an overlay
Override certain parameters for a particular sub-region
Simulating subsets of a larger region for which input data files have been constructed
Select the demographics file to use as the base layer file. All nodes to be included in the simulation must be listed in this file.
In the metadata, make note of the IdReference value.
You may change this value if you desire, but all input files for a simulation must have the same IdReference value. For more information about this parameter and the structure of demographics files in general, see Demographics file structure and parameters.
Create one or more overlay files. Keep the following things in mind:
In the metadata, the value for IdReference must match the value in the base layer file and all other input data files.
Any nodes listed in an overlay but not in the base layer will not be simulated.
If the demographics files include any JSON array elements, the entire array is overridden. You cannot add or remove individual elements in an array using an overlay file.
Overriding a parameter value on a node will not affect the other parameter values applied to that node.
Values set in the Defaults section of an overlay will be applied to all nodes listed in that file, not all nodes in the entire simulation. Therefore, an overlay file that includes a Defaults section but no Nodes section will not have any effect.
Place all demographics files in the directory where the other input data files are stored.
In the configuration file, set Demographics_Filenames to an array that contains a comma-delimited list of demographics files, listing the base layer file first.
An example base layer demographics file and an overlay file is below. You can see that the overlay adds the TransmissionMatrix for Heterogeneous Intra-Node Transmission (HINT) to only three of the five nodes (which correspond to Washington state counties).
{
"Metadata": {
"Author": "ewenger",
"NodeCount": 5,
"Tool": "table_to_demographics.py",
"IdReference": "SampleContent",
"DateCreated": "2013-08-01 15:37:16.853000"
},
"Defaults": {
"NodeAttributes": {
"BirthRate_DESCRIPTION": "Replacement of stable age distribution: Birth_Rate_Dependence=DEMOGRAPHIC_DEP_RATE (i.e. 14-45 year-old PossibleMothers)",
"BirthRate": 0.00017675,
"Airport": 0,
"Region": 1,
"Altitude": 0,
"Seaport": 0
},
"IndividualAttributes": {
"AgeDistribution_DESCRIPTION": "Box between age 0 and 60 years: Age_Initialization_Distribution_Type=DISTRIBUTION_SIMPLE",
"AgeDistributionFlag": 1,
"AgeDistribution1": 0,
"AgeDistribution2": 21900,
"PrevalenceDistribution_DESCRIPTION": "No initial infections",
"PrevalenceDistributionFlag": 0,
"PrevalenceDistribution1": 0,
"PrevalenceDistribution2": 0,
"RiskDistributionFlag": 0,
"RiskDistribution1": 1,
"RiskDistribution2": 0,
"ImmunityDistributionFlag": 0,
"ImmunityDistribution1": 1,
"ImmunityDistribution2": 0,
"MigrationHeterogeneityDistributionFlag": 0,
"MigrationHeterogeneityDistribution1": 1,
"MigrationHeterogeneityDistribution2": 0,
"MortalityDistribution_DESCRIPTION": "WA state (1999-2010). Source: wonder.cdc.gov",
"MortalityDistribution": {
"NumDistributionAxes": 2,
"AxisNames": [ "gender", "age" ],
"AxisUnits": [ "male=0,female=1", "years" ],
"AxisScaleFactors": [ 1, 365 ],
"NumPopulationGroups": [ 2, 15 ],
"PopulationGroups": [
[ 0, 1 ],
[ 0, 1, 2.5, 7.5, 12.5, 17.5, 22.5, 30, 40, 50, 60, 70, 80, 90, 120 ]
],
"ResultUnits": "annual deaths per 1000 individuals",
"ResultScaleFactor": 0.00000273972602739726027397260273973,
"ResultValues": [
[ 4.8, 0.2, 0.1, 0.2, 0.5, 1.1, 1.1, 1.7, 4.1, 9.2, 19.8, 53.7, 154.2, 1000, 1000 ],
[ 4.2, 0.2, 0.2, 0.2, 0.2, 0.4, 0.5, 1.1, 2.6, 5.5, 14.3, 40.1, 129.3, 1000, 1000 ]
]
}
}
},
"Nodes": [
{
"County": "Adams",
"NodeAttributes": {
"Latitude": 47.1274,
"InitialPopulation": 19200,
"Longitude": -118.38
},
"NodeID": 1
},
{
"County": "Asotin",
"NodeAttributes": {
"Latitude": 46.3393,
"InitialPopulation": 21800,
"Longitude": -117.0482
},
"NodeID": 2
},
{
"County": "Benton",
"NodeAttributes": {
"Latitude": 46.2068,
"InitialPopulation": 183400,
"Longitude": -119.7689
},
"NodeID": 3
},
{
"County": "Chelan",
"NodeAttributes": {
"Latitude": 47.4235,
"InitialPopulation": 73600,
"Longitude": -120.3103
},
"NodeID": 4
},
{
"County": "Clallam",
"NodeAttributes": {
"Latitude": 48.1181,
"InitialPopulation": 72350,
"Longitude": -123.4307
},
"NodeID": 5
}
]
}
{
"Metadata": {
"Author": "cwiswell",
"NodeCount": 10,
"Tool": "table_to_demographics.py",
"IdReference": "SampleContent",
"DateCreated": "2013-08-01 15:37:16.853000"
},
"Defaults": {
"IndividualProperties": [
{
"Property": "Accessibility",
"Values": [ "VaccineTake", "VaccineRefuse"],
"Initial_Distribution": [ 0.85, 0.15],
"Transitions": [],
"TransmissionMatrix": {
"Route": "Contact",
"Matrix": [
[1.1, 0.3],
[0.3, 5.0]
]
}
},
{
"Property": "Age_Bin",
"Age_Bin_Edges_In_Years": [ 0, 5, 13, -1 ],
"Transitions": [],
"TransmissionMatrix": {
"Route": "Contact",
"Matrix": [
[1.4, 1.0, 1.0],
[1.0, 2.5, 0.7],
[1.0, 0.7, 1.0]
]
}
}
]
},
"Nodes": [
{
"NodeID": 1
},
{
"NodeID": 3
},
{
"NodeID": 5
}
]
}
Configure heterogeneous populations¶
Demographics files are used to add heterogeneity to a population. You can define the initial distribution to use for age, prevalence, risk and more. You can also define values for accessibility, age, geography, risk, and other properties and assign individuals to groups based on those property values. This topic describes how to configure the population distribution and define heterogeneous groups.
After you set up the groups, you may want to add more parameters to the demographics file to configure how individuals transition into and out of groups, how transmission occurs between different groups, or how to target interventions to specific groups. For more information about the parameters and structure of demographics files, see Demographics file structure and parameters.
For example, you might want to divide a population up into different groups based on age so you can target interventions to individuals in a particular age range. Another common use is to configure treatment coverage to be higher for individuals in areas that are easy to access and lower for individuals in areas that are difficult to access. For more information on targeting interventions to particular groups, see Target interventions to nodes or groups. For information on how tp configure disease transmission among groups created with any of the properties, see Configure heterogeneous disease transmission.
Nodes are subregions within the geographic area being modeled. Values in the Defaults section of the demographics file will be applied to all nodes, while values in the Nodes section will be applied only to the specified node. Node-level settings take precedence.
You can configure attribute distributions in the population using the IndividualAttributes parameter in the demographics file. The initial value for an individual is a randomly selected value from the distribution. For example, if you use a uniform distribution to initialize age, the initial ages of individuals in the simulation will be evenly distributed between some minimum and maximum value.
In the demographics file, add the IndividualAttributes parameter and assign it an empty JSON object. If you want the groups to apply to all nodes, add it in the Defaults section; if you want the groups to be applied to specific nodes, add it to the Nodes section.
Within this object, add the parameters for different attributes (age, prevalence, etc.) and assign values for available distributions (constant, Gaussian, etc.) See Demographics file structure and parameters for more information about the available parameters.
The example below shows how to set up the age distribution for all nodes in a simulation.
{
"Defaults": {
"IndividualAttributes": {
"AgeDistributionFlag": 3,
"AgeDistribution1": 0.8,
"AgeDistribution2": 0.1
}
}
}
Assigning individuals to different groups based on properties, such as accessibility or risk, uses the IndividualProperties parameter in the demographics file. See Demographics file structure and parameters for a list of supported properties. The values you assign to properties are use-defined and can be applied to individuals in all nodes or only in particular nodes in a simulation.
In the demographics file, add the IndividualProperties parameter and set it to an empty array. If you want the groups to apply to all nodes, add it in the Defaults section; if you want the groups to be applied to specific nodes, add it to the Nodes section.
In the array, add an empty JSON object. Within it, do the following:
Add the Property parameter and set it to one of the supported values.
Add the Values parameter and set it to an array of possible values that can be assigned to individuals. You can define any value here.
Add the Initial_Distribution parameter and set it to an array of numbers that add up to 1. This configures the initial distribution of individuals assigned to each of the groups.
To define how individuals transition into and a out of each group, add the Transitions parameter and set it to it an empty array. Within it, do the following:
Add an empty JSON object and set parameters that define the group that individuals transition from, the group they transition to, the event that triggers the transition, the probability of transition, and more. See doc:parameter-demographics for a list of supported Transitions parameters and values.
If you want to add another property and associated groups, add a new JSON object in the IndividualProperties array as above.
Note
Multiple properties must be defined in one file. They can be defined in either the base layer demographics file or an overlay file, but they cannot be split between the files. The maximum number of property types that can be added is two.
Creating groups based on age ranges works a little differently than creating groups based on other properties. Age_Bin is tied to the simulated age of an individual rather than being an independent property. Some of its characteristics, such as initial distribution and transitions, are dependent on information from the demographics file and EMOD that manages individual aging during the simulation. Because of this, the parameters and structures in IndividualProperties are slightly different.
In the demographics file, add the IndividualProperties parameter and set it to an empty array. If you want the groups to apply to all nodes, add it in the Defaults section; if you want the groups to be applied to specific nodes, add it to the Nodes section.
In the array, add an empty JSON object. Within it, do the following:
Add the Property parameter and set it to “Age_Bin”.
Add the Age_Bin_Edges_In_Years parameter and set it to an array that contains a comma- delimited list of integers in ascending order that define the boundaries used for each of the age bins, in years. The first number must always be 0 (zero) to indicate the age at birth and the last number must be -1 to indicate the maximum age in the simulation.
To define how individuals transition into and a out of each group, add the Transitions parameter and set it to it an empty array. Aging during the simulation will be handled by EMOD.
The example below shows how to set up several groups based on disease risk and physical place, and how to move individuals among these groups. It also defines three age bins: 0 to 5 years, older than 5 to 13, and older than 13 to the maximum age.
{
"Defaults": {
"IndividualProperties": [{
"Property": "Risk",
"Values": ["Low", "Medium", "High"],
"Initial_Distribution": [0.7, 0.2, 0.1],
"Transitions": [{
"From": "High",
"To": "Medium",
"Type": "At_Age",
"Coverage": 1,
"Probability_Per_Timestep": 0.3,
"Timestep_Restriction": 20,
"Age_In_Years": 5,
"Timesteps_Until_Reversion": 0
}, {
"From": "Medium",
"To": "Low",
"Type": "At_Age",
"Coverage": 1,
"Probability_Per_Timestep": 0.3,
"Timestep_Restriction": 20,
"Age_In_Years": 12,
"Timesteps_Until_Reversion": 0
}]
}, {
"Property": "Place",
"Values": ["Community", "School", "Work", "Vacation"],
"Initial_Distribution": [0.3, 0.2, 0.4, 0.1],
"Transitions": [{
"From": "School",
"To": "Vacation",
"Type": "At_Timestep",
"Coverage": 1,
"Timestep_Restriction": {
"Start": 20
},
"Age_In_Years_Restriction": {},
"Probability_Per_Timestep": 1,
"Timesteps_Until_Reversion": 20
}]
}, {
"Property": "Age_Bin",
"Age_Bin_Edges_In_Years": [0, 5, 13, -1],
"Transitions": []
}]
}
}
Configure heterogeneous disease transmission¶
Only generic simulations can use Heterogeneous Intra-Node Transmission (HINT) to manually configure heterogeneous disease transmission within each node. All other simulation types have preconfigured solutions for heterogeneous disease transmission based on the biological processes of the disease being modeled. Because HINT cannot be used with this simulation type, the parameter Enable_Heterogeneous_Intranode_Transmission in the configuration file must be set to 0 (zero).
This topic describes the mathematics governing the Heterogeneous Intra-Node Transmission (HINT) feature. However, only generic simulations can use HINT to manually configure heterogeneous disease transmission within each node. All other simulation types have preconfigured solutions for heterogeneous disease transmission based on the biological processes of the disease being modeled. See the generic disease documentation for more information about HINT.
Create a configuration file¶
You define the general configuration and processing of a simulation through a JSON-formatted configuration file, typically called config.json. Two forms of the simulation configuration file are used: a flattened version and a hierarchical version. This topic describes how to create a configuration file of other form. The EMOD executable (Eradication.exe) and Eradication binary requires a flattened version of the simulation configuration file. The hierarchical version allows you to organize parameters into logical groups, making them easier to manage. If you use hierarchical configuration files, you must flatten them prior to running a simulation.
The EMOD Regression directory contains many different subdirectories that contain configuration, campaign, and other associated files to run simulations that represent real-world scenarios. Within the each subdirectory, there is usually a hierarchical overlay file (param_overrides.json) and a flattened file (config.json), which has been created by combining param_overrides.json with one of the default files in Regression/defaults. The default files are also organized hierarchically. The naming of these files is an arbitrary convention used at IDM; you may name this files anything you choose. See Use configuration overlay files for more information about flattening hierarchical files.
For a complete list of configuration parameters that are available to use with this simulation type, see Configuration parameters. For more information about JSON, see EMOD parameter reference.
Flattened configuration file¶
A flattened configuration file is generally a single-depth JSON file with configuration parameters listed alphabetically. This is the configuration file format that Eradication.exe requires for running simulations.
However, there may be some hierarchical elements in the flattened version. For example, Vector_Species_Params and Malaria_Drug_Params have nested JSON objects.
Below is an example of a flattened configuration file:
{
"parameters": {
"Acquisition_Blocking_Immunity_Decay_Rate": 0.1,
"Acquisition_Blocking_Immunity_Duration_Before_Decay": 60,
"Age_Initialization_Distribution_Type": "DISTRIBUTION_SIMPLE",
"Animal_Reservoir_Type": "NO_ZOONOSIS",
"Base_Incubation_Period": 0,
"Base_Individual_Sample_Rate": 1,
"Base_Infectious_Period": 4,
"Base_Infectivity": 3.5,
"Base_Mortality": 0,
"Base_Population_Scale_Factor": 1,
"Birth_Rate_Dependence": "POPULATION_DEP_RATE",
"Birth_Rate_Time_Dependence": "NONE",
"Burnin_Cache_Mode": "none",
"Burnin_Cache_Period": 0,
"Burnin_Name": "",
"Campaign_Filename": "campaign.json",
"Climate_Model": "CLIMATE_OFF",
"Config_Name": "00_DEFAULT",
"Custom_Reports_Filename": "NoCustomReports",
"Death_Rate_Dependence": "NONDISEASE_MORTALITY_OFF",
"Default_Geography_Initial_Node_Population": 1000,
"Default_Geography_Torus_Size": 10,
"Demographics_Filenames": [
"../00_Default/demographics.json"
],
"Enable_Absolute_Time": "NO",
"Enable_Aging": 1,
"Enable_Birth": 1,
"Enable_Default_Reporting": 1,
"Enable_Default_Shedding_Function": 1,
"Enable_Demographics_Birth": 0,
"Enable_Demographics_Builtin": 0,
"Enable_Demographics_Gender": 1,
"Enable_Demographics_Other": 0,
"Enable_Demographics_Reporting": 1,
"Enable_Disease_Mortality": 0,
"Enable_Heterogeneous_Intranode_Transmission": 0,
"Enable_Immune_Decay": 0,
"Enable_Immunity": 1,
"Enable_Interventions": 1,
"Enable_Maternal_Transmission": 0,
"Enable_Property_Output": 0,
"Enable_Spatial_Output": 0,
"Enable_Superinfection": 0,
"Enable_Vital_Dynamics": 0,
"Geography": "",
"Immunity_Acquisition_Factor": 0,
"Immunity_Initialization_Distribution_Type": "DISTRIBUTION_OFF",
"Immunity_Mortality_Factor": 0,
"Immunity_Transmission_Factor": 0,
"Incubation_Period_Distribution": "FIXED_DURATION",
"Individual_Sampling_Type": "TRACK_ALL",
"Infection_Updates_Per_Timestep": 1,
"Infectious_Period_Distribution": "EXPONENTIAL_DURATION",
"Infectivity_Scale_Type": "CONSTANT_INFECTIVITY",
"Job_Node_Groups": "Chassis08",
"Job_Priority": "BELOWNORMAL",
"Listed_Events": [],
"Load_Balance_Filename": "",
"Local_Simulation": 0,
"Maternal_Transmission_Probability": 0,
"Max_Individual_Infections": 1,
"Max_Node_Population_Samples": 40,
"Migration_Model": "NO_MIGRATION",
"Minimum_Adult_Age_Years": 15,
"Mortality_Blocking_Immunity_Decay_Rate": 0.001,
"Mortality_Blocking_Immunity_Duration_Before_Decay": 60,
"Mortality_Time_Course": "DAILY_MORTALITY",
"Node_Grid_Size": 0.042,
"Num_Cores": 1,
"Number_Basestrains": 1,
"Number_Substrains": 1,
"PKPD_Model": "FIXED_DURATION_CONSTANT_EFFECT",
"Population_Density_C50": 30,
"Population_Density_Infectivity_Correction": "CONSTANT_INFECTIVITY",
"Population_Scale_Type": "USE_INPUT_FILE",
"Report_Event_Recorder": 0,
"Run_Number": 1,
"Sample_Rate_0_18mo": 1,
"Sample_Rate_10_14": 1,
"Sample_Rate_15_19": 1,
"Sample_Rate_18mo_4yr": 1,
"Sample_Rate_20_Plus": 1,
"Sample_Rate_5_9": 1,
"Sample_Rate_Birth": 1,
"Serialization_Test_Cycles": 0,
"Simulation_Duration": 90,
"Simulation_Timestep": 1,
"Simulation_Type": "GENERIC_SIM",
"Start_Time": 0,
"Susceptibility_Scale_Type": "CONSTANT_SUSCEPTIBILITY",
"Transmission_Blocking_Immunity_Decay_Rate": 0.1,
"Transmission_Blocking_Immunity_Duration_Before_Decay": 60,
"x_Air_Migration": 1,
"x_Birth": 1,
"x_Local_Migration": 1,
"x_Other_Mortality": 1,
"x_Population_Immunity": 1,
"x_Regional_Migration": 1,
"x_Sea_Migration": 1,
"x_Temporary_Larval_Habitat": 1
}
}
Hierarchical configuration file¶
The hierarchical version of a configuration file has a more complex structure. As a way of sorting parameters into logical groups, the parameters can be contained inside nested JSON objects. The names you use to create these logical categories are unimportant; the scripts used to flatten the files will search through the hierarchies and retain only the “leaf” values in the resulting flattened file. See Use configuration overlay files for more information on flattening files.
Below is an example of a hierarchical configuration file:
{
"parameters": {
"CAMPAIGNS": {
"Campaign_Filename": "campaign.json",
"Enable_Interventions": 1,
"Listed_Events": [],
"PKPD_Model": "FIXED_DURATION_CONSTANT_EFFECT"
},
"CLIMATE": {
"Climate_Model": "CLIMATE_OFF"
},
"DEMOGRAPHICS": {
"Age_Initialization_Distribution_Type": "DISTRIBUTION_SIMPLE",
"Base_Population_Scale_Factor": 1,
"Birth_Rate_Dependence": "DEMOGRAPHIC_DEP_RATE",
"Birth_Rate_Time_Dependence": "NONE",
"Demographics_Filenames": ["NO_DEFAULT_DEMOGRAPHICS"],
"Default_Geography_Initial_Node_Population": 1000,
"Default_Geography_Torus_Size": 10,
"Enable_Aging": 1,
"Enable_Birth": 1,
"Enable_Demographics_Birth": 0,
"Enable_Demographics_Gender": 1,
"Enable_Demographics_Builtin": 0,
"Enable_Demographics_Other": 0,
"Enable_Demographics_Reporting": 0,
"Enable_Vital_Dynamics": 1,
"Minimum_Adult_Age_Years": 15,
"IMMUNITY": {
"Acquisition_Blocking_Immunity_Decay_Rate": 0.1,
"Acquisition_Blocking_Immunity_Duration_Before_Decay": 60,
"Enable_Immune_Decay": 1,
"Enable_Immunity": 1,
"Immunity_Acquisition_Factor": 0,
"Immunity_Transmission_Factor": 0,
"Immunity_Initialization_Distribution_Type": "DISTRIBUTION_OFF",
"Susceptibility_Scale_Type": "CONSTANT_SUSCEPTIBILITY",
"Transmission_Blocking_Immunity_Decay_Rate": 0.1,
"Transmission_Blocking_Immunity_Duration_Before_Decay": 60
},
"MORTALITY": {
"Base_Mortality": 0,
"Enable_Disease_Mortality": 0,
"Death_Rate_Dependence": "NONDISEASE_MORTALITY_BY_AGE_AND_GENDER",
"Immunity_Mortality_Factor": 0,
"Mortality_Blocking_Immunity_Decay_Rate": 0.001,
"Mortality_Blocking_Immunity_Duration_Before_Decay": 60,
"Mortality_Time_Course": "DAILY_MORTALITY"
},
"Population_Density_C50": 30,
"Population_Scale_Type": "USE_INPUT_FILE",
"SAMPLING": {
"Base_Individual_Sample_Rate": 1,
"Individual_Sampling_Type": "TRACK_ALL",
"Max_Node_Population_Samples": 40,
"Sample_Rate_0_18mo": 1,
"Sample_Rate_10_14": 1,
"Sample_Rate_15_19": 1,
"Sample_Rate_18mo_4yr": 1,
"Sample_Rate_20_Plus": 1,
"Sample_Rate_5_9": 1,
"Sample_Rate_Birth": 2
}
},
"DISEASE": {
"Animal_Reservoir_Type": "NO_ZOONOSIS",
"Enable_Superinfection": 0,
"INCUBATION": {
"Base_Incubation_Period": 3,
"Incubation_Period_Distribution": "FIXED_DURATION"
},
"INFECTIOUSNESS": {
"Base_Infectious_Period": 7,
"Base_Infectivity": 0.3,
"Infectious_Period_Distribution": "EXPONENTIAL_DURATION",
"Infectivity_Scale_Type": "CONSTANT_INFECTIVITY",
"Population_Density_Infectivity_Correction": "CONSTANT_INFECTIVITY"
},
"Infection_Updates_Per_Timestep": 1,
"Max_Individual_Infections": 1,
"TRANSMISSION": {
"Enable_Maternal_Transmission": 0,
"Maternal_Transmission_Probability": 0
}
},
"FUDGE_FACTORS": {
"x_Air_Migration": 1,
"x_Birth": 1,
"x_Local_Migration": 1,
"x_Other_Mortality": 1,
"x_Population_Immunity": 1,
"x_Regional_Migration": 1,
"x_Sea_Migration": 1,
"x_Temporary_Larval_Habitat": 1
},
"HPC": {
"Job_Node_Groups": "Chassis08",
"Job_Priority": "BELOWNORMAL",
"Load_Balance_Filename": "",
"Local_Simulation": 0,
"Num_Cores": 1
},
"INTRANODE_TRANSMISSION": {
"Enable_Default_Shedding_Function": 1,
"Enable_Heterogeneous_Intranode_Transmission": 0
},
"MIGRATION": {
"Migration_Model": "NO_MIGRATION"
},
"OUTPUT": {
"Custom_Reports_Filename": "NoCustomReports",
"Report_Event_Recorder": 0,
"Enable_Default_Reporting": 1,
"Enable_Property_Output": 0,
"Enable_Spatial_Output": 0
},
"POLIO": {},
"PRIMARY": {
"Config_Name": "DEFAULT_CONFIG_NAME_SHOULD_BE_SET",
"ENUMS": {
"Simulation_Type": "GENERIC_SIM"
},
"Geography": "DEFAULT_GEOGRAPHY_SHOULD_BE_SET",
"Node_Grid_Size": 0.042,
"Run_Number": 0,
"Simulation_Duration": 365,
"Simulation_Timestep": 1,
"Start_Time": 0,
"Enable_Absolute_Time": "NO"
},
"SERIALIZATION": {
"Burnin_Cache_Mode": "none",
"Burnin_Cache_Period": 0,
"Burnin_Name": "",
"Serialization_Test_Cycles": 0
},
"STRAIN_TRACKING": {
"Number_Basestrains": 1,
"Number_Substrains": 1
}
}
}
Create or modify configuration files¶
Although you can create configuration files entirely from scratch, it is often easier to start from an existing configuration file and modify it to meet your needs. Any of the configuration files in the Regression directory may be used; in particular the Regression/defaults directory contains hierarchical configuration files with the common parameter settings used with different simulation types.
The simplest method is to use a text editor to directly edit the parameters or parameter values in the JSON file. However, while direct editing of files may be sufficient for small and infrequent changes, you will likely find that scripting tools are more powerful and reliable for both creating and modifying files.
The following example shows how to read a configuration file to a Python dictionary, modify a parameter, and write it back out to the file:
import json
# load the current config.json
config_file = open( "config.json" )
config_json = json.load( config_file )
config_file.close()
# modify one of the parameter values, e.g. "base_infectivity"
config_json["parameters"]["base_infectivity"] = 0.5
# write the modified config file
modified_file = open( "modified_config.json", "w" )
json.dump( config_json, modified_file, sort_keys=True, indent=4 )
modified_file.close()
The following example shows how to modify a configuration file in MATLAB:
addpath Matlab
addpath Matlab\test
% load the simulation configuration file into MATLAB structure
configjson = loadJson( "config.json" );
% modify one of the values
configjson.parameters.Base_Infectivity = 08.5;
% save the new configuration to file
saveJson( "modified_config.json", configjson );
Generate a list of all available parameters (a schema)¶
You can generate a schema from the EMOD executable (Eradication.exe) or Eradication binary that defines all configuration parameters and campaign parameters available in the version of EMOD that is installed, for all available simulation types. This includes parameter names, data types, defaults, ranges, and short descriptions.
Logging information is also produced as part of the schema. If you are using EMOD source and have added or modified configuration parameters or intervention code, this logging information can help you troubleshoot any errors that may occur.
Warning
If you modify the source code to add or remove configuration or campaign parameters, you may need to update the code used to produce the schema. You must also verify that your simulations are still scientifically valid.
The following command-line options are available for providing information about EMOD.
Long form |
Short form |
Description |
|---|---|---|
|
|
Shows help options for Eradication.exe. |
|
|
Shows the version information and supported simulation types. Note capitalization of short alternative. |
|
|
Outputs the configuration and campaign parameters. Unless |
|
|
When used with |
Open a Command Prompt window and navigate to the directory where Eradication.exe is installed.
To output the schema to the Command Prompt window, enter:
Eradication.exe --get-schema
To output the schema to a file, do one of the following (replacing <filename> with the desired filename):
To output a text file that includes logging information, enter:
Eradication.exe --get-schema > <filename>.txt
To display logging in the Command Prompt window and output a text file that does not include logging information, enter:
Eradication.exe --get-schema --schema-path <filename>.txt
To output the schema to a JSON file that includes logging information, enter:
Eradication.exe --get-schema > <filename>.json
To display logging in the Command Prompt window and output a JSON file that does not include logging information, enter:
Eradication.exe --get-schema --schema-path <filename>.json
Perform a parameter sweep¶
Parameter sweeps iteratively update the values of parameters to exhaustively search through the parameter space for a simulation. EMOD does not currently support automated parameter sweeps. However, you can write your own code, such as a Python or MATLAB script, that iterates through the values you want for a particular parameter. This topic describes how to perform a parameter sweep.
For example, you can run simulations using a Python script that uses the parameter values specified in a JSON parameter sweep file to iteratively update the configuration or campaign parameter values. The IDM test team performs parameter sweeps as part of regression testing. See the following examples to see how this is implemented.
The following JSON example illustrates sweeping through configuration parameter values.
{
"sweep" :
{
"path": "43_Vector_Garki_MultiCore_VectorMigration",
"param_name" : "Run_Number",
"param_values" : [ 1, 2, 3, 5, 8, 13, 21, 34, 55, 89, 144 ]
}
}
The following Python example illustrates how to update the configuration file with the above values. This is excerpted from the regression_test.py script in the Regression directory.
elif "sweep" in reglistjson:
print( "Running sweep...\n" )
param_name = reglistjson["sweep"]["param_name"]
for param_value in reglistjson["sweep"]["param_values"]:
os.chdir(ru.cache_cwd)
sim_timestamp = str(datetime.datetime.now()).replace('-', '_' ).replace( ' ', '_' ).replace( ':', '_' ).replace( '.', '_' )
if regression_id == None:
regression_id = sim_timestamp
configjson = ru.flattenConfig( os.path.join( reglistjson["sweep"]["path"], "param_overrides.json" ) )
if configjson is None:
print("Error flattening config. Skipping " + simcfg["path"])
ru.final_warnings += "Error flattening config. Skipped " + simcfg["path"] + "\n"
continue
# override sweep parameter
configjson["parameters"][param_name] = param_value
campjson_file = open( os.path.join( reglistjson["sweep"]["path"],"campaign.json" ) )
campjson = json.loads( campjson_file.read().replace( "u'", "'" ).replace( "'", '"' ).strip( '"' ) )
campjson_file.close()
configjson["campaign_json"] = str(campjson)
report_fn = os.path.join( reglistjson["sweep"]["path"],"custom_reports.json" )
if os.path.exists( report_fn ) == True:
reportjson_file = open( report_fn )
reportjson = json.loads( reportjson_file.read().replace( "u'", "'" ).replace( "'", '"' ).strip( '"' ) )
reportjson_file.close()
configjson["custom_reports_json"] = str(reportjson)
else:
configjson["custom_reports_json"] = None
thread = runner.commissionFromConfigJson( sim_timestamp, configjson, reglistjson["sweep"]["path"], None, 'sweep' )
ru.reg_threads.append( thread )
else:
print "Unknown state"
sys.exit(0)
Use configuration overlay files¶
You can use two configuration files when setting up a simulation. One file contains default parameter settings and an overlay file contains additional parameters or different parameter values that override the values in the default file. This topic describes how to set up the default and overlay file and flatten them into a single file before running a simulation.
Overlay files allow you to easily separate a subset of parameters that are of particular interest from the rest of the parameters needed to run a simulation. You can easily modify the parameters in the overlay file without needing to maintain a complete configuration or campaign file. This can be especially helpful when you want to experiment with the values set in certain parameters of interest without modifying the rest of the settings. You can have one default file and many different overlay files for different configurations. It also allows you to easily update the default values across multiple simulations.
In addition to being used for model experimentation, overlay files are used when testing the software functionality after making source code changes. If you run the EMOD regression tests using regression_test.py, configuration files will be flattened as part of those tests. However, this may take several hours if run locally. More guidance on modifying the EMOD source code is in the “Advance the model” section.
Follow the steps below to quickly flatten configuration files.
Create the default configuration file in JSON. You may, though it is not required, organize the parameters into logical categories of nested JSON objects to make managing the parameters easier. See Configuration parameters for a complete list of all parameters that are available. See the example default configuration file below.
{ "parameters": { "CAMPAIGNS": { "Campaign_Filename": "campaign.json", "Enable_Interventions": 1, "Listed_Events": [], "PKPD_Model": "FIXED_DURATION_CONSTANT_EFFECT" }, "CLIMATE": { "Climate_Model": "CLIMATE_OFF" }, "DEMOGRAPHICS": { "Age_Initialization_Distribution_Type": "DISTRIBUTION_SIMPLE", "Base_Population_Scale_Factor": 1, "Birth_Rate_Dependence": "DEMOGRAPHIC_DEP_RATE", "Birth_Rate_Time_Dependence": "NONE", "Demographics_Filenames": ["NO_DEFAULT_DEMOGRAPHICS"], "Default_Geography_Initial_Node_Population": 1000, "Default_Geography_Torus_Size": 10, "Enable_Aging": 1, "Enable_Birth": 1, "Enable_Demographics_Birth": 0, "Enable_Demographics_Gender": 1, "Enable_Demographics_Builtin": 0, "Enable_Demographics_Other": 0, "Enable_Demographics_Reporting": 0, "Enable_Vital_Dynamics": 1, "Minimum_Adult_Age_Years": 15, "IMMUNITY": { "Acquisition_Blocking_Immunity_Decay_Rate": 0.1, "Acquisition_Blocking_Immunity_Duration_Before_Decay": 60, "Enable_Immune_Decay": 1, "Enable_Immunity": 1, "Immunity_Acquisition_Factor": 0, "Immunity_Transmission_Factor": 0, "Immunity_Initialization_Distribution_Type": "DISTRIBUTION_OFF", "Susceptibility_Scale_Type": "CONSTANT_SUSCEPTIBILITY", "Transmission_Blocking_Immunity_Decay_Rate": 0.1, "Transmission_Blocking_Immunity_Duration_Before_Decay": 60 }, "MORTALITY": { "Base_Mortality": 0, "Enable_Disease_Mortality": 0, "Death_Rate_Dependence": "NONDISEASE_MORTALITY_BY_AGE_AND_GENDER", "Immunity_Mortality_Factor": 0, "Mortality_Blocking_Immunity_Decay_Rate": 0.001, "Mortality_Blocking_Immunity_Duration_Before_Decay": 60, "Mortality_Time_Course": "DAILY_MORTALITY" }, "Population_Density_C50": 30, "Population_Scale_Type": "USE_INPUT_FILE", "SAMPLING": { "Base_Individual_Sample_Rate": 1, "Individual_Sampling_Type": "TRACK_ALL", "Max_Node_Population_Samples": 40, "Sample_Rate_0_18mo": 1, "Sample_Rate_10_14": 1, "Sample_Rate_15_19": 1, "Sample_Rate_18mo_4yr": 1, "Sample_Rate_20_Plus": 1, "Sample_Rate_5_9": 1, "Sample_Rate_Birth": 2 } }, "DISEASE": { "Animal_Reservoir_Type": "NO_ZOONOSIS", "Enable_Superinfection": 0, "INCUBATION": { "Base_Incubation_Period": 3, "Incubation_Period_Distribution": "FIXED_DURATION" }, "INFECTIOUSNESS": { "Base_Infectious_Period": 7, "Base_Infectivity": 0.3, "Infectious_Period_Distribution": "EXPONENTIAL_DURATION", "Infectivity_Scale_Type": "CONSTANT_INFECTIVITY", "Population_Density_Infectivity_Correction": "CONSTANT_INFECTIVITY" }, "Infection_Updates_Per_Timestep": 1, "Max_Individual_Infections": 1, "TRANSMISSION": { "Enable_Maternal_Transmission": 0, "Maternal_Transmission_Probability": 0 } }, "FUDGE_FACTORS": { "x_Air_Migration": 1, "x_Birth": 1, "x_Local_Migration": 1, "x_Other_Mortality": 1, "x_Population_Immunity": 1, "x_Regional_Migration": 1, "x_Sea_Migration": 1, "x_Temporary_Larval_Habitat": 1 }, "HPC": { "Job_Node_Groups": "Chassis08", "Job_Priority": "BELOWNORMAL", "Load_Balance_Filename": "", "Local_Simulation": 0, "Num_Cores": 1 }, "INTRANODE_TRANSMISSION": { "Enable_Default_Shedding_Function": 1, "Enable_Heterogeneous_Intranode_Transmission": 0 }, "MIGRATION": { "Migration_Model": "NO_MIGRATION" }, "OUTPUT": { "Custom_Reports_Filename": "NoCustomReports", "Report_Event_Recorder": 0, "Enable_Default_Reporting": 1, "Enable_Property_Output": 0, "Enable_Spatial_Output": 0 }, "POLIO": {}, "PRIMARY": { "Config_Name": "DEFAULT_CONFIG_NAME_SHOULD_BE_SET", "ENUMS": { "Simulation_Type": "GENERIC_SIM" }, "Geography": "DEFAULT_GEOGRAPHY_SHOULD_BE_SET", "Node_Grid_Size": 0.042, "Run_Number": 0, "Simulation_Duration": 365, "Simulation_Timestep": 1, "Start_Time": 0, "Enable_Absolute_Time": "NO" }, "SERIALIZATION": { "Burnin_Cache_Mode": "none", "Burnin_Cache_Period": 0, "Burnin_Name": "", "Serialization_Test_Cycles": 0 }, "STRAIN_TRACKING": { "Number_Basestrains": 1, "Number_Substrains": 1 } } }
Create the overlay configuration file in JSON. This file must include the parameter Default_Config_Path, set to the path to the default configuration file, relative to the location of the flatten_config.py script in the EMOD Regression folder. Again, you may organize the parameters into logical categories if you desire. See the example overlay configuration file below.
{ "Default_Config_Path": "defaults/generic-default-config.json", "parameters": { "DEMOGRAPHICS": { "Enable_Demographics_Builtin": 0, "Birth_Rate_Dependence": "POPULATION_DEP_RATE", "Death_Rate_Dependence" : "NONDISEASE_MORTALITY_OFF", "Enable_Vital_Dynamics": 0, "Sample_Rate_Birth": 1, "Enable_Demographics_Reporting": 1 }, "DISEASE": { "Base_Incubation_Period": 0, "Base_Infectious_Period": 4, "Base_Infectivity": 3.5, "Enable_Immune_Decay": 0 }, "PRIMARY": { "Config_Name": "00_Generic_DEFAULT", "Demographics_Filenames": ["demographics.json"], "Geography": "", "Run_Number": 1, "Simulation_Duration": 90 } } }
In a Command Prompt window, navigate to the Regression folder.
Run the flatten_config.py script, providing the relative path to the overlay file:
python flatten_config.py experiment/param_overlay.json
Open the resulting config.json file in the same folder as param_overlay.json and see that it has been flattened into a single layer with all parameters listed alphabetically and any logical categories removed. Eradication.exe will not accept a configuration file with nested JSON objects.
{ "parameters": { "Acquisition_Blocking_Immunity_Decay_Rate": 0.1, "Acquisition_Blocking_Immunity_Duration_Before_Decay": 60, "Age_Initialization_Distribution_Type": "DISTRIBUTION_SIMPLE", "Animal_Reservoir_Type": "NO_ZOONOSIS", "Base_Incubation_Period": 0, "Base_Individual_Sample_Rate": 1, "Base_Infectious_Period": 4, "Base_Infectivity": 3.5, "Base_Mortality": 0, "Base_Population_Scale_Factor": 1, "Birth_Rate_Dependence": "POPULATION_DEP_RATE", "Birth_Rate_Time_Dependence": "NONE", "Burnin_Cache_Mode": "none", "Burnin_Cache_Period": 0, "Burnin_Name": "", "Campaign_Filename": "campaign.json", "Climate_Model": "CLIMATE_OFF", "Config_Name": "00_DEFAULT", "Custom_Reports_Filename": "NoCustomReports", "Death_Rate_Dependence": "NONDISEASE_MORTALITY_OFF", "Default_Geography_Initial_Node_Population": 1000, "Default_Geography_Torus_Size": 10, "Demographics_Filenames": [ "../00_Default/demographics.json" ], "Enable_Absolute_Time": "NO", "Enable_Aging": 1, "Enable_Birth": 1, "Enable_Default_Reporting": 1, "Enable_Default_Shedding_Function": 1, "Enable_Demographics_Birth": 0, "Enable_Demographics_Builtin": 0, "Enable_Demographics_Gender": 1, "Enable_Demographics_Other": 0, "Enable_Demographics_Reporting": 1, "Enable_Disease_Mortality": 0, "Enable_Heterogeneous_Intranode_Transmission": 0, "Enable_Immune_Decay": 0, "Enable_Immunity": 1, "Enable_Interventions": 1, "Enable_Maternal_Transmission": 0, "Enable_Property_Output": 0, "Enable_Spatial_Output": 0, "Enable_Superinfection": 0, "Enable_Vital_Dynamics": 0, "Geography": "", "Immunity_Acquisition_Factor": 0, "Immunity_Initialization_Distribution_Type": "DISTRIBUTION_OFF", "Immunity_Mortality_Factor": 0, "Immunity_Transmission_Factor": 0, "Incubation_Period_Distribution": "FIXED_DURATION", "Individual_Sampling_Type": "TRACK_ALL", "Infection_Updates_Per_Timestep": 1, "Infectious_Period_Distribution": "EXPONENTIAL_DURATION", "Infectivity_Scale_Type": "CONSTANT_INFECTIVITY", "Job_Node_Groups": "Chassis08", "Job_Priority": "BELOWNORMAL", "Listed_Events": [], "Load_Balance_Filename": "", "Local_Simulation": 0, "Maternal_Transmission_Probability": 0, "Max_Individual_Infections": 1, "Max_Node_Population_Samples": 40, "Migration_Model": "NO_MIGRATION", "Minimum_Adult_Age_Years": 15, "Mortality_Blocking_Immunity_Decay_Rate": 0.001, "Mortality_Blocking_Immunity_Duration_Before_Decay": 60, "Mortality_Time_Course": "DAILY_MORTALITY", "Node_Grid_Size": 0.042, "Num_Cores": 1, "Number_Basestrains": 1, "Number_Substrains": 1, "PKPD_Model": "FIXED_DURATION_CONSTANT_EFFECT", "Population_Density_C50": 30, "Population_Density_Infectivity_Correction": "CONSTANT_INFECTIVITY", "Population_Scale_Type": "USE_INPUT_FILE", "Report_Event_Recorder": 0, "Run_Number": 1, "Sample_Rate_0_18mo": 1, "Sample_Rate_10_14": 1, "Sample_Rate_15_19": 1, "Sample_Rate_18mo_4yr": 1, "Sample_Rate_20_Plus": 1, "Sample_Rate_5_9": 1, "Sample_Rate_Birth": 1, "Serialization_Test_Cycles": 0, "Simulation_Duration": 90, "Simulation_Timestep": 1, "Simulation_Type": "GENERIC_SIM", "Start_Time": 0, "Susceptibility_Scale_Type": "CONSTANT_SUSCEPTIBILITY", "Transmission_Blocking_Immunity_Decay_Rate": 0.1, "Transmission_Blocking_Immunity_Duration_Before_Decay": 60, "x_Air_Migration": 1, "x_Birth": 1, "x_Local_Migration": 1, "x_Other_Mortality": 1, "x_Population_Immunity": 1, "x_Regional_Migration": 1, "x_Sea_Migration": 1, "x_Temporary_Larval_Habitat": 1 } }
Note
These same steps can also be used to flatted a single configuration file that has been hierarchically organized into logical categories.
Malaria or vector configuration tips¶
The following tips are useful for the initial configuration of the vector or malaria model. Configuration tips for a specific microsolver in the model will be found on their respective component pages. For more information on using the software, or configuring input files, see How to use the software.
There are several options for running a vector-based model. To run a basic model that functions as an enhanced SEIRS model, you will need to set the parameter Simulation_Type to the value “VECTOR_SIM”. However, to add complexity and include the full system of microsolvers for within- host parasite dynamics and other immune system interactions, set the parameter Simulation_Type to the value “MALARIA_SIM”.
To select between the cohort model versus individual mosquito model, set the Vector_Sampling_Type parameter. For the cohort model, the possible values are “VECTOR_COMPARTMENTS_NUMBER” or “VECTOR_COMPARTMENTS_PERCENT”. For the individual mosquito model, the values are “TRACK_ALL_VECTORS” or “SAMPLE_IND_VECTORS”.
The simulation configuration file has a field Vector_Species_Names which specifies the vector species that will be present in the simulation. For example, the following specifies that Anopheles arabiensis, An. funestus, and An. gambiae will be added to the simulation:
{
"Vector_Species_Names": [
"arabiensis",
"funestus",
"gambiae"
]
}
For each species listed, a “VectorPopulation” object will be added to the simulation at each node. Each species will be defined by parameters in the simulation configuration file for the vector ecology and behavior of the species. This allows for a mechanistic description of vector abundances and behavior through the effects of climate and weather on different preferred larval habitats.
The following example shows the syntax for configuring species parameters for Anopheles arabiensis. Note that you would also need to configure parameters for An. fuenstus and An. gambiae, if you are using a three species model as from the above example.
{
"Vector_Species_Params": {
"arabiensis": {
"Acquire_Modifier": 0.2,
"Adult_Life_Expectancy": 10,
"Anthropophily": 0.95,
"Aquatic_Arrhenius_1": 84200000000,
"Aquatic_Arrhenius_2": 8328,
"Aquatic_Mortality_Rate": 0.1,
"Days_Between_Feeds": 3,
"Egg_Batch_Size": 100,
"Immature_Duration": 2,
"Indoor_Feeding_Fraction": 1,
"Infected_Arrhenius_1": 117000000000,
"Infected_Arrhenius_2": 8336,
"Infected_Egg_Batch_Factor": 0.8,
"Infectious_Human_Feed_Mortality_Factor": 1.5,
"Larval_Habitat_Types": {
"TEMPORARY_RAINFALL": 11250000000
},
"Transmission_Rate": 0.5
}
}
}
While the above example shows the mosquito residing in one habitat, it is worth noting that mosquitoes can reside in multiple habitat types. This linear-combination of habitat types allows for greater control over local vector ecology (especially during dry seasons as habitat availability may be drastically different). The following example demonstrates the syntax to include several habitat types.
{
"Vector_Species_Params": {
"arabiensis": {
"Acquire_Modifier": 0.2,
"Adult_Life_Expectancy": 10,
"Anthropophily": 0.95,
"Aquatic_Arrhenius_1": 84200000000,
"Aquatic_Arrhenius_2": 8328,
"Aquatic_Mortality_Rate": 0.1,
"Days_Between_Feeds": 3,
"Egg_Batch_Size": 100,
"Immature_Duration": 2,
"Indoor_Feeding_Fraction": 1,
"Infected_Arrhenius_1": 117000000000,
"Infected_Arrhenius_2": 8336,
"Infected_Egg_Batch_Factor": 0.8,
"Infectious_Human_Feed_Mortality_Factor": 1.5,
"Larval_Habitat_Types": {
"TEMPORARY_RAINFALL": 11250000000,
"WATER_VEGETATION": 6000000000
},
"Transmission_Rate": 0.5
}
}
}
There are example configuration and campaign files available to run the vector and malaria models. To download these files, go to the Vector and Malaria Scenarios. You can download the executable file to run these configuration and campaign files here: Quickstart.
Descriptions of the scenarios these files simulate can be found in the Vector and malaria tutorials.
Create a campaign file¶
You define the initial disease outbreak and interventions used a campaign for a simulation through a JSON-formatted campaign file, typically called campaign.json. It is hierarchically organized into logical groups of parameters that can have arbitrary levels of nesting. For some interventions, there can be a very complex hierarchical structure, including recursion. The campaign file must be in the same directory as the configuration file and the name must be specified by the configuration parameter Campaign_Filename. This topic describes how to create the campaign file.
The EMOD Regression directory contains many different subdirectories that contain configuration, campaign, and other associated files to run simulations that represent real-world scenarios. Within the each subdirectory, there is usually a single campaign file (campaign.json), though some directories also include a campaign overlay file (campaign_overrides.json), which has been created by combining campaign_overrides.json with one of the default files in Regression/defaults. The naming of these files is an arbitrary convention used at IDM; you may name this files anything you choose. See Use campaign overlay files for more information flattening two campaign files.
For a complete list of campaign parameters that are available to use with this simulation type, see Campaign parameters. For more information about JSON, see EMOD parameter reference.
To distribute an intervention, you must configure the following:
- Campaign event
A JSON object that determines when and where an intervention is distributed during a campaign.
- Event coordinator
A JSON object that determines who will receive a particular intervention during a campaign.
- Intervention
A JSON object that determines what will be distributed to reduce the spread of a disease. An intervention can be distributed either to an individual (such as a vaccine, drug, or bednet) or to a node (such as a larvicide). Sometimes this can be an intermediate intervention that schedules another intervention.
The following is an example of campaign file that has two events (SimpleVaccine and Outbreak) that occur in all nodes at day 1 and day 30, respectively. Each event contains an event coordinator that describes who receives the intervention (everyone, with the vaccine repeated three times) and the configuration for the intervention itself.
Warning
The Outbreak must be the last event in the campaign file or none of the interventions will take place.
{
"Campaign_Name": "Vaccine",
"Use_Defaults": 1,
"Events":
[
{
"Event_Name": "SimpleVaccine",
"Event_Coordinator_Config": {
"Demographic_Coverage": 0.5,
"Intervention_Config": {
"Cost_To_Consumer": 10,
"Waning_Config": {
"class": "WaningEffectMapLinear",
"Initial_Effect" : 1.0,
"Expire_At_Durability_Map_End" : 0,
"Durability_Map" : {
"Times" : [ 0, 30, 60, 90, 120 ],
"Values" : [ 0.9, 0.3, 0.9, 0.6, 1.0 ]
}
},
"Vaccine_Take": 1,
"Vaccine_Type": "AcquisitionBlocking",
"class": "SimpleVaccine"
},
"Number_Repetitions": 3,
"Target_Demographic": "Everyone",
"Timesteps_Between_Repetitions": 7,
"class": "StandardInterventionDistributionEventCoordinator"
},
"Nodeset_Config": {
"class": "NodeSetAll"
},
"Start_Day": 1,
"class": "CampaignEvent",
},
{
"Event_Name": "Outbreak",
"Event_Coordinator_Config": {
"Demographic_Coverage": 0.001,
"Intervention_Config": {
"Antigen": 0,
"Genome": 0,
"Outbreak_Source": "PrevalenceIncrease",
"class": "OutbreakIndividual"
},
"Target_Demographic": "Everyone",
"class": "StandardInterventionDistributionEventCoordinator"
},
"Nodeset_Config": {
"class": "NodeSetAll"
},
"Start_Day": 30,
"class": "CampaignEvent"
}
]
}
Multiple interventions¶
When creating multiple interventions, either of the same type or different types, they will generally be distributed independently without regard to whether a person has already received another intervention. For example, say you create two SimpleBednet interventions and both interventions have Demographic_Coverage set to 0.5 (50% demographic coverage). This value is the probability that each individual in the target population will receive the intervention. It does not guarantee that the exact fraction of the target population set by Demographic_Coverage receives the intervention.
By default, each individual in the simulation will have a 50% chance of receiving a bednet in both of the distributions and the two distributions will be independent. Therefore, each individual has a 75% chance of receiving at least one bednet.
Create a new campaign event¶
Although you can create campaign files entirely from scratch, it is often easier to start from an existing campaign file and modify it to meet your needs. Any of the campaign files in the Regression directory may be used. The simplest method is to edit the parameters or parameter values in the JSON file in any text editor. You may also create or modify the files using a scripting language, as with configuration files. See Create a configuration file for examples.
In the Events array, create a new empty JSON object. This object will contain all parameters for the new campaign event.
In this object, add and set values for the Event_Name and class parameters.
Add the Nodeset_Config parameter to define in which geographic nodes this even will occur.
Add the Event_Coordinator_Config parameter and assign a JSON object that will contain multiple parameters that control who receives the intervention.
In this object, add the Intervention_Config parameter and assign a JSON object that will contain multiple parameters that control which intervention is distributed.
Use campaign overlay files¶
You can use two campaign files when setting up a simulation. One file contains default campaign settings and an overlay file contains additional parameters or different parameter values that override the values in the default file. This topic describes how to set up the default and overlay file and flatten them into a single file before running a simulation.
Overlay files allow you to easily separate a subset of parameters that are of particular interest from the rest of the parameters needed to run a simulation. You can easily modify the parameters in the overlay file without needing to maintain a complete campaign file. This can be especially helpful when you want to experiment with different interventions without modifying the rest of the settings. You can have one default file and many different overlay files for different intervention settings. It also allows you to easily update the default values across multiple simulations.
In addition to being used for model experimentation, overlay files are used when testing the software functionality after making source code changes. If you run the EMOD regression tests using regression_test.py, campaign files will be flattened as part of those tests. However, this may take several hours if run locally. More guidance on modifying the EMOD source code is in the “Advance the model” section.
Follow the steps below to quickly flatten campaign files.
Create the default campaign file in JSON. You may, though it is not required, organize the parameters into logical categories of nested JSON objects to make managing the parameters easier. See Campaign parameters for a complete list of all parameters that are available. See the example default campaign file below.
{ "Campaign_Name": "Initial Seeding", "Events": [ { "Event_Coordinator_Config": { "Demographic_Coverage": 0.001, "Intervention_Config": { "Antigen": 0, "Genome": 0, "Outbreak_Source": "PrevalenceIncrease", "class": "OutbreakIndividual" }, "Target_Demographic": "Everyone", "class": "StandardInterventionDistributionEventCoordinator" }, "Event_Name": "Outbreak", "Nodeset_Config": { "class": "NodeSetAll" }, "Start_Day": 30, "class": "CampaignEvent" } ], "Use_Defaults": 1 }
Create the overlay campaign file in JSON. This file must include the parameter Default_Campaign_Path, set to the path to the default campaign file, relative to the location of the flatten_campaign.py script in the EMOD Regression folder. Again, you may organize these into logical categories if you desire. See the example overlay campaign file below.
{ "Campaign_Name": "Vaccine", "Default_Campaign_Path": "defaults/generic-default-campaign.json", "Events": [ { "VACCINATION": "BEGIN", "Event_Name": "SimpleVaccine", "Event_Coordinator_Config": { "Demographic_Coverage": 0.5, "Intervention_Config": { "Cost_To_Consumer": 10, "Waning_Config": { "class": "WaningEffectMapLinear", "Initial_Effect" : 1.0, "Expire_At_Durability_Map_End" : 0, "Durability_Map" : { "Times" : [ 0, 30, 60, 90, 120 ], "Values" : [ 0.9, 0.3, 0.9, 0.6, 1.0 ] } }, "Vaccine_Take": 1, "Vaccine_Type": "AcquisitionBlocking", "class": "SimpleVaccine" }, "Number_Repetitions": 3, "Target_Demographic": "Everyone", "Timesteps_Between_Repetitions": 7, "class": "StandardInterventionDistributionEventCoordinator" }, "Nodeset_Config": { "class": "NodeSetAll" }, "Start_Day": 1, "class": "CampaignEvent", "VACCINATION": "END" } ] }
In a Command Prompt window, navigate to the Regression folder.
Run the flatten_campaign.py script, providing the relative path to the overlay file and the path to and name of the new flattened campaign file that will be saved, using the arguments as shown below:
python flatten_campaign.py --overlay experiment/campaign_overlay.json --saveto experiment/campaign.json
Open the resulting campaign.json file and see that it has been flattened into a single file with nested JSON objects and any logical categories retained.
{ "Campaign_Name": "Vaccine", "Default_Campaign_Path": "defaults/generic-default-campaign.json", "Use_Defaults": 1, "Events": [ { "VACCINATION": "BEGIN", "Event_Name": "SimpleVaccine", "Event_Coordinator_Config": { "Demographic_Coverage": 0.5, "Intervention_Config": { "Cost_To_Consumer": 10, "Waning_Config": { "class": "WaningEffectMapLinear", "Initial_Effect" : 1.0, "Expire_At_Durability_Map_End" : 0, "Durability_Map" : { "Times" : [ 0, 30, 60, 90, 120 ], "Values" : [ 0.9, 0.3, 0.9, 0.6, 1.0 ] } }, "Vaccine_Take": 1, "Vaccine_Type": "AcquisitionBlocking", "class": "SimpleVaccine" }, "Number_Repetitions": 3, "Target_Demographic": "Everyone", "Timesteps_Between_Repetitions": 7, "class": "StandardInterventionDistributionEventCoordinator" }, "Nodeset_Config": { "class": "NodeSetAll" }, "Start_Day": 1, "class": "CampaignEvent", "VACCINATION": "END" }, { "Event_Name": "Outbreak", "Event_Coordinator_Config": { "Demographic_Coverage": 0.001, "Intervention_Config": { "Antigen": 0, "Genome": 0, "Outbreak_Source": "PrevalenceIncrease", "class": "OutbreakIndividual" }, "Target_Demographic": "Everyone", "class": "StandardInterventionDistributionEventCoordinator" }, "Nodeset_Config": { "class": "NodeSetAll" }, "Start_Day": 30, "class": "CampaignEvent" } ] }
After the overlay files and default files are combined into a single campaign file, you can run a simulation using the EMOD executable (Eradication.exe).
Target interventions to nodes or groups¶
Generally, you want to target your outbreaks and campaign interventions to specific regions or individuals who meet certain criteria. For example, you may want to distribute bednets only to areas where a mosquito-borne disease is endemic or vaccinate only young children who are at highest risk. This topic describes how to distribute interventions to specific geographic nodes or groups of individuals.
Targeting geographic nodes with a particular intervention is controlled by the Nodeset_Config parameter in the campaign file. To distribute the intervention to all nodes, simply set it to “NodeSetAll”. To distribute to a subset of nodes, follow the steps below.
In the JSON object for the campaign event, set Nodeset_Config using one of the following two options:
Set it to an empty JSON object. Within that object, set the following:
Set class to “NodeSetNodeList”.
Set Node_List to an array that contains a comma-delimited list of nodes that set where the intervention will be distributed.
Set it to an empty JSON object. Within that object, set the following:
Set class to “NodeSetPolygon”.
Set Polygon_Format to “SHAPE”.
Set Vertices to a comma-delimited list of latitude and longitude pairs that define the outer boundary of the region you want to target. The intervention will be distributed to all nodes within the defined bounds.
See the example below.
{
"Use_Defaults": 1,
"Events": [{
"Event_Name": "Outbreak",
"Nodeset_Config": {
"class": "NodeSetNodeList",
"Node_List": [1, 3, 5]
},
"Start_Day": 10,
"class": "CampaignEvent"
}]
}
To target interventions to particular groups in a population, you must first create those groups in the demographics file using IndividualProperties. Then, in the campaign file, you can target an intervention or outbreak to a group of individuals based on the properties applied to them. See Configure heterogeneous populations for instructions on creating the groups.
Just as creating groups based on age works a little differently than groups based on other properties, targeting an intervention to a particular age range works a little differently than targeting an intervention to other properties.
To target interventions to properties other than age:
In the campaign file, in the event you want to target, set Target_Demographic to “Everyone”.
Add the Property_Restrictions parameter and set to an empty array.
In that array, add a list of JSON key-value pairs of the property type and value that specifies the groups to apply the intervention to. If the name of the element is not valid, EMOD will ignore the property restriction.
To target interventions to age ranges:
In the campaign file, in the event you want to target, set Target_Demographic to “ExplicitAgeRanges.”
Add the Target_Age_Min and set to the lower age bound.
Add the Target_Age_Max and set to the upper age bound.
Both of these values must match one the values listed in Age_Bin_Edges_In_Years in the demographics file. EMOD does not verify this range.
The following examples illustrate how to target interventions to different groups. This includes how to configure interventions when there are multiple relevant properties, such as targeting individuals who are both low-risk and in a suburban setting or individuals who are either low- risk or living in suburban settings.
The following examples show how to target interventions based on a single property value.
At day 0 of the simulation, an outbreak starts. Target_Demographic is set to “Everyone” but Property_Restrictions restricts the start of the outbreak to the “Urban” group.
{
"Events": [{
"Event_Coordinator_Config": {
"Number_Distributions": -1,
"Intervention_Config": {
"Antigen": 0,
"Genome": 0,
"Outbreak_Source": "PrevalenceIncrease",
"class": "OutbreakIndividual"
},
"Demographic_Coverage": 1,
"Target_Demographic": "Everyone",
"Property_Restrictions": [
"Place:Urban"
],
"class": "StandardInterventionDistributionEventCoordinator"
},
"Nodeset_Config": {
"class": "NodeSetAll"
},
"Start_Day": 0,
"class": "CampaignEvent"
}]
}
Even if you have multiple properties defined in the demographics file, you can target interventions to a single property value in the same way. Individuals can have membership in any of the groups from the other property type.
In this example, groups are defined for both the “Risk” and “Place” property types. The outbreak only targets the “Suburban” group using the “Place” property type. Individuals can have membership in either of the “Risk” groups.
{
"Events": [{
"Event_Coordinator_Config": {
"Number_Distributions": -1,
"Intervention_Config": {
"Antigen": 0,
"Genome": 0,
"Outbreak_Source": "PrevalenceIncrease",
"class": "OutbreakIndividual"
},
"Demographic_Coverage": 1,
"Target_Demographic": "Everyone",
"Property_Restrictions": [
"Place:Urban"
],
"class": "StandardInterventionDistributionEventCoordinator"
},
"Nodeset_Config": {
"class": "NodeSetAll"
},
"Start_Day": 0,
"class": "CampaignEvent"
}]
}
If you want to target multiple groups with the same property type, such as both “Urban” and “Rural” values with the “Place” property, you must use multiple interventions. You cannot have more than one group with the same property value in one intervention or outbreak.
In this example, an outbreak starts at day 0 in the both the “Rural” and “Urban” groups.
{
"Events": [{
"Event_Coordinator_Config": {
"Number_Distributions": -1,
"Intervention_Config": {
"Antigen": 0,
"Genome": 0,
"Outbreak_Source": "PrevalenceIncrease",
"class": "OutbreakIndividual"
},
"Demographic_Coverage": 1,
"Target_Demographic": "Everyone",
"Property_Restrictions": [
"Place:Rural"
],
"class": "StandardInterventionDistributionEventCoordinator"
},
"Nodeset_Config": {
"class": "NodeSetAll"
},
"Start_Day": 0,
"class": "CampaignEvent"
}, {
"Event_Coordinator_Config": {
"Number_Distributions": -1,
"Intervention_Config": {
"Antigen": 0,
"Genome": 0,
"Outbreak_Source": "PrevalenceIncrease",
"class": "OutbreakIndividual"
},
"Demographic_Coverage": 1,
"Target_Demographic": "Everyone",
"Property_Restrictions": [
"Place:Urban"
],
"class": "StandardInterventionDistributionEventCoordinator"
},
"Nodeset_Config": {
"class": "NodeSetAll"
},
"Start_Day": 0,
"class": "CampaignEvent"
}]
}
To target individuals who belong to multiple groups defined by different property types, such as people who are both urban and low risk, you can use a single intervention or outbreak. When two groups from multiple properties are targeted in one intervention or outbreak, the event is only applied to individuals that belong to both of the groups.
In this example, a vaccine intervention is targeted at the “Low” risk group and the “Suburban” place group. Individuals that are targeted to receive the vaccine must be in both the “Suburban” group and the “Low” risk group.
{
"Events": [{
"Event_Coordinator_Config": {
"Demographic_Coverage": 1.0,
"Intervention_Config": {
"Cost_To_Consumer": 10,
"Durability_Time_Profile": "BOXDECAYDURABILITY",
"Primary_Decay_Time_Constant": 3650,
"Reduced_Acquire": 1,
"Reduced_Transmit": 0,
"Secondary_Decay_Time_Constant": 3650,
"Vaccine_Take": 1,
"Vaccine_Type": "StrainSpecific",
"class": "SimpleImmunoglobulin"
},
"Target_Demographic": "Everyone",
"Property_Restrictions": [
"Risk:Low",
"Place:Suburban"
],
"class": "StandardInterventionDistributionEventCoordinator"
},
"Nodeset_Config": {
"class": "NodeSetAll"
},
"Start_Day": 100,
"class": "CampaignEvent"
}]
}
However, if you want to target multiple groups, but individuals need to be a member of only one of the specified groups to qualify for the intervention, you must create an intervention for each of the targeted groups. The events are applied separately and are not restricted to individuals that belong to both of the groups. Instead, individuals can belong to either of the groups.
In this example, an outbreak is targeted at the “Low” risk group in the first intervention and is targeted at the “Suburban” group in the second intervention.
{
"Events": [{
"Event_Coordinator_Config": {
"Number_Distributions": -1,
"Intervention_Config": {
"Antigen": 0,
"Genome": 0,
"Outbreak_Source": "PrevalenceIncrease",
"class": "OutbreakIndividual"
},
"Demographic_Coverage": 1,
"Target_Demographic": "Everyone",
"Property_Restrictions": [
"Risk:Low"
],
"class": "StandardInterventionDistributionEventCoordinator"
},
"Nodeset_Config": {
"class": "NodeSetAll"
},
"Start_Day": 0,
"class": "CampaignEvent"
}, {
"Event_Coordinator_Config": {
"Number_Distributions": -1,
"Intervention_Config": {
"Antigen": 0,
"Genome": 0,
"Outbreak_Source": "PrevalenceIncrease",
"class": "OutbreakIndividual"
},
"Demographic_Coverage": 1,
"Target_Demographic": "Everyone",
"Property_Restrictions": [
"Place:Suburban"
],
"class": "StandardInterventionDistributionEventCoordinator"
},
"Nodeset_Config": {
"class": "NodeSetAll"
},
"Start_Day": 0,
"class": "CampaignEvent"
}]
}
Targeting an intervention to an age range is set up differently than targeting an intervention to other property types. However, you can combine both kinds of restrictions. In this example, a vaccine campaign is targeted at the “Urban” group for individuals who are age 0 to 5.
{
"Events": [{
"Event_Coordinator_Config": {
"Demographic_Coverage": 1.0,
"Intervention_Config": {
"Cost_To_Consumer": 10,
"Durability_Time_Profile": "BOXDECAYDURABILITY",
"Primary_Decay_Time_Constant": 3650,
"Reduced_Acquire": 1,
"Reduced_Transmit": 0,
"Secondary_Decay_Time_Constant": 3650,
"Vaccine_Take": 1,
"Vaccine_Type": "StrainSpecific",
"class": "SimpleImmunoglobulin"
},
"Target_Demographic": "ExplicitAgeRanges",
"Target_Age_Min": 0,
"Target_Age_Max": 5,
"Property_Restrictions": [
"Place:Urban"
],
"class": "StandardInterventionDistributionEventCoordinator"
},
"Nodeset_Config": {
"class": "NodeSetAll"
},
"Start_Day": 30,
"class": "CampaignEvent"
}]
}
Run simulations¶
There are a number of ways to run an EMOD simulation, whether locally or on a remote HPC cluster. The simplest way is to run a simulation at the command line, however that limits you to running one local simulation at a time. Because the EMOD model is stochastic, you must run many simulations before you can interpret the outcome. You may want to use a scripting language like Python or MATLAB or an application like mpiexec to run multiple simulations at once.
In addition, if you are modifying the EMOD source code to add functionality to the model, you can run simulations in Visual Studio as part of debugging. This process is described in Run debug simulations in Visual Studio.
No matter how you choose to run simulations, you must have a built copy of the EMOD executable (Eradication.exe) or Eradication binary, either downloaded directly from GitHub or built from the EMOD source code. See EMOD installation. In addition, you must know the paths to the configuration file and input data files and where you want to store the output files. You will pass this information as arguments to Eradication.exe.
Directory structure¶
Although there are many ways you can structure the files needed to run a simulation, we recommend the following to keep your files organized and simplify the file paths set in the configuration file or passed as arguments to Eradication.exe.
Place the configuration and campaign files needed for a simulation in the same directory. This is also known as the working directory.
However, if you are using overlay files, you may want the default configuration or campaign file in a separate directory so they can be used with different overlay files for other simulations.
Place all input data files for a given region in the same directory.
Place output for a simulation in a subdirectory of the directory containing configuration and campaign files.
It is not important where you install Eradication.exe or the Eradication binary.
Run a simulation using the command line¶
Using command-line options is the simplest way to run an EMOD simulation. This topic describes the commands available for running simulations.
The EMOD executable (Eradication.exe) and Eradication binary also provide commands that provide information about the version of EMOD that is installed, such as available parameters and simulation types. For more information, see Generate a list of all available parameters (a schema). The examples below show the Windows Eradication.exe, but the options are the same for the Eradication binary on CentOS on Azure.
The following command-line options are available for running EMOD simulations. Paths can be absolute or relative to the directory you are in when running the commands, unless otherwise specified.
Long form |
Short form |
Description |
|---|---|---|
|
|
Path to the configuration file. If not specified, EMOD will look for a file named default-config.json in the current directory. |
|
|
Path to the directory containing input data files. If not specified, EMOD will look for files in the current directory. |
|
|
Path to the directory where output files will be written. If not specified, EMOD will create an “output” directory and overwrite any previous output in that directory. |
|
|
Path to the EMODule root directory. For more information, see Custom reporters. |
The following options are for monitoring the progress of simulations running on an high-performance computing (HPC) cluster. They are optional for any simulation, but they must be used together. To monitor progress, listen for User Datagram Protocol (UDP) messages on the specified host and port.
Long form |
Description |
|---|---|
|
The IP address of the commissioning/monitoring host. Set to “none” for no IP address. |
|
The port of the commissioning/monitoring host. Set to “0” for no port. |
|
The unique ID for this simulation. This ID is needed for self-identification to the UDP host. Set to “none” for no simulation ID. |
Open a Command Prompt window and navigate to the working directory, which contains the configuration and campaign files.
Enter a command like the one illustrated below, substituting the appropriate paths and file names for your simulation:
../Eradication.exe -C config.json -I C:\|EMOD_s|\InputFiles -O Sim1Output
If you do not specify anything when invoking Eradication.exe, it will not execute with all defaults, but will instead tell you how to invoke the
--helpcommand.EMOD will display logging information, including an errors that occur, while running the simulation. See Error and logging files for more information.
Run a simulation using mpiexec¶
The application mpiexec is used to run multi-node simulations in parallel. Eradication.exe is “single threaded”, so it uses only one core for processing. If you run a simulation with multiple geographic nodes using mpiexec instead of invoking Eradication.exe directly, multiple copies of Eradication.exe will be running, with one copy per core processing data for a single node at a time. Message Passing Interface (MPI) communicates between the cores when handling the migration of individuals from one node to another.
Although mpiexec can be used to run a simulation in parallel on your local computer, it is more often used to run complex simulations in parallel on an HPC cluster or several linked computers. Mpiexec is part of the Microsoft HPC Pack 2012 SDK (64-bit) installed as a prerequisite for building Eradication.exe from the EMOD source code. See EMOD source code installation for more information.
Note
If you get an error that the mpiexec command cannot be found, you must add the path to mpiexec to the PATH environment variable. For example, open Control Panel and add the path C:\Program Files\Microsoft HPC Pack 2012\Bin to PATH.
Take note of the number of cores on your computer or cluster.
If running locally, we recommend running mpiexec with one fewer cores than are available, so one core is reserved for the operating system. The simulation can be run on all available cores and will complete faster, but the desktop will not be responsive.
Open a Command Prompt window and navigate to the directory that contains the configuration and campaign files for the simulation.
Invoke Eradication.exe using mpiexec as follows, replacing the number of cores, paths, and command options as necessary for your environment. See Run a simulation using the command line for more information about the command options available for use with Eradication.exe.
mpiexec -n 3 ..\Eradication.exe --config config.json --input-path ..\InputDirectory\Garki --output-path OutputGarki
Mpiexec will start multiple copies of Eradication.exe as specified by -n. Those instances will
communicate with each other via MPI. If all cores are on a single computational node or host, they
will use internal networking to carry the MPI packets.
Note
The EMOD executable (Eradication.exe) does NOT use the configuration file parameter Num_Cores, which is used by the infrastructure that runs the simulation, such as an HPC cluster or regression script.
You can also link together several computers with MPI using the mpiexec -host option. For
example, if you were using six cores on two computers, you could run three copies of Eradication.exe
on the first computer, and three more could be run on the second computer. Again, this assumes that
each computer has at least three cores.
For more information about mpiexec, see MSDN.
Run a simulation using Python¶
If you used Python to create or modify JSON files as shown in Create a configuration file or Create a campaign file, it may be convenient to invoke Eradication.exe to run a simulation from a Python script. One way of doing this is shown below using the subprocess package.
import subprocess
# specify paths
binary_path = "binDirectory\Eradication.exe"
input_path = "inputDirectory\Namawala\
# commission job
subprocess.call( [binary_path, "-C", "config.json", "--input", input_path] )
See Run a simulation using the command line for more information about the command options available for use with Eradication.exe.
Run a simulation using MATLAB¶
If you used MATLAB to create or modify JSON files as shown in Create a configuration file or Create a campaign file, it may be convenient to invoke Eradication.exe to run a simulation from a MATLAB script. One way of doing this is shown below using the dos command.
exe_name = fullfile( binDirectory, 'Eradication.exe' );
exe_args = [ '-C config.json -I ', fullfile( inputDirectory, 'Namawala' ) ];
[status,result] = dos( ['cd ', WorkDirectory, ' && ', exe_name, ' ', exe_args ], '-echo' );
See Run a simulation using the command line for more information about the command options available for use with Eradication.exe.
Improve EMOD performance¶
While small simulations can be run quickly on a local computer, the time and memory needed to complete a simulation can grow significantly when simulations become larger and more complex. For example, a single geographical node simulation might use 3 GB with one core in order to run successfully in one minute or less, while another simulation job may require 32 GB in a dual-core system in order to complete in approximately the same amount of time.
As your simulations grow, you will likely want to run simulations on an HPC cluster or take other steps to improve performance and reduce processing time. This topic describes many of the steps you can take to speed up EMOD simulations.
The EMOD executable (Eradication.exe) is “single threaded”, meaning that for processing, it will use only one core. However, you can use the Message Passing Interface (MPI) to run multiple copies of Eradication.exe in parallel, either locally or on an HPC cluster. To run a simulation in parallel, you must invoke Eradication.exe with the mpiexec command. For more information, see Run a simulation using mpiexec.
You can also set various parameters in the configuration file that will improve performance by scaling down the amount of data used or optimizing data processing. See the parameters listed in the sections below for guidance on making performance adjustments. See Configuration parameters for more information about each of the parameters.
- Simulation_Duration
Obviously, simulating a shorter timespan will take less processing time. However, the processing time is often driven by the number of infections or immune updates, so running a simulation after all infections have cleared may not increase processing time much.
- Individual_Sampling_Type and Base_Individual_Sample_Rate
Instead of the default Individual_Sampling_Type setting of “TRACK_ALL”, you can speed up performance by using “FIXED_SAMPLING” with the Base_Individual_Sample_Rate set to less than one. For example, a simulation with a population of 1 million and sample rate of 0.1 would simulate 100,000 individuals, with each given a weight of 10. There are also sampling parameters that sample the population at different rates by age and by the population of the node.
However, you should be especially careful not to undersample simulations to the point where they are overly sensitive to rare stochastic events.
- Population_Scale_Type and Base_Population_Scale_Factor
Alternatively, you can simply reduce the total population of the simulation using Population_Scale_Type set to “FIXED_SCALING” and Base_Population_Scale_Factor set to less than one.
- Num_Cores
For large, spatially distributed simulations, running the intra-node dynamics (for example, infection and immune dynamics) in parallel on multiple cores may be very advantageous. Ideally, the timing would be reduced inversely to the number of cores. However, there are costs to serializing individuals for migration over MPI, as well as considerations for balancing the CPU load on each core. These issues may be mitigated using a load balancing input file that is suitable to the geography being simulated. See Load-balancing file structure for more information.
- Simulation_Type
“VECTOR_SIM” will run much faster than “MALARIA_SIM” if detailed intra-host dynamics are not “required.
- Vector_Sampling_Type and Mosquito_Weight
The individual vector simulation (“TRACK_ALL_VECTORS”) will run more slowly than the cohort model (“VECTOR_COMPARTMENTS_NUMBER”). If the individual vector model must be used, you can improve performance by sampling some fraction of all vectors (“SAMPLE_IND_VECTORS”) where each mosquito has a weight configured by Mosquito_Weight, which is analogous to individual human sampling.
- Infection_Updates_Per_Timestep
Using fewer intra-day updates to the malaria intra-host model will speed up simulation time. However, updates longer than three hours, such as eight per day, result in increasingly inaccurate dynamics.
Use output files¶
At the end of the simulation, you will notice that a number of files have been written to the output directory. Some are logging or error output files, which you can read about more in Error and logging files. The rest are output reports that contain data from the simulation itself, usually in JSON or CSV format. This topic describes how to parse and use the output reports.
By default, the output report InsetChart.json is always produced, which contains per- timestep values accumulated over the simulation in a variety of reporting channels, for example, “New Infections”, “Adult Vectors”, and “Parasite Prevalence”. EMOD provides several other built-in reports that you can produce if you enable them in the configuration file. See Output report structure for additional information about the reports and how to enable them.
If none of the built-in output reports provide the data you need, you can use a custom reporter that plugs in to the Eradication.exe as a EMODule dynamic link library (DLL). For more information, see Custom reporters.
In order to interpret the output of EMOD simulations, you will find it useful to parse the output reports into an analyzable structure. For example, you can use a Python or MATLAB script to create graphs and charts for analysis.
Use Python to parse data¶
The example below uses the Python package JSON to parse the file and the Python package matplotlib.pyplot to plot the output. This is a very simple example and not likely the most robust or elegant. Be sure to set the actual path to your working directory.
import os
import json
import matplotlib.pyplot as plt
# open and parse InsetChart.json
ic_json = json.loads( open( os.path.join( WorkingDirectoryLocation, "output", "InsetChart.json" ) ).read() )
ic_json_allchannels = ic_json["Channels"]
ic_json_birthdata = ic_json["Channels"]["Births"]
# plot "Births" channel by time step
plt.plot( ic_json_birthdata[ "Data" ], 'b-' )
plt.title( "Births" )
plt.show()
Use MATLAB to parse data¶
The example below uses the MATLAB toolbox JSONlab to parse an InsetChart.json file and plot one channel. This script uses JSONLab to parse the file into a usable form in MATLAB. This is a very simple example and not likely the most robust or elegant. Be sure to set the actual paths to JSONlab and your working directory.
% this sample uses JSONLab toolbox
addpath('PATH TO/jsonlab');
% open and parse InsetChart.json
ic_json = loadjson( fullfile( 'WorkingDirectoryLocation', 'output', 'InsetChart.json' ));
ic_json_allchannels = ic_json.Channels;
ic_json_birthinfo = ic_json_allchannels.Births;
ic_json_birthdata = ic_json_birthinfo.Data;
M = num2cell(ic_json_birthdata);
% plot "Births" channel by time step
plot(cell2mat(M));
title( 'Births' );
Custom reporters¶
Reporters extract simulation data, aggregate it, and output it to a file known as an output report. You can process the report data using tools, such as Python or MATLAB, to create graphs and charts for analysis. EMOD provides built-in reporters that are part of the EMOD executable (Eradication.exe) and can be enabled or disabled by setting parameters in the configuration file. See Output report structure for a list of the reports that are available using built-in reporters.
In addition, you can use custom reporters that extract data from the simulation and aren’t part of the Eradication.exe. A custom reporter is an EMODule that you plug in to EMOD. Custom reporters are not supported for CentOS on Azure. There are several reporters in the GitHub reporters directory that you can use. You may also want to build your own custom reporter to create a new output report.
The Eradication.exe must load the reporter when running a simulation to use it. If it is loaded, the output report will be automatically generated at the end of the simulation. There are three ways to specify their location so Eradication.exe can load them. Eradication.exe attempts to load the file using the following methods, in the order listed:
Define the location of the dynamic link library (DLL) in a JSON-formatted file (emodules_map.json).
Point to the location using the
--dll-pathcommand-line option when invoking Eradication.exe.Place the DLL in the working directory.
First, Eradication.exe looks for an emodules_map.json file in the working directory. This map file lists specific reporter DLLs and their exact location. This is useful if you want to store all custom reporters in the same directory, but only want certain reporters to be used with each simulation.
The emodules_map.json file has three sections: diseases, interventions and reporters. Each section is a JSON array and contains the path to each EMODule in the array. Although reporter EMODules are the only type currently supported, empty disease and intervention sections must be included. You must use either a double backslash (\) or a single forward slash (/) to represent a single backslash for the path. For example:
{
"disease_plugins": [
],
"interventions": [
],
"reporter_plugins": [
"C:\\src\\EMOD\\x64\\Release\\emodules\\reporter_plugins\\libreportpluginbasic.dll"
]
}
If a map file is not provided, you can include the --dll-path option when you invoke Eradication.exe. It
must point to a directory with a reporter_plugins subdirectory that contains the reporters. The name
of the directory does not matter, but the reporter must be in a subdirectory named reporter_plugins.
Note that this method will load all reporters in the root directory specified.
If you build custom reporters as described here, the DLLs will be saved to the following directories, depending on the build configuration:
<EMOD source directory path>x64Releasereporter_plugins
<EMOD source directory path>x64Debugreporter_plugins
Therefore, if you do not move your reports, you will set --dll-path to one of the following paths:
<EMOD source directory path>x64Release
<EMOD source directory path>x64Debug
For more information on EMOD command-line options, see Run a simulation using the command line.
Finally, if neither the the map file or the command-line option are provided, Eradication.exe looks for the reporter plug-in in the current working directory.
Malaria disease overview¶
Welcome to malaria modeling at the Institute for Disease Modeling (IDM)! We are committed to supporting data-driven malaria control and elimination efforts. This page will provide information about the disease itself. The following pages will provide information about the the vector class of Epidemiological MODeling software (EMOD), the software developed by IDM to aid in malaria elimination.
Malaria biology¶
Malaria is a mosquito-borne disease caused by plasmodium parasites transmitted through the bite of blood-feeding Anopheles mosquitos. There are roughly 30 species of Anophelenes capable of serving as malaria vectors, and only females bite (i.e. seek blood meals) to gain nutrients necessary for oviposition. After taking blood meals, females lay eggs in water sources, where the emerging larvae hatch and mature. Each species of Anopheles has different aquatic habitat preferences, ranging from small, ephemeral pools to large swampy areas, and in some cases, brackish water.
Transmission of malaria depends on both biotic and abiotic factors. Mosquito lifespan is positively correlated to transmission intensity. Mosquito populations, and therefore transmission, tend to be seasonal with dependence on climatic conditions such as rainfall patterns, humidity, and temperature. Typically, malaria transmission peaks during and after the rainy season, when conditions are prime for mosquito reproduction. Additionally, transmission is dependent on human immunity: people living in areas of high exposure to malaria can develop partial immunity, which reduces the risk of developing severe disease. Unfortunately, such partial immunity may mask symptoms of disease (and obscure infection), which may hinder control or elimination efforts.
Once an infected mosquito bites a human and transmits parasites, the parasites multiply in the host’s liver, then infect and destroy red blood cells (erythrocytes). During the blood stage, the parasites that infect and destroy erythrocytes release merozoites which then infect other erythrocytes. The gametocyte form of the blood-stage parasite are ingested by female Anophelenes during future blood meals, where the mosquito-based life cycle continues.
In the mosquito, the ingested gametocytes generate zygotes in the mosquito’s gut. Those zygotes undergo development, embed into the mosquito’s midgut wall, and mature into oocytes. Oocytes develop and in turn rupture, releasing the sporozoite form of the parasite. The sporozoites travel to and reside in the mosquito’s salivary glands. Then, when the female takes a blood meal, the sporozoites are injected into the host, starting another infection cycle. The mosquito is the vector that transmits pathogens between human hosts. See the below figure and figure caption for an illustrated description of this process.
Plasmodium life cycle, adapted from the CDC, Malaria Biology. During a blood meal, a malaria-infected female Anopheles mosquito inoculates sporozoites into the human host (1) . Sporozoites infect liver cells (2) and mature into schizonts (3), which rupture and release merozoites (4). After this initial replication in the liver (exo-erythrocytic schizogony, “A”), the parasites undergo asexual multiplication in the erythrocytes (erythrocytic schizogony, “B”). Merozoites infect red blood cells (5). Some parasites differentiate into sexual erythrocytic stages (gametocytes) (7). The gametocytes are ingested by an Anopheles mosquito during a blood meal (8). The parasites’ multiplication in the mosquito is known as the sporogonic cycle, “C”. While in the mosquito’s stomach, the gametocytes generate zygotes (9). The zygotes in turn become motile and elongated (ookinetes) (10) which invade the midgut wall of the mosquito where they develop into oocysts (11). The oocysts grow, rupture, and release sporozoites (12), which make their way to the mosquito’s salivary glands. Inoculation of the sporozoites into a new human host (1) perpetuates the malaria life cycle.¶
Symptoms¶
All the clinical symptoms associated with malaria are caused by the asexual erythrocytic (or blood stage) parasites. Malaria has a long incubation period, so symptoms do not develop immediately after a person has been bitten by an infectious mosquito; they can occur 7 days after infection, but this may vary up to 30 days.
Symptoms of malaria range in severity, but include fever, headache, body-ache, chills, and vomiting. Severe malaria can develop when the infection is not treated, and may result in organ failure or even death. Examples of severe malaria include cerebral malaria, severe anemia, respiratory distress (due to accumulated fluid in the lungs), kidney failure, metabolic acidosis, hypoglycemia, and other maladies. Death can occur from anemia, hypoglycemia, or cerebral malaria (where capillaries leading to the brain are blocked; typically resulting in coma, brain damage or learning disabilities).
Confirmed diagnosis of malaria is done through observation of parasites in the blood, as seen on the below microscopic image. In highly endemic areas, treatment typically occurs prior to diagnostic confirmation, as malaria is more easily cured with prompt treatment.
Human red blood cells infected with Plasmodium falciparum (the parasite is shown in dark purple). Photo credit Le Roch lab, UC Riverside.¶
Prevention & control efforts¶
There are numerous types of strategies used to control malaria. As a vector-borne disease, there are multiple stages at which the transmission cycle can be broken.
Vector control: strategies that take into account vector ecology. These include:
Chemical control, such as insecticide spraying or use of larvicides.
Reduction of or elimination of mosquito larval habitat, through drainage or use of biological controls.
Potential use of genetic modification (with tools such as CRISPR) to create mosquitoes that are resistant to infection from Plasmodium parasites.
Potential use of the bacterium Wolbachia to prevent mosquitoes from becoming infectious.
Personal protection: strategies that avoid infection (by avoiding bites by infectious mosquitoes), or by preventing disease. These include:
The use of insecticide-treated nets (ITN)
Administration of antimalarial drugs to particularly vulnerable groups, such as children or pregnant women
Global malaria burden¶
While progress towards reducing the malaria burden has been largely successful, malaria nevertheless remains a major health problem and target of focused, global efforts for elimination and eradication. According to the CDC, 3.2 billion people worldwide are at risk of malaria. In 2015, there were 214 million cases with 483,000 deaths. Malaria is especially harmful to children: more than 70% of all malaria deaths occur in children under the age of 5. To put that in perspective, a child dies from malaria roughly every 2 minutes.
While malaria is a global problem, the burden is not distributed equally across the globe. Sub-Saharan Africa experiences a disproportionately high burden, with about 76% of all cases and 75% of all deaths. South East Asia, Latin America, and the Middle East are also at high risk for malaria.
Benefits of mathematical models in malaria control and eradication¶
Controlling and eventual eradication of malaria will require multifaceted and geographically-specific intervention efforts. Heterogeneity in transmission, and transmission potential, creates a need for combinations of interventions that can adapt to the particular malaria epidemiology of the target area.
For malaria, mathematical modeling and simulations are key to achieving eradication. Species- specific vector ecology is a fundamental driver of transmission, and transmission also impacted by the interactions of climate, human behavior, and land usage, across varying spatial scales. Modeling these factors enables accurate representations of baseline transmission, which in turn provides a platform to test various interventions (such as insecticide-treated nets (ITN), indoor residual spraying (IRS), or mass drug administration (MDA)) and combinations of interventions. These simulation results can inform policy to develop effective–and cost effective–strategies by exploring the many possible dimensions of coverage, frequency of distribution, and combinations of interventions targeted to particular locations.
It should be noted that such a multifaceted and integrated approach to vector control and health management is the most likely path towards elimination, as heavy reliance on singular approaches can be problematic. For example, vector control has been widely implemented and quite successful; however, mosquitoes are developing resistance to pyrethroids, one of the most common classes of insecticides currently recommended for ITN or IRS to control mosquito populations. Through modeling approaches, researchers will be able to develop strategies that lessen dependence on particular insecticides while maintaining successful control efforts.
A brief history of malaria modeling¶
Malaria has a long history of posing risks to public health, and as such, also has long been the target of mathematical models tasked with providing solutions to ease the burden. For a more detailed history of malaria models, see Smith et al. PLOS Pathogens 2012 8(4), and Smith et al. Trans R Soc Trop Med Hyg 2014 108(4). The EMOD malaria model builds upon this rich history of disease modeling to provide a novel and rigorous approach to help guide efforts towards malaria elimination and eradication. To fully understand the strengths of EMOD, it is helpful to understand the modeling background from which EMOD developed.
Arguably, the quantitative analysis of mosquito-borne diseases, specifically malaria, began with Ronald Ross. In 1897 Ross confirmed that mosquitoes serve as the vector for malaria parasites, and embarked on a career focused on disease prevention through vector control. Further, Ross developed the modeling framework that serves as the basis for studying malaria transmission dynamics.
Ross’s work inspired researchers focused on controlling mosquito-borne diseases. His emphasis on the development of metrics useful for measuring the intensity of transmission led to the development of :term:` entomological inoculation rate (EIR)` and motivated research aimed at understanding mosquito movement and the relevant spatial scales for mosquito ecology. Later, in the 1950s, George MacDonald formalized Ross’s models, and introduced the concepts of vectorial capacity and reproductive number. This framework is now known as the Ross-Macdonald model, and is still widely implemented in current modeling work. In fact, the vast majority of models created in the last 40 years share most of their assumptions with the Ross-Macdonald model (see Reiner et al. 2013, Journal of the Royal Society Interface.)
The Garki project¶
The Garki project was a major milestone in malaria research. From 1969 to 1976 this study was conducted by the World Health Organization (WHO) and the local government of the Jigawa State, Nigeria, to understand the impacts of indoor residual spraying (IRS) and mass drug administration (MDA) on malaria transmission, as well as to evaluate the utility of mathematical modeling. While the interventions used did not interrupt transmission at the desired level, the model proved to be a success. The epidemiology of the Plasmodium parasite was realistically replicated, even with a simplistic model, and as a result understanding of the parasite’s epidemiology was greatly increased.
While the project was largely considered a failure in terms of malaria control, much was gained from the project. Basic tenets of malaria control were learned– namely, that malaria ecology needs to be fundamentally altered: either by modification of mosquito ecology, or by changing human ecology such as by improving housing conditions. The superinfection, a fundamental malaria concept modeled by Macdonald in his first malaria model publication [Macdonald-1950], was more correctly described by Dietz [Dietz-1974]. Further, the data set from the project is publicly available, and has become a fundamental tool for use in malaria modeling. For more information on this project, see The Garki Project.
Moving forward¶
The Ross-Macdonald model, while extremely useful and influential, has some shortcomings. The model assumes homogeneous transmission within a well-mixed population. Host-vector ratios and numerous aspects of mosquito feeding and development biology is assumed to remain constant. Further, all hosts are assumed to be identical, with equal exposure to pathogens, and the probability of transmission is proportional to host and vector densities. Interestingly, despite work demonstrating that these assumptions are not realistic, many models still utilize them.
To improve on the Ross- Macdonald model, it is necessary to understand and implement increased complexity into the transmission dynamics of the model. Transmission dynamics are not linear, and instead depend on fine-scale heterogeneities; ecological as well as social contexts are extremely relevant for potential host-vector interactions, and both human and mosquito movement across relevant spatial scales also impact malaria transmission.
Modern research has continued to improve our understanding of both parasite biology (and genetics) as well as within-host immune system interactions. To fully understand the complex nature of malaria epidemiology and how to control it, it is important to include these aspects into models.
The malaria EMOD leverages the history of malaria modeling by staring with these important fundamentals and building upon them. EMOD combines detailed vector population dynamics and interactions with human populations, and includes microsimulations for human immunity and within- host parasite dynamics. The model builds on the work of Ross and MacDonald, leverages the Garki model, and incorporates current modeling efforts to model multiple vector species simultaneously interacting with a human population.
References¶
- Macdonald-1950
Macdonald G (1950) The analysis of infection rates in diseases in which superinfection occurs. Trop Dis Bull 47: 907–915.
- Dietz-1974
Dietz K, Molineaux L, Thomas A (1974) A malaria model tested in the African savannah. Bull World Health Organ 50: 347–357
Malaria model overview¶
The Epidemiological MODeling software (EMOD), malaria model, is an agent-based, discrete time, Monte Carlo simulator used to simulate malaria conditions in order to evaluate the effectiveness of eradication or mitigation approaches. The model utilizes microsolvers to combine detailed vector population dynamics, human population dynamics, human immunity, within-host parasite dynamics, effects of antimalarial drugs, and other aspects of malaria biology to simulate malaria transmission. EMOD can be calibrated to particular geographic locations, and the microsolver framework enables the model’s functionality to be highly modifiable. Further, the framework includes the ability to add intervention campaigns, and those interventions can be specified to target particular populations or sub-populations of human or vector groups. Interventions can be deployed within simulations for a variety of transmission settings with different transmission intensities, vector behaviors, and seasonally-driven ecologies.
The following pages describe the structure of the model, and then explore each of the model components. Additionally, the specifics of the model are discussed in detail in the articles Eckhoff, Malaria Journal 2011, 10:303; Eckhoff, Malaria Journal 2012, 11:419; Eckhoff, PLoS ONE 2012, 7(9); and Eckhoff, Am. J. Trop. Med. Hyg. 2013, 88(5).
The generic software architecture of EMOD can be found on Introduction to the software. The malaria EMOD is based on the same architecture, with modifications that transform the generic model into a vector-based model. For example, transmission occurs via vectors (and the associated climatalogical impacts on vector biology), individuals can have multiple infections, and campaign interventions can act at the individual level or at the node level.
Typical SEIR models are described in Introduction to disease modeling. Vectors add a level of complexity to the interactions, as pathogens are transmitted via vector - human - vector. For malaria, the SEIR model would appear as follows:
While EMOD is an agent-based model, both the simulated humans and vectors move through the various infection states analogously to the compartmental model illustrated above. To account for the real-world complexity of malaria transmission, numerous parameters have been added to EMOD to increase its predictive power. Further, these parameters and their associated microsolvers allow EMOD to model aspects of malaria infections and population dynamics that do not readily fit into the SEIRS framework (for example, within-host parasite dynamics, parasitalogical immunity, and innate and adaptive responses to antigens). Finally, as an agent- based model, EMOD enables the addition of spatial and temporal properties to the simulation. The following pages describe this complexity in detail.
Malaria model general information¶
Here we describe the broad overview of how the malaria model functions. As there are numerous parameters in the model, it is useful to categorize parameters into components, which are described in the Malaria model structure and framework section.
The core of the simulation consists of solvers for mosquito population dynamics, pathogen transmission to and from the human host population, and dynamics of the infection within the human host. While the human population is fully individual-based, the mosquito population can be represented by discrete cohorts or individual agent mosquitoes. As an agent-based model, interactions between humans and vectors advance during each time step. The mosquito population is advanced, for example, through discrete 1-day time steps during which mosquitoes may engage in host- seeking or feeding behaviors, mating, maturation through a life-stage, or even death. Successful feeds on humans can result in infection of the human host, or conversely, susceptible mosquitoes may become infected from an infectious human.
Model implementation structure¶
There are two categories of possible implementations of the basic model, each with different computational efficiencies, resolutions, and flexibilities. The first is an individual model, where it simulates every individual mosquito in the population or can utilize a sampled subset of mosquitoes to represent the population as the whole. The second is a modified cohort simulation, with or without explicit mosquito ages.
Individual mosquito model¶
This basic model can be implemented through simulation of every individual mosquito, or by simulation of a subset of individual mosquitoes to represent the full population. Each mosquito’s state contains status (susceptible, latently infected, and infectious), timers for transition to adult from immature and infected to infectious, mating status and Wolbachia infection, and age. An oviposition timer to enforce a fixed feeding cycle may be included as well. If mosquitoes are sampled and a subset used to represent the local population, each sampled mosquito will have an associated sampling weight as well.
Cohort model¶
In the modified cohort simulation, rather than representing the entire population by three compartments for susceptible, latently infected, and infectious mosquitoes, the simulation dynamically allocates a cohort for every distinct state, and the cohort maintains the count of all mosquitoes in that state. This allows temperature-dependent progression through sporogony, even with a different mean temperature each day, with no mosquitoes passing from susceptible to infectious before the full discrete latency. For the cohort simulation with explicit ages, in order to allow modeling of senescence, mosquito age is part of the state definition, and many more cohorts are required to represent the population.
Discrete and continuous processes¶
The model implements a hybrid of discrete and continuous processes that work together to capture system latencies and discrete events inherent in the population dynamics of humans, mosquitoes, and parasites. Discrete events, such as latencies in the infected hepatocyte stage, the length of the asexual cycle from merozoite invasion to schizont rupture, and gametocyte maturation have particular time durations. State changes occur after the completion of the required amount of time (with specific time durations set for each state). See the figure of the Plasmodium life-cycle in Malaria disease overview for more information regarding the life-cycle stages used in this example.
Other processes, such as the decay of antibodies and the clearance of parasites, are represented by continuous-time processes. These are solved with a one-hour time step Euler method. All parasite quantities, such as the number of hepatocytes, number of merozoites, number of infected red blood cells of each antigenic variant, and gametocytes of each stage are represented as discrete integers. The infection is not cleared until each category is reduced to zero. This allows resolution of model dynamics at subpatent levels.
Spatial-scale dynamics and migration¶
Extensive multi-node simulations with location-specific climate, intervention deployments, larval habitat, and migration may be configured within the EMOD framework. Nodes can be configured to best represent the spatial scale being simulated: a node can be the household level, village level, county level, country level, or other desired geographic scale. Migration can be enabled between nodes, such that EMOD can most accurately represent the dynamics of the location. By varying spatial-scale resolution, relevant factors for transmission dynamics can be more clearly elucidated.
The following visualization shows an example of a gridded representation of malaria transmission on the island of Madagascar.
For information on configuring nodes and migration parameters, see Migration file structure.
Malaria model structure and framework¶
The malaria EMOD is complex, with numerous configurable parameters. The following network diagram breaks down the model into various model components, and illustrates how they interact with one another. The components on the network diagram correspond to the structural components listed below. Note that there is not perfect overlap between the labels on the network diagram and the structural components; this is because the network is drawn with increased detail in order to provide clarity in how the model functions and the components interact. The following pages will describe in detail how the structural components function.
Network diagram illustrating the malaria model and its constituent components.¶
See the EMOD parameter reference for a complete description of all configurable parameters for the malaria model.
Malaria model structural components¶
The model’s modular framework includes the following components:
Climate and demographics spatial data¶
Climate is especially important in vector models, as it is a key determinant of the geographic distribution and seasonality of malaria. Climate, which includes rainfall, temperature, and humidity, directly influences availability of larval habitat, determines mosquito developmental and survival rates, and impacts human behavior that leads to contact with infectious mosquitoes.
This framework explores the impact of different locations and weather patterns in single node and multi-node geographies. It explores demographics data for births, deaths, age initialization and immunity. Climate and demographics spatial data is configured for a geographic location, such as a rainfall-driven larval habitat, to create a climate-based distribution model of malaria transmission. The seasonality of species abundances are considered, as are rainfall, temperature, humidity and larval dynamics. Mosquito species and seasonality are depicted by habitat preferences that are, for example, temporary, semi-permanent or permanent.
Clearly, climate and habitat type are inextricably linked. However, it is possible to have several types of habitat within a given climate region; for this reason, climate is configured separately from habitat type. You may choose from a variety of options to configure climate, ranging from pre-set classification systems to uploaded user data. Further, it is possible to enable stochasticity in climate as well as in rainfall patterns. For example, rainfall data may be based off of monthly totals averaged per day. In the simulation configuration file, the parameter Enable_Rainfall_Stochasticity will set up daily rainfalls drawn from an exponential distribution with the same daily mean. This preserves approximate monthly totals, but produces a more realistic rainfall pattern for the simulation. For more information on how to configure climate, see Climate file structure. And for more information on the particular habitat types that are configurable in EMOD see Larval habitat.
The EMOD vector model utilizes site-specific climate and demographics data to accurately simulate vector transmission in a given geographic location.¶
See Input data files for more information on configuring demographics. For a complete list of demographics parameters, go to Demographics file structure and parameters.
Larval habitat¶
Available larval habitat is a primary driver of local mosquito populations, and different mosquito species can have different habitat preferences. Rainfall and humidity can strongly affect available larval habitat, although this depends on the mosquito species and its particular habitat preference. For example, Anopheles funestus, which prefers more semi-permanent larval habitat, exhibits rainfall dependence, partly due to vegetation on edges of water and the interaction of rainfall with agricultural schedules for crops such as rice. Further, preference can vary within species: An. funestus exhibits differences in population responses to rainfall which are correlated with chromosomal diversity.
In addition to available habitat, mosquito populations depend on larval development and mortality rates. These in turn are affected by a variety of factors, including climate, short term weather, and densities of other larvae.
In the present model, different models for larval habitat are developed for temporary, semi- permanent, permanent, and human-driven habitats. The duration of larval development is a decreasing function of temperature, and the present model utilizes an Arrhenius temperature-dependent rate a1(a2/TK). In some cases, this temperature-dependent rate must be modified by local larval density. Rainfall and temperature then combine through habitat creation and larval development to create varying local patterns of distribution by larval instar, and larval mortality and development duration determine pupal rates.
The creation of the habitat model is described in further detail in the articles Eckhoff, Malaria Journal 2011, 10:303, Eckhoff, Malaria Journal 2012, 11:419, and Eckhoff, Am. J. Trop. Med. Hyg. 2013, 88(5). The following sections provide information regarding specific larval habitat types and how to configure these parameters in EMOD.
For a complete list of configuration parameters that are used in the malaria model, see the Configuration parameters.
The framework facilitates simulating multiple species of Anopheles mosquitoes simultaneously. This allows for a mechanistic description of vector abundances through the effects of climate and weather on different preferred larval habitats. Each species is configured separately according to its ecological and behavioral preferences. For example, the female An. arabiensis deposits eggs primarily in temporary, rainfall-driven habitat and has a higher propensity to feed outdoors or on livestock. In contrast, An. farauti larvae live in brackish lagoons in the Solomon Islands, where lagoons fill with rain, and larvae are washed out into the ocean due to excess rain. The temporary rainfall relationship between larval habitat and rainfall cannot reproduce abrupt (sub-week) changes in biting rates, and an additional rainfall-driven larval mortality term has been implemented to capture the non-linear rain-to-habitat relationship.
Vector abundance depends on larval habitat availability¶
Larval habitat type determines how rainfall and temperature influence vector populations¶
The parameter Larval_Habitat_Types is a dictionary that contains one or more string and integer pairs. The user indicates the vector species, habitat type(s), and the array of scaling factors used to determine the volume of that habitat needed for larvae to develop. Note that multiple species may utilize the same habitat, with each species existing independently. The scaling factor represents the number of larvae in a 1x1-degree area. The factor multiplicatively scales the resulting weather or population dependent functional form. The following example shows how to use this parameter, using Anopheles funestus as the vector and CONSTANT as the scaled habitat type.
{
"Vector_Species_params": {
"funestus": {
"Larval_Habitat_Types": {
"CONSTANT": 10000000
},
"Acquire_Modifier": 0.2
}
}
}
The following are possible values for Larval_Habitat_Types. They are described in detail below.
TEMPORARY_RAINFALL
WATER_VEGETATION
CONSTANT
BRACKISH_SWAMP
HUMAN_POPULATION
LINEAR_SPLINE
TEMPORARY_RAINFALL habitat corresponds to mosquitoes such as Anopheles arabiensis or Anopheles gambiae which breed primarily in temporary puddles which are replenished by rainfall and drain through evaporation and infiltration. We have developed a series of update equations for the larval carrying capacity which scale the functional form for evaporation rates. This creates a relationship in which evaporation and infiltration are higher when the weather is hot and dry.
The basic update equations appear in the article Eckhoff, Malaria Journal 2011, 10:303 and are as follows: Temporary habitat Htemp in a grid of diameter Dcell increases with rainfall Prain and decays with a rate \(\tau\)temp proportional to the evaporation rate driven by temperature T(K) and humidity RH:
in which the exponential results from the Clausius-Clayperon relation, the root is from the expression for vapor evaporation rates due to molecular mass given a partial pressure, and the constant is the Clausius-Clayperon integration constant multiplied by a factor ktempdecay to relate mass evaporation per unity area to habitat loss.
Examining the two equations more closely, the first is a time step update equation, in which available habitat is increased by a multiple of the rainfall in the past time step, scaled by the size of the node in the simulation. For example, a 5 km square node will have 25 times the habitat of a 1 km square node. Then existing habitat decays with a time constant defined by the second equation. The basic functional form is based on an equation for evaporation rates, which is not fully realistic in that it neglects boundary layer effects. However, we are looking for a rate of habitat loss, not the rate of evaporation, and we also want to take infiltration into account. The parameter ktempdecay converts raw evaporation rates in kg/m^2/s into habitat loss per day.
The value of ktempdecay is initially chosen to set the habitat half-lives near 1 day for hot and dry conditions and 2-3 weeks for more tropical conditions. The parameter ktempdecay in the article Eckhoff, Malaria Journal 2011, 10:303 corresponds to Temporary_Habitat_Decay_Factor in the simulation configuration file, and it applies to all local species using temporary habitat. Each species in a given location can adjust the parameter ktemp from the paper, or equivalently the scaling factor value in Larval_Habitat_Types in the simulation configuration file (nested under the species name), to adjust the overall scaling of the time series. So ktemp is specific for each species, but ktempdecay is a single value for all temporary habitat species in that location. Basically, the local weather and ktempdecay set the overall time profile for larval habitat which can feed-forward into adult population levels and biting rates. ktempdecay can be adjusted to achieve a best fit for the time profile for a given location. Factors such as how sandy or clay-like the soil is will affect this value. Once the time profile is correct, ktemp can be adjusted to yield the correct annual entomological inoculation rate (EIR) for that species.
The second type of larval habitat, WATER_VEGETATION, is a semi-permanent habitat, which corresponds to developing vegetation on the edges of semi-permanent habitat. This could be seen for swamp-like or rice cultivation settings, in which the development of vegetation lags the rainfall and will be closest to its peak towards the end of the rainy season. The decay is specified as a loss of habitat per day, and this slower decay constant means that species with this habitat type will tend to have a proportionately higher dry-season population relative to the temporary habitat species.
In the model, semi-permanent habitat increases with a constant KsemiDcell2Prain and decays with a longer time constant \(\tau\)semi(Semipermanent_Habitat_Decay_Rate in the simulation configuration file). The parameter Ksemi is the scale factor value in Larval_Habitat_Types for species with a Larval_Habitat_Types of “WATER_VEGETATION”. Permanent habitat is fixed at KpermDcell2, and human population-driven habitat is calculated as population N* Kpop. The values of ktempdecay, \(\tau\)semi, Ksemi, and Ktemp can be fit to local data on vector abundance by species over time or to local data on EIR to tailor a simulation to a specific setting.
For habitat type CONSTANT, larval carrying capacity is constant throughout the year and does not depend on weather. However, there will be a seasonal signal in adult population levels due to the effects of temperature upon aquatic development times. For a given carrying capacity, a faster development time will allow a local habitat to have a higher larval “through-put” with corresponding impacts on the adult population. The scaling factor value in the Larval_Habitat_Types parameter is used to specify the carrying capacity per unit area Dcell2.
The BRACKISH_SWAMP habitat setting deals with the dynamics of how rain fills a brackish swamp, how it decays and the associated parameter for rainfall-driven mortality threshold. This habitat type is observable in the Solomon Islands where an entire larval generation is flushed away as a result of the overflowing lagoon geography. Since this is the case, the TEMPORARY_RAINFALL relationship between larval habitat and rainfall cannot reproduce these abrupt changes (on 0 week timescales) of biting rates. So, an additional rainfall-driven larval mortality term has been implemented that captures the non-linear rain-to- habitat relationship. This habitat uses the same decay rate as the semi-permanent water vegetation habitat.
Rainfall-driven mortality during swamp flushing¶
The habitat type HUMAN_POPULATION scales with correlates of urban development, for example, water that is available due to water pots in urban areas. This type is configured by multiplying the number of people in the node’s population times the capacity value set in Larval_Habitat_Types. Further, climate is not involved in setting the capacity, as available water is not dependent on rainfall (it is instead dependent on human-provided water sources).
LINEAR_SPLINE is a recent addition to the types of habitat supported. Instead of specifying a habitat type, users may utilize data that tracks the number of larvae measured throughout the year to estimate the daily larval population using linear interpolation. In other words, this option enables the user to use data collected from a specific site, or create data files to match a particular location. This option does not replace the climatological parameters, but it instead adds a more flexible option in which you do not need climate data.
LINEAR_SPLINE has different syntax than the other habitat types, as the user is now required to provide a distribution of the larval population. The following example provides the syntax for configuring this habitat type:
{
"Vector_Species_params": {
"funestus": {
"Larval_Habitat_Types": {
"LINEAR_SPLINE": {
"Capacity_Distribution_Per_Year": {
"Times": [0.0, 60.833, 121.667, 182.5, 243.333, 304.167],
"Values": [0.0, 0.0, 0.2, 1.0, 0.5, 0.0]
},
"Max_Larval_Capacity": 10000000000.0
}
}
}
}
}
Ultimately, habitat is configured in order to create mosquito populations that realistically emulate observed entomological inoculation rate (EIR). Briefly, available habitat is directly related to mosquito abundance, and mosquito abundance in turn is directly related to biting rate. In order to calibrate the model there are several options for configuring habitat. You can first set habitat parameters and modify them directly using the habitat scalar and decay rate. Then, after those initial parameters are set, you can modify habitat with overall scaling parameters.
The following two figures demonstrate the effects of varying the habitat scalar (Larval_Habitat_Types) and the Temporary_Habitat_Decay_Factor (ktempdecay) for a single species with temporary habitat. Changing the habitat scalar will scale the resulting adult population size and biting rate by a similar factor.
Effect of varying the habitat scalar, Larval_Habitat_Types¶
Lowering ktempdecay causes the resulting rainfall-driven habitat to decay at a slower rate and thus increases \(\tau\)temp. A slower decay rate will result in higher larval habitat on average, and higher resulting adult population biting rates.
Effect of varying the decay rate, Temporary_Habitat_Decay_Factor (ktempdecay)¶
These two parameters can be co-varied to produce an appropriate temporal profile. If rainfall is constant, there is only one degree of freedom, as the habitat is always at equilibrium. The two degrees of freedom become important when there is a distinct rainy season. For example, scaling habitat produces the same ratio between biting rates in the wet season and dry season as seen in Effect of Varying Habitat Scalar.
However, when varying the decay rate, the ratio between habitats is different in the wet season versus the dry season. As seen in the graph “Effect of Varying Temporary Habitat Decay Scalar”, a given ratio in rainy season will become exaggerated in the start of dry season (with a higher ratio the drier the season), as the slower decay rate extends habitat longer into the dry season.
Often, it is desired to study a similar location, with the same species, temporal profile, etc., but with a different annual entomological inoculation rate (EIR). EIR depends not only on climate, but on larval habitat availability as well. So as an alternative for modifying all climate and habitat parameters to change EIR, it is simpler to use the habitat scaling parameter, x_Temporary_Larval_Habitat found in the configuration file. This parameter will scale all habitat parameters without changing the temporal dynamics, so that a new transmission is achieved with the same ratios among the species, and same time profile. For example, setting x_Temporary_Larval_Habitat to 0.1 would simulate low EIR (or a low transmission setting) by reducing available habitat to 10%; a value of 1 could be used to simulate high EIR (or a high transmission setting), and there would be no reduction in available habitat.
An alternative to x_Temporary_Larval_Habitat is LarvalHabitatMulitpilier. LarvalHabitatMultiplier is a parameter in the demographics file, and can be applied to all habitat types configured for the simulation, specific habitat types, or individual mosquito species within particular habitat types in the Nodes array of the demographics file (see NodeAttributes in Demographics file structure and parameters). The following example shows the syntax:
{
"Nodes": {
"NodeID": 340461476,
"NodeAttributes": {
"LarvalHabitatMultiplier": 10.0
}
}
}
It should be noted that LarvalHabitatMultiplier enables habitat availability to be modified independently for each species within a shared habitat. This is an upgrade over previous versions of EMOD in which the modifier would be applied equally to all species within a shared habitat.
Vector transmission model¶
The vector model can be run in one of two modes: as a cohort model or as an individual mosquito model (see Malaria model overview for more information). The vector model inherits the same human-infection model structure from the generic simulation type: uninfected, latent incubation, infectious, multiple immune variables, and super-infection. However, the transmission of infections is not between individual humans, but rather via the human-to-vector and vector-to-human pathways. The model is a closed-loop feeding cycle where a successful vector (mosquito) feeds (both indoors and outdoors), mates, then produce eggs which become larva and then adult mosquitoes.
For example, vector blood-feeding branches into various probabilities that are calculated once per time step. These calculations are based on species parameters and aggregated vector interventions (see feeding decision tree).
Vector transmission can be thought of in terms of vector ecology, and in the EMOD framework is comprised of several components: 1. Vector tracking and the mosquito life-cycle, 2. mosquito behavior, which involves feeding (see Modeling feeding cycle outcomes), and 3. transmission of malaria-causing Plasmodium parasites between humans and mosquitoes (see Transmission between humans and vectors). Each of these components is described in more detail below.
For a complete list of configuration parameters that are used in the malaria model, see the Configuration parameters.
Populations are tracked as cohorts throughout the full mosquito life-cycle:
From eggs to larvae with a temperature-dependent development period preceding emergence
A brief immature phase including sugar-feeding and mating
Repeating multi-day gonotrophic cycles during which mosquitoes may be exposed to infection by gametocytes
A temperature-dependent latency for sporogony
Life-cycle progression is modulated through a set of required species-specific parameters.
Vector life cycle¶
Adult vectors enter a cycle of host-seeking, feeding, and egg-laying that continues until death. Successful vector blood-feeding branches into various probabilities that are calculated once per time step. These calculations are based on species parameters (for example, An. arabiensis has a higher propensity to feed outdoors or on livestock) and aggregated vector interventions.
Various feeding cycle outcomes are calculated from branching trees of conditional probabilities, and individual interventions modulate the probability of choosing between branches. Feeding cycle outcomes include:
Death before, during, or after feeding
Host unavailable
Successful human feed
The allocation of mosquitoes to feeding-cycle outcomes is based on end-state probabilities that have been aggregated over the individual humans in the simulation.
Multiple simultaneous interventions can target various branches in the vector feeding tree. For example, when both indoor residual spraying (IRS) and insecticide-treated nets (ITN) are applied against indoor host-seeking mosquitoes, IRS can discourage mosquitoes from entering the house and kill mosquitoes before feeding. The fraction of mosquitoes that survive can be blocked by the ITN, which may also kill a subset of the blocked fraction. Those mosquitoes who survive the feeding attempt may be killed by IRS post-feed. This is how deterrent and toxic effects of multiple interventions can be represented simultaneously.
To interact with these parameters and visualize the workings of this microsolver, see the decision tree visualization below:
To get started, press the play button. You can also pause the visualization at any time. Parameters in blue are vector species parameters, while parameters in green are types of campaign interventions. Information on these parameters can be found in Configuration parameters and Campaign parameters. The two pink points on the tree illustrate when transmission of malaria parasites is possible.
When the simulation starts, the initial mosquito population is set at 100 individuals. The starting population for day two has an initial seeding of 50 mosquitoes, and also includes all mosquitoes that either live without feeding or feed and oviposit. The simulation includes parameters that determine the lifespan of mosquitoes and the time it takes for oviposited eggs to hatch and mature to adulthood. As time progresses, the population will be comprised of only mosquitoes that are generated through the oviposition cycle in the model.
The counters on the right side of the visualization keep track of current and total mosquitoes that have “spawned” (generated in the simulation), died, lived without feeding, and fed and oviposited.
As an example, let’s simulate Anopheles gambiae. Set Life_Expectancy to 10 (most are thought to live approximately 1-2 weeks in nature), Egg_to_Adult to 5 (this is their minimum duration in the aquatic phase), Days_between_Feeds to 3, and Anthropophily and Indoor_Feeding_Fraction to 0.8. These mosquitoes prefer to primarily feed on humans, and preferentially feed indoors. Now, by changing the interventions, you can see how effective interventions (or combinations of interventions) need to be in order to disrupt (and reduce) mosquito feeding and oviposition. Note that the slider bars for interventions range from 0 - 1, with 1 conferring 100% effectiveness. When mosquito ecology is sufficiently disrupted, malaria transmission can be controlled. You can also manipulate the species parameters to see how mosquito ecology impacts the need for particular types of interventions.
If you are interested in simulating other mosquito species, more information on their relevant attributes can be found in the articles Made-to-measure malaria vector control strategies: rational design based on insecticide properties and coverage of blood resources for mosquitoes, by Killeen et al., 2014, Malaria Journal 13:146, and A global bionomic database for the dominant vectors of human malaria, by Massey et al., 2016, Nature.
Transmission between humans and vectors can only occur when mosquitoes successfully feed on humans (see Modeling feeding cycle outcomes). Infectious humans can infect susceptible mosquitoes when mosquitoes ingest gametocytes during the blood meal. Conversely, infectious mosquitoes can infect susceptible humans by injecting sporozoites into the blood during feeding. Infectiousness of individuals to mosquitoes is proportional to the number of mature gametocytes absorbed in a blood meal, but modulated by a factor that inactivates gametocytes at high cytokine densities. Peak infectiousness typically follows peak parasitemia by about ten days. See the Malaria infection and immune model for more information regarding within-host parasite dynamics, and the figure on the parasite life-cycle in the Malaria disease overview for more information on the parasite life stages.
Transmission between humans and vectors¶
Malaria infection and immune model¶
To study aspects of malaria infections and population dynamics that do not fit into the framework of a simple SEIRS model with enhancements, EMOD contains a detailed microsolver using within-host parasite dynamics. In the microsolver, the development of clinical and parasitological immunity is tracked through innate and adaptive immune responses to specific antigens. A detailed parasite count over time is tracked for each infected individual. This permits the simulation of measured prevalence over time by slide microscopy, RDTs (Rapid Diagnostic Tests) or other diagnostics. Gametocyte production and decay is included to study the human infectious reservoir.
Immune variables, such as antibody levels for each antigen, inflammatory cytokine levels, and immunological memory are represented by continuous floating-point variables. Most model parameters have a specific physical interpretation and can be rationally constrained by defined experimental measurements.
For a complete list of configuration parameters that are used in the malaria model, see the Configuration parameters.
Within the microsolver framework, each new infection begins with a hepatic (liver-stage) latency of fixed duration (7 days) and proceeds through cycles of asexual replication (with a fixed duration of 2 days) and gametocyte production (which takes 10 days). The model accounts for several antigenic components to which the immune system may develop immunity: the merozoite surface protein (MSP) variant, the Plasmodium falciparum erythrocyte membrane protein 1 (PfEMP1) presented on the surface of the infected red-blood cell (IRBC), and less immunogenic minor surface epitopes (nonspecific epitopes).
Within-host dynamics¶
The model advances gametocytes through five stages of development that are characterized by different drug susceptibilities. A fraction of gametocytes die at each developmental stage, and gametocytes can be inactivated by cytokines after uptake in a blood meal. Gametocytes reach maturity after ten days and remain in the bloodstream with a 2.5-day half-life. EMOD does not model acquired immunity to gametocytes.
Gametocyte development¶
A single clonal infection is modeled with an antigenic repertoire of 50 unique PfEMP-1 variants where each is associated with one of five repeating minor epitopes. In the first asexual cycle, the first five variants are expressed in equal numbers. At each update, the immune system is stimulated by the Infected Red Blood Cell (IRBC) count of each antigenic variant, and concurrently, the IRBC counts are decremented on account of immune and drug killing effects.
At the end of each asexual cycle, the model calculates the fraction of merozoites (16 for each previous IRBC) killed by specific recognition of the MSP variant, and the fraction that is differentiated into male and female gametocytes. To capture the dynamics of the parasite’s immune evasion strategy, the model imposes a constant per-parasite switching rate on the remaining merozoites to advance to subsequent antigenic variants in the repertoire.
Parasite strain switching¶
The number of population-level variants affect the age-pattern of natural immunity acquisition. The full set of population-level antigenic variants (out of which a single infection’s repertoire is randomly drawn), consists of 100 MSP variants, 20 sets of five minor epitopes, and 1000 PfEMP-1 variants. Some parameters drive the asymptotic levels of adult detected parasitemia while others primarily govern the transition between child and adult detected prevalence rates, and provide that the number of PfEMP-1 variants is substantially more than an individual would experience in a year. A multi-year burn-in period ensures immune initialization dynamics approximate long-term asymptotic behavior where individuals in the simulation build antibody responses to a broad repertoire of parasite antigens appropriate for their age.
The immune response to infection is characterized by innate inflammatory and specific antibody components that limit maximum parasite density and work to clear infection. As the antibody response increases on a dimensionless scale from 0 to 1, the initial inflammatory response is suppressed. After the antibody response appears, continued antigenic stimulation will drive up the adapted response until that antigenic variant is cleared, at which point both the antibody levels and the capacity to respond will decay over time. The larger the antigenic population, the more infections, and thus time, it takes to acquire broad parasitological immunity.
Innate and adaptive immunity work together to limit asexual parasite numbers¶
The innate immune response is modeled to depend on a temporary contribution from a rupturing schizont at the end of each asexual cycle, as well as the concentration of IRBC surface antigens to which an antibody response has not yet been developed. The innate response that is suppressed by the presence of specific antibodies is responsible for driving febrile symptoms and broad-spectrum parasite suppression.
Innate immune response¶
The capacity to generate specific antibodies grows in response to the concentration of each novel antigen. Above a capacity threshold level, antibodies are produced in increasing concentration until the corresponding antigenic variant is cleared. At this time, the capacity will decay to a non-zero memory level. The mechanism by which the antibody capacity evolves captures the time delay of specific antibody response on re-infection.
Adaptive immune response to PfEMP1 and minor epitopes¶
Adaptive immune response to PfEMP1 and minor epitopesAdaptive immune response to MSP antigens¶
Adaptive immune response to CSP¶
Malaria symptoms and diagnostics¶
In EMOD, parameters for multiple types of malaria symptoms are included. Symptoms indicate the presence of disease, and one of the most prevalent types of data used to inform models is case data. By leveraging such data, models can help elucidate the processes that cause disease and enable calculations of the likely percent of the population that is infected (or infectious). Similarly, in order to model treatment-seeking behaviors (which will change the transmission dynamics of malaria), it is necessary to have a spectrum of disease severity (and hence, a spectrum of symptom parameters). Finally, disease severity is related to past infection and immune response, so inclusion of symptoms can help to further inform transmission dynamics of the target population.
For a complete list of configuration parameters that are used in the malaria model, see the Configuration parameters.
Fever is triggered by cytokine production through the innate immune response. A clinical case begins when fever surpasses a certain threshold and the clinical incident continues until fever subsides below another threshold.
Fever and clinical case definition¶
Rupture of infected red blood cells (RBCs) destroys nearby RBCs and leads to anemia. Erythropoiesis is stimulated in anemic individuals to increase the rate of RBC production.
Anemia¶
Antimalarial drugs¶
Antimalarial drugs are a powerful tool for malaria control and elimination. Modeling mass antimalarial campaigns can elucidate how to most effectively deploy drug-based interventions and quantitatively compare the effects of cure, prophylaxis, and transmission-blocking in suppressing parasite prevalence.
Individuals in a simulation can receive antimalarial drugs as treatment for clinical or severe malaria or during a mass drug administration (MDA) campaign. The following information explains how to configure antimalarial drug parameters in a configuration file. For a complete list of configuration parameters that are used in the malaria model, see the Configuration parameters.
To specify where, when, and to whom the antimalarial drugs are administered, you must create an AntiMalarialDrug intervention in the campaign file. See Campaign parameters for more information.
Pharmacokinetics (PK) of antimalarial drugs can be modeled with a box profile or with a double exponential decay. Pharmacodynamics (PD) of antimalarial drugs are modeled as an on/off switch for box PK or with a simple sigmoid for double exponential PK. If multiple drugs are given in a treatment, each drug is considered independently for both PK and PD.
To model a box PK, set the parameter PKPD_Model to FIXED_DURATION_CONSTANT_EFFECT. Drugs will have full killing efficacy for the duration set by Drug_Decay_T1. Dosing multiple times leads to multiple pulses of drug killing if dose separation is longer than Drug_Decay_T1. If dose separation (Drug_Dose_Interval) is shorter than Drug_Decay_T1, the box time is reset by each new dose such that drug killing is switched off after time Drug_Decay_T1 after the last dose.
Modeling drug killing effects with a box profile. (A) A single dose of drug kills parasites at maximum efficacy for duration set by Drug_Decay_T1. (B) A drug given in multiple doses results in multiple parasite killing pulses of duration Drug_Decay_T1 if the dose separation Drug_Dose_Interval is longer than Drug_Decay_T1. (C) If a drug’s dosing regimen results in doses taken before time Drug_Decay_T1 has elapsed, the drug kills parasites at maximum efficacy the entire time between the first dose and time Drug_Decay_T1 after the last dose.¶
The following example provides the syntax:
{
"Malaria_Drug_Params": {
"Artemether_Lumefantrine": {
"Drug_Cmax": 1000,
"Drug_Decay_T1": 1,
"Drug_Decay_T2": 1,
"Drug_Dose_Interval": 1,
"Drug_Fulltreatment_Doses": 3,
"Drug_Gametocyte02_Killrate": 2.3,
"Drug_Gametocyte34_Killrate": 2.3,
"Drug_GametocyteM_Killrate": 0,
"Drug_Hepatocyte_Killrate": 0,
"Drug_PKPD_C50": 100,
"Drug_Vd": 10,
"Max_Drug_IRBC_Kill": 4.61,
"Bodyweight_Exponent": 0
},
"Artemisinin": {},
"Chloroquine": {},
"GenPreerythrocytic": {},
"GenTransBlocking": {},
"Primaquine": {},
"Quinine": {},
"SP": {},
"Tafenoquine": {}
}
}
To model a double exponential PK, set the parameter PKPD_Model to CONCENTRATION_VERSUS_TIME. The double exponential PK approximates a one- or two-compartment PK model. To model a drug with 1-compartment PK, set Drug_Decay_T2 equal to Drug_Decay_T1 (Drug_Vd becomes irrelevant). For PD, the Drug_PKPD_C50 determines the drug concentrations where parasite killing is effective (drug concentration > C50) and ineffective (drug concentration < C50). EMOD sets the shape of the parasite killing curve based on parasite kill rate and C50. Drug_Cmax, Drug_Decay_T1, Drug_Decay_T2, Drug_Vd, and Drug_PKPD_C50 together determine when the drug is effectively killing parasites. Changing each of these parameters will affect how long the parasites are exposed to a strong killing effect. Adding additional doses will also increase the duration of the parasite- killing window.
Children can be dosed with a fraction of the adult dose according to Fractional_Dose_By_Upper_Age, where each dose fraction is paired with the maximum age of children receiving that dose. Depending on specific PK characteristics of each drug, bodyweight may affect the rate of drug clearance. In the double exponential PK model, bodyweight is directly determined by age, and Drug_Cmax is multiplied by the inverse of the bodyweight raised to the power of Bodyweight_Exponent. See the paper on drug PK and PD ( Mass campaigns with antimalarial drugs: a modelling comparison of artemether-lumefantrine and DHA-piperaquine with and without primaquine as tools for malaria control and elimination) for more details on how to set PK and PD parameter values in EMOD.
Modeling drug killing effects with a double exponential PK and sigmoid PD. (A) Two-compartment PK models include both distribution and elimination of drug. One-compartment models do not have a distribution phase. (B) A double exponential PK approximates the distribution and elimination phases of a two- compartment model. (C) Parasite killing is determined by drug concentration, Drug_PKPD_C50, and parasite stage- specific maximum kill rates. (D) PK and PD together determine the duration over which a dose of drug is effective.¶
Malaria research at IDM¶
Malaria research at IDM is progressive and dynamic. For the current research, research team profiles, and a full list of published papers from IDM’s malaria group, see the IDM Malaria Research Website.
IDM malaria publications¶
Information specific to the components of the model described previously can be found in the following list of select IDM malaria publications.
Vector model¶
Eckhoff, 2013. Mathematical models of within-host and transmission dynamics to determine effects of malaria interventions in a variety of transmission settings. Am J Trop Med Hyg. 88(5):817-27
Eckhoff, 2011. A malaria transmission-directed model of mosquito life cycle and ecology. Malaria Journal. 10:303
Infection and immunity: Within-host parasite dynamics¶
Lawniczak & Eckhoff, 2016. A computational lens for sexual-stage transmission, reproduction, fitness and kinetics in Plasmodium falciparum. Malaria Journal. 15:487
Cameron, et al., 2015. Defining the relationship between infection prevalence and clinical incidence of Plasmodium falciparum malaria. Nature Communications. 6:8170
Ouedraogo, et al., 2015. Dynamics of the Human Infectious Reservoir for Malaria Determined by Mosquito Feeding Assays and Ultrasensitive Malaria Diagnosis in Burkina Faso. Journal of Infectious Diseases. 213(1):90-99
Gerardin, Ouedraogo, McCarthy, Eckhoff and Wenger, 2015. Characterization of the infectious reservoir of malaria with an agent-based model calibrated to age-stratified parasite densities and infectiousness. Malaria Journal. 14:231
Eckhoff, 2012. Malaria parasite diversity and transmission intensity affect development of parasitological immunity in a mathematical model. Malaria Journal. 11:419
Eckhoff, 2012. P. falciparum Infection Durations and Infectiousness Are Shaped by Antigenic Variation and Innate and Adaptive Host Immunity in a Mathematical Model. PLOS one. 7(9)
Outcrossing of parasite genetics¶
Daniels, et al., 2015. Modeling malaria genomics reveals transmission decline and rebound in Senegal. PNAS. 112(22):7067-7072
Antimalarial drugs¶
Bellinger, et al., 2016. Oral, ultra–long-lasting drug delivery: Application toward malaria elimination goals. Science Translational Medicine. 8(365).
Gerardin, Eckhoff and Wenger, 2015. Mass campaigns with antimalarial drugs: a modeling comparison of artemether-lumefantrine and DHA-piperaquine with and without primaquine as tools for malaria control and elimination. BMC Infectious Diseases. 15:144
Vaccines¶
Penny, et al., 2016. Public health impact and cost-effectiveness of the RTS,S/AS01 malaria vaccine: a systematic comparison of predictions from four mathematical models. The Lancet. 387(100016): 367-375
McCarthy, Wenger, Huynh and Eckhoff, 2015. Calibration of an intrahost malaria model and parameter ensemble evaluation of a pre-erythrocytic vaccine. Malaria Journal. 14:6
Wenger & Eckhoff, 2013. A mathematical model of the impact of present and future malaria vaccines. Malaria Journal. 12:126
Spatial models¶
Nikolov, et al., 2016. Malaria Elimination Campaigns in the Lake Kariba Region of Zambia: a Spatial Dynamical Model. PLOS Computational Biology.
Marshall, et al., 2016. Key traveler groups of relevance to spatial malaria transmission: a survey of movement patterns in four sub-Saharan African countries. Malaria Journal. 15(200)
Gerardin, et al., 2016. Optimal Population-Level Infection Detection Strategies for Malaria Control and Elimination in a Spatial Model of Malaria Transmission. PLOS Computational Biology.
Eckhoff, et al., 2015. From puddles to planet: modeling approaches to vector-borne diseases at varying resolution and scale. Current Opinion in Insect Science. 10:118-123
Interventions¶
Eckhoff, Wenger, Godfray and Burt, 2016. Impact of mosquito gene drive on malaria elimination in a computational model with explicit spatial and temporal dynamics. PNAS. 114(2)
Bhatt, et al., 2015. The effect of malaria control on Plasmodium falciparum in Africa between 2000 and 2015. Nature. 526:207-211
EMOD parameter reference¶
The configurable parameters described in this reference section determine the behavior of a particular simulation and interventions. The parameter descriptions are intended to guide you in correctly configuring simulations to match your disease scenario of interest.
The reference section contains only the parameters supported for use with the malaria simulation type. To see the parameters for a different simulation type, see the corresponding documentation.
These parameters include three major groups: parameters for the demographics file, parameters for the configuration file, and parameters for the campaign file.
If you would rather use a text file for parameter reference than this web documentation, you can also generate a schema with the EMOD executable (Eradication.exe) that defines all configuration and campaign parameters that can be used in a simulation, including names, data types, defaults, ranges, and short descriptions. For instructions, see Generate a list of all available parameters (a schema).
JSON formatting overview¶
All of these parameters are contained in JavaScript Object Notation (JSON) formatted files. JSON is a data format that is human-readable and easy for software to read and generate. It can be created and edited using a simple text editor.
JSON has two basic structures: a collection of key-value pairs or an ordered list of values (an array). These structures are within a JSON object, which is enclosed in a set of braces. You will notice that each of these files begins with a left brace ({) and ends with a right brace (}).
A value can be a string, number, Boolean, array, or an object. The campaign and data input files often use nested objects and arrays. See www.json.org for more information on JSON.
A few important details to call out when creating JSON files:
Add a comma after every key-value pair or array except for the last key-value pair in an array or object.
The keys (parameters) are case-sensitive. For example, “NodeID” is not the same as “NodeId”.
Decimals require a 0 (zero) before the decimal point.
EMOD does not support Booleans (“True”, “False”). Instead, EMOD use the integer 1 for “True” and 0 for “False”.
Every opening brace or bracket ({ or [) requires a corresponding closing brace or bracket (} or ]).
The following is an example of a JSON formatted file.
{
"A_Complex_Key": {
"An_Array_with_a_Nested_Object_Value": [
{
"A_Simple_Key": "Value",
"A_Simple_Array": [ "Value1", "Value2" ],
"An_Array_with_Number_Values": [ 0.1, 0.2 ],
"A_Nested_Object": {
"Another_Simple_Key": "Value",
"Nested_Arrays": [
[ 10, 0.1 ],
[ 0.1, 1 ]
]
}
}
]
}
}
Demographics file structure and parameters¶
The parameters described in this reference section can be added to the JavaScript Object Notation (JSON) formatted demographics file to determine the demographics of the population within each geographic node in a simulation. For example, the number of individuals and the distribution for age, gender, immunity, risk, and mortality. This is in contrast to the configuration parameters related to demographics that are simulation-wide and generally control whether certain events, such as births or deaths, are enabled in a simulation.
At least one demographics file is required for every simulation unless you set the parameter Enable_Demographics_Builtin to 0 (zero) in the configuration file. This is generally only used for testing and validating code pathways by using a standard testing sandbox instead of actual demographics information. If Migration_Model is set to “FIXED_RATE_MIGRATION”, built-in demographics will use a form of local migration where individuals migrate only to adjacent nodes.
Demographics files are usually the only input data file you will modify. Demographics files are named using the name of the region, appended with “_demographics.json”.
Additionally, you can use more than one demographics file, with one serving as the base layer and the one or more others acting as overlays that override the values in the base layer.
Contents
Demographic file structure¶
Demographics files are organized into three sections:
- Metadata
Information such as data source, created date or region name. Metadata is used to provide provenance information.
- Defaults
Default parameter values applied to all nodes.
- Nodes
Parameter values specific to individual geographic nodes, which override the default values. There is one entry for each node in the simulation.
The following example shows the skeletal format of a demographics file.
{
"Metadata": {
"DateCreated": "dateTime",
"Tool": "scriptUsedToGenerate",
"Author": "author",
"IdReference": "ID reference",
"NodeCount": 2
},
"Defaults": {
"NodeAttributes": {},
"IndividualAttributes": {},
"IndividualProperties": {}
},
"Nodes": [{
"NodeID": 1,
"NodeAttributes": {},
"IndividualAttributes": {},
"IndividualProperties": {}
}, {
"NodeID": 2,
"NodeAttributes": {},
"IndividualAttributes": {},
"IndividualProperties": {}
}]
}
Most of the items in the Metadata section informational only and are not used by EMOD. This includes information such as author, date created, and tool used to create the file. However, the following two values are used by EMOD:
- NodeCount
The number of nodes to expect in the demographics file.
- IdReference
The unique string that indicates the method used for generating the NodeID, the identifier used for each node in the simulation.
See Input data file structure for more information about IdReference and NodeID generation.
Parameter values in the Defaults section is applied to all nodes. However, any node-level information provided in the Nodes section will take precedence for that node. The Defaults section is optional and can be completely omitted, as is often the case of single- node simulations. The Defaults section can contain any of the parameters listed below.
Parameter values in the Nodes section is applied to specific nodes. The intent of the Nodes section is to have only the information that is unique to each node, such as identifier and location (longitude and latitude). The section contains an array with each element in the array representing a node that is identified by its NodeID. The NodeID is a unique integer value. If a parameter appears in both the Defaults and Nodes sections, the value in the Nodes section will take precedence.
Parameters¶
The demographic parameters are divided into the following broad categories. The parameters will be contained in a nested JSON object of the same name.
Warning
Parameters are case-sensitive. For Boolean parameters, set to 1 for true or 0 for false.
NodeAttributes are a JSON object that contains parameters that add or modify information regarding the location, migration, habitat, and population of a simulation. Some NodeAttributes depend on values set in the configuration parameters.
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
AbovePoverty |
float |
0 |
1 |
1 |
The proportion of the population that is living above the poverty line. |
{
"AbovePoverty": 0.5
}
|
Airport |
bool |
0 |
1 |
0 |
Set to 1 to indicate that the node is connected by air migration. |
{
"Airport": 0
}
|
Altitude |
float |
-3.40282e+038 |
3.40E+38 |
0 |
Required. Only used when Climate_Model is set to “CLIMATE_KOPPEN”. The unit of measurement is in meters. |
{
"Altitude": 250
}
|
BirthRate |
float |
0 |
1 |
0.25 |
The birth rate. In the configuration file, Enable_Birth must be set to 1 and Birth_Rate_Dependence will affect how this rate is used. |
{
"BirthRate": 0.0001
}
|
InitialPopulation |
integer |
0 |
2.15E+09 |
The number of people that will be populated into the node at the beginning of the simulation. You can scale this number using Base_Population_Scale_Factor in the configuration file. |
{
"InitialPopulation": 1000
}
|
|
LarvalHabitatMultiplier |
float or nested JSON object |
The value by which to scale the larval habitat availability specified in the configuration file with Larval_Habitat_Types across all habitat types, for specific habitat types, or for specific mosquito species within each habitat type. |
The following example scales the larval habitat equally across all habitat types and mosquito species. {
"LarvalHabitatMultiplier": 2.0
}
The following example scales the larval habitat only in the temporary rainfall habitat for all mosquito species. {
"LarvalHabitatMultiplier": {
"TEMPORARY_RAINFALL": 2.0
}
}
The following example scales the larval habitat independently for A. gambiae in the temporary rainfall habitat and for A. arabiensis in the brackish swamp habitat. {
"LarvalHabitatMultiplier": {
"TEMPORARY_RAINFALL": {
"gambiae": 2.0
},
"BRACKISH_SWAMP": {
"arabiensis": 2.5
}
}
}
|
|||
Latitude |
float |
-3.40282e+038 |
3.40E+38 |
Latitude of the node in decimal degrees. This can be used for several things including determining infectiousness by latitude and defining the size of grid cells. |
{
"Latitude": 12.4
}
|
|
Longitude |
float |
-3.40282e+038 |
3.40E+38 |
Longitude of the node in decimal degrees. This can be used for several things including defining the size of grid cells. |
{
"Longitude": 9.1
}
|
|
Region |
bool |
0 |
1 |
0 |
Set to 1 to indicate that the node is connected by regional migration. |
{
"Region": 1
}
|
Seaport |
bool |
0 |
1 |
0 |
Set to 1 to indicate that the node is connected by sea migration. |
{
"Seaport": 1
}
|
Urban |
bool |
0 |
1 |
0 |
Set to 1 to enable urban settings. In the configuration file, Enable_Demographics_Other must be set to 1 and Birth_Rate_Dependence must be set to “INDIVIDUAL_PREGNANCIES_BY_URBAN_AND_AGE”. |
{
"Urban": 1
}
|
Zoonosis |
float |
0 |
1 |
0 |
The daily rate of zoonotic infection per individual. If Animal_Reservoir_Type is set to “ZOONOSIS_FROM_DEMOGRAPHICS”, the value for Zoonosis overrides the value set for Zoonosis_Rate in the configuration file. |
{
"Zoonosis": 0
}
|
IndividualAttributes are a JSON object that contains parameters that add or modify the distribution of attributes across individuals in a simulation. For example, the age or immunity distribution. An initial value for an individual is a randomly selected value from a distribution. For example, if you use a uniform distribution to initialize age, the initial ages of individuals in the simulation will be evenly distributed between some minimum and maximum value.
There are two types of distributions: simple and complex. Simple distributions are defined by three parameters where one is the distribution type and the other two are used to further define the distribution. You can use complex distributions to define a distribution that does not fit some standard. For example, a complex distribution is useful when trying to represent real-world data.
The table below shows the simple distributions and their associated parameters.
Distribution |
Demographic parameters |
Related configuration parameters |
|---|---|---|
Age |
|
Age_Initialization_Distribution_Type must be set to “DISTRIBUTION_SIMPLE”. |
Prevalence |
|
Enable_Demographics_Other must be set to 1. |
Immunity |
|
|
Risk |
|
Enable_Demographics_Other must be set to 1. |
Migration heterogeneity |
|
Enable_Migration_Heterogeneity must be set to 1. |
The table below shows the available probability distribution types to use with simple distributions and how to set the other two demographic parameters to initialize the distribution.
Distribution type |
Flag value |
Distribution1 value |
Distribution2 value |
|---|---|---|---|
Constant |
0 |
Value |
N/A (set to 0) |
Uniform |
1 |
Minimum |
Maximum |
Gaussian |
2 |
Mean |
Standard deviation |
Exponential |
3 |
Mean |
N/A (set to 0) |
Poisson |
4 |
Mean |
N/A (set to 0) |
Log normal |
5 |
Mean |
Log (standard deviation) |
Bimodal |
6 |
Proportion of time to return the value specified by Distribution2. Value must be between 0 and 1. |
A positive value to be returned. If the value is not positive, a value of 1 is returned. |
Weibull |
7 |
Scale parameter |
Shape parameter |
EMOD also supports the following complex distributions:
AgeDistribution
FertilityDistribution
MortalityDistribution
MortalityDistributionMale
MortalityDistributionFemale
HIVCoinfectionDistribution
HIVMortalityDistribution
MSP_mean_antibody_distribution
nonspec_mean_antibody_distribution
PfEMP1_mean_antibody_distribution
MSP_variance_antibody_distribution
nonspec_variance_antibody_distribution Malaria
PfEMP1_variance_antibody_distribution
IndividualProperties are a JSON array that contains parameters that add properties to individuals in a simulation as a means of setting up groups. For example, you can define values for accessibility, age, geography, risk, and other properties and assign individuals to different groups based on those property values.
Parameter |
Data type |
Description |
Example |
|---|---|---|---|
Property |
string |
The property type for which you will assign values to create groups. Accepted values are:
|
{
"IndividualProperties": [{
"Property": "Risk"
}]
}
|
Values |
array |
An array of the user-defined values that can be assigned to individuals for this property. You can have up to 125 values for the Geographic property type and up to 5 values for all other types. For “Age_Bin” property types, omit and use Age_Bin_In_Years instead. |
{
"IndividualProperties": [{
"Values": ["Low", "Medium", "High"]
}]
}
|
Initial_Distribution |
array |
An array of numbers that define the proportion of individuals to assign to each group at the beginning of the simulation. Their total must equal 1 and the number of members in this array must match the number of members in Values. For “Age_Bin” property types, omit this parameter as the demographics file controls the age distribution. |
{
"IndividualProperties": [{
"Initial_Distribution": [0.2, 0.4, 0.4]
}]
}
|
Age_Bin_In_Years |
array |
An array of integers that represents the ages, in years, at which to demarcate the age groups. Used only with the “Age_Bin” property type. The first number must be 0, the last must be -1, and they must be listed in ascending order. |
The following example creates three age groups: 0 to 5, older than 5 to 13, and older than 13 to the maximum age. {
"IndividualProperties": [{
"Age_Bin_In_Years": [0, 5, 13, -1]
}]
}
|
Transitions |
array |
An array contains multiple JSON objects that each define how an individual transitions from one property group to another. See the following table for information about the parameters to include in the object. For “Age_Bin” property types, set to an empty array, as individuals will transition to the next age bin based on the passing of time. |
{
"IndividualProperties": [{
"Transitions": []
}]
}
|
TransmissionMatrix |
array |
An object that contains Route and Matrix parameters that define how to scale the base infectivity from individuals in one group to individuals in another. See Configure heterogeneous disease transmission for more information. |
{
"IndividualProperties": [{
"TransmissionMatrix": {
"Route": "Contact",
"Matrix": [
[10, 0.1],
[0.1, 1]
]
}
}]
}
|
The following table contains the parameters that are available to use in the Transitions array.
Parameter |
Data type |
Description |
Example |
|---|---|---|---|
From |
string |
Which group an individual will transition from. |
{
"Transitions": [{
"From": "Low"
}]
}
|
To |
string |
Which group an individual will transition to. |
{
"Transitions": [{
"To": "High"
}]
}
|
Type |
string |
The type of condition that starts transitioning individuals. Set to “At_Age” or “At_Timestep”. The other parameters you must set depend on the condition type set here. |
{
"Transitions": [{
"Type": "At_Age"
}]
}
|
Age_In_Years_Restriction |
nested JSON object |
The age when an individual is eligible to transition. Min is optional and Max is required. Do not use this with “At_Age” types because it will conflict with other age parameters in Age_In_Years. This is required for “At_Timestep” types, though it can be empty for no age restriction. |
{
"Transitions": [{
"Type": "At_Timestep",
"Age_In_Years_Restriction": {
"Min": 5,
"Max": 40
}
}]
}
|
Timestep_Restriction |
nested JSON object |
The time step when transitioning starts and stops. Required for both “At_Age” and “At_Timestep” types. Start is required and End is optional. |
{
"Transitions": [{
"Timestep_Restriction": {
"Start": 20
}
}]
}
|
Coverage |
float |
A value between 0 and 1 that indicates the proportion of the population that EMOD will attempt to transition. The calculation used with Probability_Per_Timeset only uses the proportion of the population specified in Cover to calculate the probability of an individual transitioning at a given time step. Coverage has no effect when the value is set to 1. Required for both “At_Age” and “At_Timestep” types. |
{
"Transitions": [{
"Coverage": 0.5
}]
}
|
Probability_Per_Timestep |
float |
A value between 0 and 1 that is used to calculate the probability of an individual transitioning at the current time step given the number of individuals remaining in the group. Required for both “At_Age” and “At_Timestep” types. |
{
"Transitions": [{
"Probability_Per_Timestep": 0.25
}]
}
|
Age_In_Years |
integer |
The age at which individuals are eligible to transition. Do not use for “At_Timestep” types. Required for “At_Age” types. |
{
"Transitions": [{
"Type": "At_Age",
"Age_In_Years": 10
}]
}
|
Timesteps_Until_Reversion |
integer |
The number of time steps after the start of transitioning when individuals revert back to their original group. The start of transitioning is specified by the Start parameter in Timestep_Restriction. |
{
"Transitions": [{
"Timestep_Restriction": {
"Start": 20
},
"Timesteps_Until_Reversion": 100
}]
}
|
Example¶
The example below shows a complete multi-node demographics file that uses a gridded IdReference type.
{
"Metadata": {
"DateCreated": "03/03/2016",
"Tool": "DemographicsWorker",
"Author": "Institute for Disease Modeling",
"NodeType": "Grid",
"ProjectName": "IDM-Madagascar",
"RegionName": "Madagascar",
"IdReference": "Gridded world grump2.5arcmin",
"NodeCount": 15,
"DataProvenance": "Generated by internally-developed demographic tools",
"Resolution": 150,
"UpperLatitude": -10.6666666666667,
"LeftLongitude": 41.6666666666667,
"BottomLatitude": -27.3333333333333,
"RightLongitude": 51.3333333333333
},
"Defaults": {
"NodeAttributes": {
"Altitude": 0,
"Airport": 0,
"Region": 1,
"Seaport": 0,
"BirthRate": 0.000102,
"AbovePoverty": 0
},
"IndividualAttributes": {
"AgeDistributionFlag": 3,
"AgeDistribution1": 0.00015,
"AgeDistribution2": 0,
"PrevalenceDistributionFlag": 1,
"PrevalenceDistribution1": 0.1,
"PrevalenceDistribution2": 0.2,
"ImmunityDistributionFlag": 0,
"ImmunityDistribution1": 1,
"ImmunityDistribution2": 0,
"RiskDistributionFlag": 0,
"RiskDistribution1": 1,
"RiskDistribution2": 0,
"MigrationHeterogeneityDistributionFlag": 0,
"MigrationHeterogeneityDistribution1": 1,
"MigrationHeterogeneityDistribution2": 0,
"MortalityDistribution": {
"NumDistributionAxes": 2,
"AxisNames": ["gender", "age"],
"AxisUnits": ["male=0,female=1", "years"],
"AxisScaleFactors": [1, 365],
"NumPopulationGroups": [2, 3],
"PopulationGroups": [
[0, 1],
[0, 100, 2000]
],
"ResultUnits": "annual deaths per 1000 individuals",
"ResultScaleFactor": 0.00000273972602739726027397260273973,
"ResultValues": [
[0, 20.0000035, 400.00007],
[0, 20.0000035, 400.00007]
]
}
}
},
"Nodes": [{
"NodeID": 358876977,
"NodeAttributes": {
"Latitude": -13.3125,
"Longitude": 48.1875,
"Altitude": 34,
"Area": 25,
"InitialPopulation": 2088,
"BirthRate": 0.000113117808219178,
"AbovePoverty": 0,
"Urban": 0
}
}, {
"NodeID": 358876978,
"NodeAttributes": {
"Latitude": -13.2708330154419,
"Longitude": 48.1875,
"Altitude": 19,
"Area": 25,
"InitialPopulation": 1682,
"BirthRate": 0.000113117808219178,
"AbovePoverty": 0,
"Urban": 0
}
}, {
"NodeID": 358942512,
"NodeAttributes": {
"Latitude": -13.3541669845581,
"Longitude": 48.2291679382324,
"Altitude": 127,
"Area": 25,
"InitialPopulation": 5160,
"BirthRate": 0.000113117808219178,
"AbovePoverty": 0,
"Urban": 0
}
}, {
"NodeID": 358942513,
"NodeAttributes": {
"Latitude": -13.3125,
"Longitude": 48.2291679382324,
"Altitude": 128,
"Area": 25,
"InitialPopulation": 2860,
"BirthRate": 0.000113117808219178,
"AbovePoverty": 0,
"Urban": 0
}
}, {
"NodeID": 358942514,
"NodeAttributes": {
"Latitude": -13.2708330154419,
"Longitude": 48.2291679382324,
"Altitude": 26,
"Area": 25,
"InitialPopulation": 790,
"BirthRate": 0.000113117808219178,
"AbovePoverty": 0,
"Urban": 0
}
}, {
"NodeID": 359008047,
"NodeAttributes": {
"Latitude": -13.3958330154419,
"Longitude": 48.2708320617676,
"Altitude": 48,
"Area": 25,
"InitialPopulation": 6436,
"BirthRate": 0.000113117808219178,
"AbovePoverty": 0,
"Urban": 0
}
}, {
"NodeID": 359008048,
"NodeAttributes": {
"Latitude": -13.3541669845581,
"Longitude": 48.2708320617676,
"Altitude": 111,
"Area": 25,
"InitialPopulation": 4983,
"BirthRate": 0.000113117808219178,
"AbovePoverty": 0,
"Urban": 0
}
}, {
"NodeID": 359008049,
"NodeAttributes": {
"Latitude": -13.3125,
"Longitude": 48.2708320617676,
"Altitude": 124,
"Area": 25,
"InitialPopulation": 1245,
"BirthRate": 0.000113117808219178,
"AbovePoverty": 0,
"Urban": 0
}
}, {
"NodeID": 359008050,
"NodeAttributes": {
"Latitude": -13.2708330154419,
"Longitude": 48.2708320617676,
"Altitude": 57,
"Area": 25,
"InitialPopulation": 1387,
"BirthRate": 0.000113117808219178,
"AbovePoverty": 0,
"Urban": 0
}
}, {
"NodeID": 359008051,
"NodeAttributes": {
"Latitude": -13.2291669845581,
"Longitude": 48.2708320617676,
"Altitude": 66,
"Area": 25,
"InitialPopulation": 1516,
"BirthRate": 0.000113117808219178,
"AbovePoverty": 0,
"Urban": 0
}
}, {
"NodeID": 359073583,
"NodeAttributes": {
"Latitude": -13.3958330154419,
"Longitude": 48.3125,
"Altitude": 132,
"Area": 25,
"InitialPopulation": 5957,
"BirthRate": 0.000113117808219178,
"AbovePoverty": 0,
"Urban": 0
}
}, {
"NodeID": 359073584,
"NodeAttributes": {
"Latitude": -13.3541669845581,
"Longitude": 48.3125,
"Altitude": 44,
"Area": 25,
"InitialPopulation": 643,
"BirthRate": 0.000113117808219178,
"AbovePoverty": 0,
"Urban": 0
}
}, {
"NodeID": 359073585,
"NodeAttributes": {
"Latitude": -13.3125,
"Longitude": 48.3125,
"Altitude": 55,
"Area": 25,
"InitialPopulation": 1171,
"BirthRate": 0.000113117808219178,
"AbovePoverty": 0,
"Urban": 0
}
}, {
"NodeID": 359073587,
"NodeAttributes": {
"Latitude": -13.2291669845581,
"Longitude": 48.3125,
"Altitude": 36,
"Area": 25,
"InitialPopulation": 378,
"BirthRate": 0.000113117808219178,
"AbovePoverty": 0,
"Urban": 0
}
}, {
"NodeID": 359139117,
"NodeAttributes": {
"Latitude": -13.4791669845581,
"Longitude": 48.3541679382324,
"Altitude": 282,
"Area": 25,
"InitialPopulation": 674,
"BirthRate": 0.000113117808219178,
"AbovePoverty": 0,
"Urban": 0
}
}]
}
Configuration parameters¶
The parameters described in this reference section determine the behavior of a simulation. Simulations are configured in a configuration file. This file is a JavaScript Object Notation (JSON) formatted file that contains mostly a flat list of JSON key-value pairs. For information on JSON, see EMOD parameter reference.
Warning
Parameters are case-sensitive. For Boolean parameters, set to 1 for true or 0 for false.
The tables below contain only parameters available when using the malaria simulation type.
Contents
Cluster options¶
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Load_Balance_Filename |
string |
NA |
NA |
UNINITIALIZED STRING |
Path to input file used when a static load balancing scheme is selected. |
|
Num_Cores |
integer |
NA |
NA |
NA |
Number of cores used to run a simulation. This is used by the infrastructure that runs the job, such as an HPC cluster or the Regression scripts, and is not used by the DTK. |
Demographics¶
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Age_Initialization_Distribution_Type |
enum |
NA |
NA |
DISTRIBUTION_OFF |
Method for initializing the age distribution in the simulated population. Possible values are: DISTRIBUTION_OFF - All individuals default to age 20-years-old. DISTRIBUTION_SIMPLE - Individual ages are drawn from a distribution whose functional form and parameters are specified in the IndividualAttributes demographics input layers: AgeDistributionFlag, AgeDistribution1, and AgeDistribution2. DISTRIBUTION_COMPLEX - Individual ages are drawn from a piecewise linear distribution specified in the IndividualAttributes demographics input layer AgeDistribution. |
|
Base_Population_Scale_Factor |
float |
0 |
3.40E+38 |
1 |
If Population_Scale_Type is set to FIXED_SCALING, the initial simulation population is uniformly scaled over the entire area to adjust for historical or future population density. |
|
Birth_Rate_Dependence |
enum |
NA |
NA |
FIXED_BIRTH_RATE |
This parameter determines how the birth rate specified in the demographics file is used. Possible values are: FIXED_BIRTH_RATE- The absolute rate at which new individuals are born. POPULATION_DEP_RATE - Scales the node population to determine the birth rate. DEMOGRAPHIC_DEP_RATE - Scales the female population within fertility age ranges to determine the birth rate. INDIVIDUAL_PREGNANCIES - Results in a similar birth rate as DEMOGRAPHIC_DEP_RATE, but pregnancies are assigned on an individual basis and result in a 40-week pregnancy for a specific individual with a birth at the end, INDIVIDUAL_PREGNANCIES_BY_URBAN_AND_AGE, and INDIVIDUAL_PREGNANCIES_BY_AGE_AND_YEAR. |
|
Death_Rate_Dependence |
enum |
NA |
NA |
NONDISEASE_MORTALITY_OFF |
This parameter determines how the death rate specified in the demographics file is used. Possible values are: NONDISEASE_MORTALITY_OFF, NONDISEASE_MORTALITY_BY_AGE_AND_GENDER and NONDISEASE_MORTALITY_BY_YEAR_AND_AGE_FOR_EACH_GENDER. |
|
Enable_Aging |
boolean |
0 |
1 |
1 |
Set to 1 to account for aging in the population. Set to 0 to assume that the population does not age. |
|
Enable_Birth |
boolean |
0 |
1 |
1 |
Set to 1 to enable individuals to be added to the simulation by birth. |
|
Enable_Demographics_Birth |
boolean |
0 |
1 |
0 |
Assumes that newborn characteristics such as sickle-cell status are heterogeneous. Set to 0 to assume by default that all newborns have identical characteristics. |
|
Enable_Demographics_Builtin |
boolean |
0 |
1 |
0 |
Uses demographics input files to configure the initial population. Enter 0 to use a standard testing sandbox which is useful for testing and validating code pathways. |
|
Enable_Demographics_Gender |
boolean |
0 |
1 |
1 |
Enable_Demographics_Gender=0 assigns gender based on a 50/50 draw, while Enable_Demographics_Gender=1 draws from a male/female ratio that is randomly smeared by a Gaussian of width 1%. |
|
Enable_Demographics_Other |
boolean |
0 |
1 |
0 |
Includes the impact of other relevant demographic factors. For example, fraction of individuals above poverty, urban/rural characteristics, heterogeneous initial immunity and risk. Set to 0 to run the simulation without other relevant demographic factors. |
|
Enable_Disease_Mortality |
boolean |
0 |
1 |
1 |
Set to 1 to enable disease mortality. Set to 0 to disable disease mortality. |
|
Enable_Heterogeneous_Intranode_Transmission |
boolean |
0 |
1 |
0 |
Set to 1 to enable heterogeneous intra-node disease transmission. Requires individual property definitions and beta matrix to be specified. Set to 0 to disable. |
|
Enable_Vital_Dynamics |
boolean |
0 |
1 |
1 |
Set to 1 as a master switch to enable vital dynamics (births and deaths). Set to 0 to disable vital dynamics. Even when set to 1, the individual toggles for births and deaths can still disable these. |
|
Population_Scale_Type |
enum |
NA |
NA |
USE_INPUT_FILE |
Either use the initial population specified in the demographics input file or a fixed scaling of this value based on the population scaling factor parameter. Possible values are: USE_INPUT_FILE and FIXED_SCALING. |
Event recorder report settings¶
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Report_Event_Recorder |
boolean |
0 |
1 |
0 |
Enables or disables the ReportEventRecorder.csv report. |
|
Report_Event_Recorder_Ignore_Events_In_List |
boolean |
0 |
1 |
0 |
f this flag is set to true, all events listed in Report_Event_Recorder_Events will be ignored. If it is set to false, all events will be recorded. |
|
Report_Event_Recorder_Individual_Properties |
Dynamic String Set |
NA |
NA |
NA |
Fraction of individuals in the target demographic that will receive this intervention. |
General disease¶
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Acquisition_Blocking_Immunity_Duration_Before_Decay |
float |
0 |
45000 |
0 |
Number of days after infection until acquisition-blocking immunity begins to decay. Relevant only when Enable_Immunity and Enable_Immune_Decay parameters are set to 1. |
|
Animal_Reservoir_Type |
enum |
NA |
NA |
NO_ZOONOSIS |
Configures whether there is an animal reservoir and how the risk of zoonosis in configured. Use of the animal reservoir sets a low constant baseline of infectivity beyond what is present in the human population. It allows a more random introduction of cases in continuous time, which is more applicable for various situations such as zoonosis. Possible values are: NO_ZOONOSIS, CONSTANT_ZOONOSIS where the daily rate of zoonotic infection is configured by the parameter Zoonosis_Rate, and ZOONOSIS_FROM_DEMOGRAPHICS where the zoonosis rate is additionally scaled by the node-specific Zoonosis value in the NodeAttributes section of a demographics overlay file. See Zoonosis_Rate. |
|
Base_Incubation_Period |
float |
0 |
3.40E+38 |
6 |
Average duration, in days, of the incubation period before infected individuals becomes infectious. |
|
Base_Infectious_Period |
float |
0 |
3.40E+38 |
6 |
Average duration, in days, of the infectious period before the infection is cleared. |
|
Base_Infectivity |
float |
0 |
1000 |
0.3 |
The Base_Infectivity parameter determines the base infectiousness of individuals before accounting for transmission-blocking effects of acquired immunity and/or campaign interventions. It has a slightly different meaning for each of the disease types. GENERIC_SIM: The average number of individuals per time step who will be exposed to infection by one infectious individual. In the case of super-infection, the infectiousness is summed over all infections. VECTOR_SIM: The probability of infecting a mosquito during a successful blood meal (modulated by the vector parameter Acquire_Modifier). For vector simulations, the sum infectiousness of an individual is not allowed to exceed 100%. MALARIA_SIM: This simple scale factor is not used. Instead, gametocyte abundances and cytokine-mediated infectiousness are modeled explicitly. |
|
Base_Mortality |
float |
0 |
1000 |
0.001 |
The base mortality of the infection before accounting for individual immune modification factors. Depending on the setting of Mortality_Time_Course, this is either the daily probability of the disease being fatal (DAILY_MORTALITY) or the probability of death at the end of the infection duration (MORTALITY_AFTER_INFECTIOUS). |
|
Enable_Immune_Decay |
boolean |
0 |
1 |
1 |
Set to 1 if immunity decays after an infection clears. Set to 0 if recovery from the disease confers complete immunity for life. |
|
Enable_Immunity |
boolean |
0 |
1 |
1 |
Set to 1 if an individual has protective immunity after an infection clears. Set to 0 if recovery from the infection does not affect the individual’s future immunity. |
|
Enable_Maternal_Transmission |
boolean |
0 |
1 |
0 |
Set to 1 to account for infection of infants at birth by infected mothers. Set to 0 to assume that infection is not transmitted by mothers to infants at birth. |
|
Enable_Superinfection |
boolean |
0 |
1 |
0 |
Set to 1 if an individual can have multiple infections with the same agent simultaneously. Set to 0 if multiple infections are not possible. See the Max_Individual_Infections parameter. |
|
Immunity_Acquisition_Factor |
float |
0 |
1000 |
0 |
Multiplicative reduction in probability of reacquiring the disease. Relevant only when Enable_Immunity and Enable_Immune_Decay parameters are set to 1. |
|
Immunity_Initialization_Distribution_Type |
enum |
NA |
NA |
DISTRIBUTION_OFF |
Method for initializing the immunity distribution in the simulated population. Possible values are: DISTRIBUTION_OFF - All individuals default to no immunity. DISTRIBUTION_SIMPLE - Individual immunities are drawn from a distribution whose functional form and parameters are specified in the IndividualAttributes demographics input layers: ImmunityDistributionFlag, ImmunityDistribution1, and ImmunityDistribution2. DISTRIBUTION_COMPLEX - Individual immnunities are drawn from an age-dependent piecewise l. |
|
Immunity_Mortality_Factor |
float |
0 |
1000 |
0 |
Multiplicative reduction in probability of dying from infection after getting re-infected. Relevant only when Enable_Immunity and Enable_Immune_Decay parameters are set to 1. |
|
Immunity_Transmission_Factor |
float |
0 |
1000 |
0 |
Multiplicative reduction in probability of transmitting infection after getting re-infected. Relevant only when Enable_Immunity and Enable_Immune_Decay parameters are set to 1. |
|
Incubation_Period_Distribution |
enum |
NA |
NA |
NOT_INITIALIZED |
Distribution of duration of incubation period. Can be fixed or exponentially distributed with an average duration from Base_Incubation_Period. Possible values are: FIXED_DURATION and EXPONENTIAL_DURATION. |
|
Incubation_Period_Max |
float |
0.6 |
3.40E+38 |
0 |
The maximum length of the incubation period. Used when Incubation_Period_Distribution is UNIFORM_DURATION. |
|
Incubation_Period_Mean |
float |
0 |
3.40E+38 |
6 |
The standard deviation used when Incubation_Period_Distribution is either GAUSSIAN_DURATION or POISSON_DURATION. |
|
Incubation_Period_Min |
float |
0 |
3.40E+38 |
0 |
The minimum length of the incubation period. Used when the Incubation_Period_Distribution is UNIFORM_DURATION. |
|
Incubation_Period_Std_Dev |
float |
0 |
3.40E+38 |
1 |
The standard deviation incubation period. Used when the Incubation_Period_Distribution is GAUSSIAN_DURATION. |
|
Infection_Updates_Per_Timestep |
integer |
0 |
144 |
1 |
Number of infection updates executed during each timestep. A timestep defaults to one day. |
|
Infectious_Period_Distribution |
enum |
NA |
NA |
NOT_INITIALIZED |
Distribution of duration of infectious period. Can be fixed or exponentially distributed with an average duration from Base_Infectious_Period. Possible values are: FIXED_DURATION, EXPONENTIAL_DURATION. |
|
Infectious_Period_Max |
float |
0.6 |
3.40E+38 |
0 |
The maximum length of the infectious period. Used when Infectious_Period_Distribution is UNIFORM_DURATION. |
|
Infectious_Period_Mean |
float |
0 |
3.40E+38 |
6 |
The mean of the infectious period. Used when Infectious_Period_Distribution is either GAUSSIAN_DURATION or POISSON_DURATION. |
|
Infectious_Period_Min |
float |
0 |
3.40E+38 |
0 |
The minimum length of the infectious period. Used when the Infectious_Period_Distribution is UNIFORM_DURATION. |
|
Infectious_Period_Std_Dev |
float |
0 |
3.40E+38 |
1 |
The standard deviation of the infectious period. Used when the Infectious_Period_Distribution is GAUSSIAN_DURATION. |
|
Infectivity_Scale_Type |
enum |
NA |
NA |
CONSTANT_INFECTIVITY |
Alter infectivity by time or season. Possible values are: CONSTANT_INFECTIVITY - No infectivity correction is applied. FUNCTION_OF_TIME_AND_LATITUDE - Infectivity is corrected for approximate seasonal forcing. The use of a seasonal infectivity correction is a proxy for the effects of varying climate. From October through March, infectivity increases in the Northern Hemisphere and decreases in the Southern Hemisphere. From April through September, the trend reverses. Regions closer to the equator have reduced forcing compared to temperate regions. This is not a substitute for the weather-driven vector dynamics of vector-borne and malaria simulations. FUNCTION_OF_CLIMATE - Allows infectivity to be modulated by weather directly, for example, relative humidity in airborne simulations or rainfall in environmental simulations. There is no default climate dependence enabled for generic simulations. EXPONENTIAL_FUNCTION_OF_TIME - To facilitate certain burn-in scenarios, infectivity ramps up from zero at the beginning of the simulation according to the functional form, 1-exp(-rate*time), where the rate is specified by the parameter Infectivity_Scaling_Rate. |
|
Maternal_Transmission_Probability |
float |
0 |
1 |
0 |
Probability of transmission of infection from mother to infant at birth. Relevant only if Enable_Maternal_Transmission is set to 1. Note: This parameter should be set to 0 and ignored for malaria and vector simulations. |
|
Max_Individual_Infections |
integer |
0 |
1000 |
1 |
Limit on the number of infections that an individual can have simultaneously. Note: Relevant only if the Enable_Superinfection parameter is set to 1 to allow multiple infections. |
|
Mortality_Blocking_Immunity_Decay_Rate |
float |
0 |
1000 |
0.001 |
Rate at which mortality-blocking immunity decays after the mortality-blocking immunity offset period. Relevant only when Enable_Immunity and Enable_Immune_Decay parameters are set to 1. |
|
Mortality_Time_Course |
enum |
NA |
NA |
DAILY_MORTALITY |
Determines whether disease deaths are calculated on every time step or once at the end of the disease duration. Possible values are: DAILY_MORTALITY and MORTALITY_AFTER_INFECTIOUS. |
|
Number_Basestrains |
integer |
1 |
10 |
1 |
The number of base strains in the simulation, such as antigenic variants. |
|
Number_Substrains |
integer |
1 |
16777200 |
256 |
The number of disease substrains for each base strain, such as genetic variants. |
|
Population_Density_Infectivity_Correction |
enum |
NA |
NA |
CONSTANT_INFECTIVITY |
Correction to alter infectivity by population density set in the Population_Density_C50 parameter. Measured in people per sq. km. Possible values are: CONSTANT_INFECTIVITY and SATURATING_FUNCTION_OF_DENSITY. Note: Sparsely populated areas have a lower infectivity, while densely populated areas have a higher infectivity which rises to saturate at the Base_Infectivity value. |
|
Transmission_Blocking_Immunity_Decay_Rate |
float |
0 |
1000 |
0.001 |
Rate at which transmission-blocking immunity decays after the base transmission-blocking immunity offset period. Relevant only when Enable_Immunity and Enable_Immune_Decay parameters are set to 1. |
|
Transmission_Blocking_Immunity_Duration_Before_Decay |
float |
0 |
45000 |
0 |
Number of days after infection until transmission-blocking immunity begins to decay. Relevant only when Enable_Immunity and Enable_Immune_Decay parameters are set to 1. |
|
Zoonosis_Rate |
float |
0 |
1 |
0 |
The daily rate of zoonotic infection per individual when the Animal_Reservoir_Type parameter is set to either CONSTANT_ZOONOSIS or ZOONOSIS_FROM_DEMOGRAPHICS. |
Geography and environment¶
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Air_Temperature_Variance |
float |
0 |
5 |
2 |
The standard deviation (in degrees Celsius) for normally distributed noise applied to the daily air temperature values when Climate_Model is configured as CLIMATE_CONSTANT or CLIMATE_BY_DATA. Only used if the Enable_Climate_Stochasticity is set to 1. |
|
Base_Air_Temperature |
float |
-55 |
45 |
22 |
The value of air temperature where Climate_Model is set to CLIMATE_CONSTANT. |
|
Base_Land_Temperature |
float |
-55 |
60 |
26 |
The value of land temperature where Climate_Model is set to CLIMATE_CONSTANT. |
|
Base_Rainfall |
float |
0 |
150 |
10 |
The value of rainfall per day in mm when Climate_Model is set to CLIMATE_CONSTANT. |
|
Base_Relative_Humidity |
float |
0 |
1 |
0.75 |
The value of humidity where Climate_Model is set to CLIMATE_CONSTANT. |
|
Climate_Model |
enum |
NA |
NA |
CLIMATE_OFF |
Determines how the climate of a simulation is configured and from what file(s). Possible values are: CLIMATE_OFF, CLIMATE_CONSTANT - Uses the conditional parameters that give the fixed values of temperature or rain for land temperature, air temperature, rainfall and humidity. CLIMATE_KOPPEN - Uses an input file that decodes Koppen codes by geographic location. CLIMATE_BY_DATA - Reads everything out of several input files with additional parameters that allow the addition of stochasticity or scale offsets. |
|
Climate_Update_Resolution |
enum |
NA |
NA |
CLIMATE_UPDATE_YEAR |
Update resolution for data in climate files. Possible values are: CLIMATE_UPDATE_YEAR, CLIMATE_UPDATE_MONTH, CLIMATE_UPDATE_WEEK, CLIMATE_UPDATE_DAY and CLIMATE_UPDATE_HOUR. |
|
Default_Geography_Initial_Node_Population |
integer |
0 |
1000000 |
1000 |
When using the default geography (i.e. Enable_Demographics_Initial = 0), this is the initial number of individuals in each node. |
|
Default_Geography_Torus_Size |
integer |
3 |
100 |
10 |
When using the default geography (i.e. Enable_Demographics_Initial = 0), this is the square root of the number of nodes in the simulation. For migration, the nodes are assumed to be a torus. |
|
Enable_Climate_Stochasticity |
boolean |
0 |
1 |
0 |
Controls overall stochasticity for climate. Use value 0 to disable all additional climate stochasticity. Use value 1 to enable additional variation as specified by the parameters: Air_Temperature_Variance, Land_Temperature_Variance, Enable_Rainfall_Stochasticity and Relative_Humidity_Variance. |
|
Enable_Rainfall_Stochasticity |
boolean |
0 |
1 |
1 |
Set to 0 to disable rainfall stochasticity. Set to 1 to enable stochastic variation of rainfall drawn from an exponential distribution (with a mean value as the daily rainfall from the Climate_Model values, CLIMATE_CONSTANT or CLIMATE_BY_DATA). |
|
Land_Temperature_Offset |
float |
-20 |
20 |
0 |
The linear shift of land surface temperature in degrees Celsius. This is only used or needed when Climate_Model is set to CLIMATE_BY_DATA. |
|
Land_Temperature_Variance |
float |
0 |
7 |
2 |
The standard deviation (in degrees Celsius) for normally distributed noise applied to the daily land temperature values when Climate_Model is configured as CLIMATE_CONSTANT or CLIMATE_BY_DATA. Only used if the Enable_Climate_Stochasticity is set to 1. |
|
Node_Grid_Size |
float |
0.004167 |
90 |
0.004167 |
Spatial resolution indicating the node grid size for a simulation in degrees. |
|
Rainfall_In_mm_To_Fill_Swamp |
float |
1 |
10000 |
1000 |
Millimeters of rain to fill larval habitat to capacity. Only used for vector species with Habitat_Type set to BRACKISH_SWAMP. |
|
Rainfall_Scale_Factor |
float |
0.1 |
10 |
1 |
The scalar used in multiplying rainfall value(s). This is only used or needed when Climate_Model is set to CLIMATE_BY_DATA. |
|
Relative_Humidity_Scale_Factor |
float |
0.1 |
10 |
1 |
The scalar used in multiplying relative humidity value(s). This is only used or needed when Climate_Model is set to CLIMATE_BY_DATA. |
|
Relative_Humidity_Variance |
float |
0 |
0.12 |
0.05 |
The standard deviation (in percentage, %) for normally distributed noise applied to the daily relative humidity values when Climate_Model is configured as CLIMATE_CONSTANT or CLIMATE_BY_DATA. Only used if the Enable_Climate_Stochasticity is set to 1. |
Input files¶
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Air_Migration_Filename |
string |
NA |
NA |
Path of input file defining patterns of migration by airplane. |
||
Air_Temperature_Filename |
string |
NA |
NA |
air_temp.json |
Path of input file defining air temperature data measured 2 meters above ground. Relevant only if Climate_Model is set to CLIMATE_BY_DATA. |
|
Campaign_Filename |
string |
NA |
NA |
Path of the simulation campaign file. This configuration file defines campaign related settings and thus, is an optional file, only required when interventions are part of the simulation. |
||
Demographics_Filenames |
Vector String |
NA |
NA |
Array of demographics files containing information on the identity and demographics of the region to simulate. |
||
Koppen_Filename |
string |
NA |
NA |
UNINITIALIZED STRING |
Path to input file used to specify Koppen climate classifications. This is only used or needed when Climate_Model is set to CLIMATE_KOPPEN. |
|
Land_Temperature_Filename |
string |
NA |
NA |
land_temp.json |
Path of input file defining temperature data measured at land surface. Relevant only if Climate_Model is set to CLIMATE_BY_DATA. |
|
Local_Migration_Filename |
string |
NA |
NA |
Path of input file defining patterns of migration to adjacent nodes by foot travel. |
||
Rainfall_Filename |
string |
NA |
NA |
rainfall.json |
Path of input file defining rainfall data. Relevant only if Climate_Model is set to CLIMATE_BY_DATA. |
|
Regional_Migration_Filename |
string |
NA |
NA |
Path of input file defining patterns of migration by vehicle via road or rail network. If the node is not on a road or rail network, regional migration focuses on the closest hub city in the network. |
||
Relative_Humidity_Filename |
string |
NA |
NA |
rel_hum.json |
Path of input file defining relative humidity data measured 2 meters above ground. Relevant only if Climate_Model is set to CLIMATE_BY_DATA. |
|
Sea_Migration_Filename |
string |
NA |
NA |
Path of input file defining patterns of migration by ship. |
Malaria¶
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Anemia_Mortality_Inverse_Width |
float |
0.1 |
1000000 |
10 |
The inverse width relative to threshold value of severe disease turn-on around threshold for anemia. |
|
Anemia_Mortality_Threshold |
double |
NA |
NA |
3 |
The threshold value on hemoglobin count in grams per deciliter (g/dL) at which 50% of individuals die per day. |
|
Anemia_Severe_Inverse_Width |
float |
0.1 |
1000000 |
10 |
The inverse width relative to threshold value of severe disease turn-on around threshold for anemia. |
|
Anemia_Severe_Threshold |
float |
0 |
100 |
5 |
The severe disease threshold level for anemia. Threshold units are in grams per deciliter (g/dL). |
|
Antibody_Capacity_Growth_Rate |
float |
0 |
1 |
0.1 |
Maximum daily rate of specific antibody capacity increase. Antibody production begins at capacity of 0.3. Hyperimmunity results at 0.4. The actual growth rate will tend to be lower and is a function of antigen concentrations. |
|
Antibody_CSP_Decay_Days |
float |
1 |
3.40E+38 |
90 |
Exponential decay time for CSP antibody concentration (in days). |
|
Antibody_CSP_Killing_Inverse_Width |
float |
1.00E-06 |
1000000 |
1.5 |
Inverse width of sigmoidal sporozoite killing function of CSP antibody concentration. |
|
Antibody_CSP_Killing_Threshold |
float |
1.00E-06 |
1000000 |
10 |
Threshold value on CSP antibody concentration for sporozoite killing. |
|
Antibody_IRBC_Kill_Rate |
double |
NA |
NA |
2 |
Factor multiplied by antibody level to produce the rate of clearance of the infected red blood cell (IRBC) population. |
|
Antibody_Memory_Level |
float |
0 |
0.35 |
0.2 |
Limiting level of antibody capacity that remains after a prolonged absence of a specific antigen. The antibody capacity decays to this level gradually after infection is cleared. The decay rate in antibody capacity is set so that hyperimmunity is lost within 4 months, and capacity continues to decay to this level. The antibody memory level is relevant for year-scale dynamics, but not for long dynamics (10-20 years). |
|
Antibody_Stimulation_C50 |
float |
0.1 |
10000 |
10 |
Concentration of an antigen, measured in IRBC/ul at which growth in antibody capacity against the antigen increases at half the maximum rate specified by Antibody_Capacity_Growth_Rate. |
|
Antigen_Switch_Rate |
double |
NA |
NA |
2.00E-09 |
Antigenic switching rate per infected red blood cell per asexual cycle. See Parasite_Switch_Type for different switching patterns. |
|
Base_Gametocyte_Fraction_Male |
double |
NA |
NA |
0.2 |
Fraction of gametocytes that are male. |
|
Base_Gametocyte_Mosquito_Survival_Rate |
float |
0 |
1 |
0.01 |
Average fraction of gametocytes in blood meal that are successful in infecting mosquito in the absence of other modulating effects, such as fever. |
|
Base_Gametocyte_Production_Rate |
double |
NA |
NA |
0.02 |
Fraction of IRBCs producing gametocytes. |
|
Base_Sporozoite_Survival_Fraction |
float |
0 |
1 |
0.25 |
Fraction of sporozoites that survive to infect a hepatocyte in the absence of anti-CSP protection. |
|
Clinical_Fever_Threshold_High |
float |
0 |
15 |
1 |
The fever threshold (in degrees Celsius above normal) to start a clinical incident. |
|
Clinical_Fever_Threshold_Low |
float |
0 |
5 |
1 |
The fever threshold (in degrees Celsius above normal) to end a clinical incident. |
|
Cytokine_Gametocyte_Inactivation |
float |
0 |
1 |
0.02 |
The strength of inflammatory response inactivation of gametocytes. |
|
Enable_Sexual_Combination |
boolean |
0 |
1 |
0 |
Set to 1 to enable sexual combination of male and female gametocytes. Set to 0 for transmission of each strain only. Note: This parameter is currently not in use. |
|
Erythropoiesis_Anemia_Effect |
float |
0 |
1000 |
3.5 |
Exponential rate of increased red-blood-cell production from reduced red-blood-cell availability. |
|
Falciparum_MSP_Variants |
integer |
0 |
1000 |
100 |
The number of distinct MSP variants for P. falciparum malaria in the overall parasite population in the simulation, not necessarily in an individual. |
|
Falciparum_Nonspecific_Types |
integer |
0 |
1000 |
20 |
The number of distinct non-specific types of P. falciparum malaria. |
|
Falciparum_PfEMP1_Variants |
integer |
0 |
100000 |
1000 |
The number of distinct PfEMP1 variants for P. falciparum malaria in the overall parasite population in the simulation. |
|
Fever_Detection_Threshold |
float |
0.5 |
5 |
1 |
Level of body temperature above normal, in degrees Celsius, corresponding to detectable fever. Normal body temperature is defined as 37 degrees Celsius. |
|
Fever_IRBC_Kill_Rate |
float |
0 |
1000 |
0.15 |
Factor multiplied by fever level to produce a decay rate of infected red blood cell population due to inflammatory innate immune response. |
|
Fever_Mortality_Inverse_Width |
float |
0.1 |
1000000 |
10 |
The inverse width relative to threshold value of mortality turn-on around threshold for fever. |
|
Fever_Mortality_Threshold |
double |
NA |
NA |
3 |
The fever mortality threshold for a simulation in units of degrees Celsius above normal body temperature. Sets the midpoint in the sigmoidal response for mortality rate associated with fever. |
|
Fever_Severe_Inverse_Width |
float |
0.1 |
1000000 |
10 |
The inverse width relative to threshold value of severe disease turn-on around threshold for fever. |
|
Fever_Severe_Threshold |
float |
0 |
10 |
1.5 |
The severe disease threshold level for fever. |
|
Gametocyte_Stage_Survival_Rate |
double |
NA |
NA |
1 |
Rate of gametocyte survival from one development stage to the next in the absence of drugs or inflammatory immune response. |
|
Innate_Immune_Variation_Type |
enum |
NA |
NA |
NONE |
Type of individual innate immune variation. Possible values are: NONE (homogeneous), PYROGENIC_THRESHOLD (parasite threshold resulting in cytokine production), PYROGENIC_THRESHOLD_VS_AGE (age-dependent threshold), CYTOKINE_KILLING (effectiveness of cytokines in killing infected red-blood cells). |
|
Malaria_Model |
enum |
NA |
NA |
MALARIA_MECHANISTIC_MODEL |
The type of malaria model used in a simulation. Possible values are: MALARIA_FIXED_DURATION, MALARIA_EXPONENTIAL_DURATION, MALARIA_REDUCEDSTATE_MODEL and MALARIA_MECHANISTIC_MODEL. Important: All malaria simulations should use MALARIA_MECHANISTIC_MODEL. The others are placeholder values. |
|
Malaria_Strain_Model |
enum |
NA |
NA |
FALCIPARUM_NONRANDOM_STRAIN |
The generator that is used to construct the antigenic repertoire of a malaria infection. Possible values are: FALCIPARUM_NONRANDOM_STRAIN, FALCIPARUM_RANDOM50_STRAIN, FALCIPARUM_RANDOM_STRAIN and FALCIPARUM_STRAIN_GENERATOR. |
|
Maternal_Antibodies_Type |
enum |
NA |
NA |
OFF |
Type of maternal antibody protection. Possible values are: OFF, SIMPLE_WANING, CONSTANT_INITIAL_IMMUNITY (independent of acquired immunity in possible mothers). |
|
Maternal_Antibody_Decay_Rate |
float |
0 |
3.40E+38 |
0.01 |
Decay rate per day in protection from maternal antibodies. |
|
Maternal_Antibody_Protection |
float |
0 |
1 |
0.1 |
Strength of protection from maternal antibodies as a multiple of full antibody killing effect. |
|
Max_MSP1_Antibody_Growthrate |
float |
0 |
1 |
0.02 |
Maximum increase in MSP1 antibody capacity during each asexual cycle. The higher this value, the sooner early clearances are observed and the earlier the parasite density envelope decreases. |
|
Mean_Sporozoites_Per_Bite |
float |
0 |
1000 |
11 |
Mean number of sporozoites per infectious mosquito bite. |
|
Merozoites_Per_Hepatocyte |
double |
NA |
NA |
15000 |
Number of IRBCs caused by a single infected hepatocyte at the start of infection. |
|
Merozoites_Per_Schizont |
double |
NA |
NA |
16 |
Number of merozoites released by a single infected schizont after each asexual cycle. The number of resulting IRBC’s depends on the RBC availability and merozoite-specific immunity. |
|
Min_Adapted_Response |
float |
0 |
1 |
0.02 |
Minimum level of antibody stimulation to novel antigen. The value sets the low-range asymptote for antibody capacity growth, which is calculated from Antibody_Capacity_Growth_Rate and antigen density, in the presence of any nonzero antigen level. |
|
Min_Days_Between_Clinical_Incidents |
float |
0 |
1000000 |
3 |
The number of days with fever below the low threshold before a new incident can start when the high fever threshold is again exceeded. |
|
MSP1_Merozoite_Kill_Fraction |
double |
NA |
NA |
0.5 |
Fraction of merozoites inhibited from invading new erythrocytes when MSP1-specific antibody level is 1. |
|
New_Diagnostic_Sensitivity |
float |
0.0001 |
100000 |
0.01 |
The number of microliters of blood tested to find single parasites in a new diagnostic (corresponds to inverse parasites/microliters sensitivity). |
|
Nonspecific_Antibody_Growth_Rate_Factor |
float |
0 |
1000 |
0.5 |
Factor that adjusts Antibody_Capacity_Growth_Rate for less immunogenic surface proteins called the minor epitopes. |
|
Nonspecific_Antigenicity_Factor |
double |
NA |
NA |
0.2 |
Nonspecific antigenicity factor that adjusts antibody IRBC killrate to account for IRBCs caused by antibody responses to antigenically weak surface proteins. |
|
Number_Of_Asexual_Cycles_Without_Gametocytes |
integer |
0 |
500 |
1 |
Number of asexual reproduction cycles that do not produce gametocytes. All later cycles will produce gametocytes according to Base_Gametocyte_Production_Rate. |
|
Parasite_Mortality_Inverse_Width |
float |
0.1 |
1000000 |
10 |
The inverse width relative to threshold value of mortality turn-on around threshold for parasite density. |
|
Parasite_Mortality_Threshold |
double |
NA |
NA |
2000000 |
The parasite-density mortality threshold for a simulation. |
|
Parasite_Severe_Inverse_Width |
float |
0.1 |
1000000 |
10 |
The inverse width relative to threshold value of severe disease turn-on around threshold for parasite density. |
|
Parasite_Severe_Threshold |
float |
0 |
3.40E+38 |
1000000 |
The severe disease threshold level for parasite density. |
|
Parasite_Smear_Sensitivity |
float |
0.0001 |
100 |
0.1 |
The number of microliters of blood tested to find single parasites in a traditional smear (corresponds to inverse parasites/microliters sensitivity). |
|
Parasite_Switch_Type |
enum |
NA |
NA |
CONSTANT_SWITCH_RATE_2VARS |
Parasite switch type for erythrocyte surface antigens. Possible values are: CONSTANT_SWITCH_RATE_2VARS, RATE_PER_PARASITE_7VARS - this is the one that is best parameterized - and finally, RATE_PER_PARASITE_5VARS_DECAYING. |
|
PKPD_Model |
enum |
NA |
NA |
FIXED_DURATION_CONSTANT_EFFECT |
Determines which pharmacokinetic pharmacodynamic module to use. Possible values are: FIXED_DURATION_CONSTANT_EFFECT and CONCENTRATION_VERSUS_TIME. |
|
Pyrogenic_Threshold |
float |
0.1 |
20000 |
1000 |
Level of bloodstream infection, measured in IRBC per microliter, at which stimulation of the innate inflammatory immune response is half its maximum value. |
|
RBC_Destruction_Multiplier |
double |
NA |
NA |
9.5 |
Number of total RBCs destroyed per infected rupturing schizont. |
Malaria drug efficacy¶
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Bodyweight_Exponent |
float |
0 |
100000 |
0 |
Effect of body weight on maximum drug concentration in an individual patient. |
|
Drug_Cmax |
float |
0 |
100000 |
1000 |
Maximum drug concentration in the same units as Drug_PKPD_C50. Note: Since drug concentrations and C50 values are often specified with different unit systems, these two parameters have been designed to be flexible, provided that the units are the same for each. |
|
Drug_Decay_T1 |
float |
0 |
100000 |
1 |
Primary drug decay rate, in days. |
|
Drug_Decay_T2 |
float |
0 |
100000 |
1 |
Secondary drug decay rate, in days. |
|
Drug_Dose_Interval |
float |
0 |
100000 |
1 |
Interval between doses, in days. |
|
Drug_Fulltreatment_Doses |
float |
1 |
100000 |
3 |
Number of doses for a full treatment. |
|
Drug_Gametocyte02_Killrate |
float |
0 |
100000 |
0 |
Log reduction per day in early-stage gametocytes. |
|
Drug_Gametocyte34_Killrate |
float |
0 |
100000 |
0 |
Log reduction per day in late-stage gametocytes. |
|
Drug_GametocyteM_Killrate |
float |
0 |
100000 |
0 |
Log reduction per day in mature gametocytes at saturated drug concentration. |
|
Drug_Hepatocyte_Killrate |
float |
0 |
100000 |
0 |
Log reduction in hepatocytes per day. |
|
Drug_PKPD_C50 |
float |
0 |
100000 |
100 |
Concentration at which drug killing rates are half of maximum. Note: This parameter and the Drug_Cmax parameter must use the same units. |
|
Drug_Vd |
float |
0 |
100000 |
10 |
Volume of distribution. This value is the ratio of the volume of the second compartment to the volume of the first compartment in a two-compartment model, and is dimensionless. |
|
Fraction_Of_Adult_Dose |
DoseMap |
0 |
1 |
NA |
The fraction of an adult drug dose given to children below the corresponding age (in years). |
|
Fractional_Dose_By_Upper_Age |
array |
NA |
NA |
NA |
An array of doses that contain Upper_Age_In_Years and Fraction_Of_Adult_Dose values, in days. |
|
Malaria_Drug_Params |
JSON object |
NA |
NA |
NA |
This JSON structure defines the name of an anti-malaria drug and the parameters that define it. |
{
"Malaria_Drug_Params": {
"Artemether": {
"Bodyweight_Exponent": 0,
"Drug_Cmax": 114,
"Drug_Decay_T1": 0.12,
"Drug_Decay_T2": 0.12,
"Drug_Vd": 1,
"Drug_PKPD_C50": 0.6,
"Drug_Dose_Interval": 0.5,
"Drug_Fulltreatment_Doses": 6,
"Drug_Gametocyte02_Killrate": 2.5,
"Drug_Gametocyte34_Killrate": 1.5,
"Drug_GametocyteM_Killrate": 0.7,
"Drug_Hepatocyte_Killrate": 0.0,
"Max_Drug_IRBC_Kill": 8.9,
"Fractional_Dose_By_Upper_Age": [
{
"Fraction_Of_Adult_Dose": 0.25,
"Upper_Age_In_Years": 3
},
{
"Fraction_Of_Adult_Dose": 0.5,
"Upper_Age_In_Years": 6
},
{
"Fraction_Of_Adult_Dose": 0.75,
"Upper_Age_In_Years": 10
},
]
}
}
}
|
Max_Drug_IRBC_Kill |
float |
5 |
100000 |
5 |
Maximum log reduction in IRBCs per day. |
|
Upper_Age_In_Years |
DoseMap |
0 |
125 |
NA |
The age (in years) below which children are given the corresponding fraction of an adult drug dose. |
Migration¶
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Air_Migration_Roundtrip_Duration |
float |
0 |
10000 |
1 |
The average time spent at the destination node during a round-trip migration by airplane. |
|
Air_Migration_Roundtrip_Probability |
float |
0 |
1 |
0.8 |
Likelihood that an individual who flies to another cell will return to the cell of origin during the next migration. Relevant only when Enable_Air_Migration is selected. |
|
Enable_Air_Migration |
boolean |
0 |
1 |
0 |
Set to 1 to account for migration by airplane in and out of cities with airports. Set to 0 to assume by default that no migration occurs by air travel. |
|
Enable_Local_Migration |
boolean |
0 |
1 |
0 |
Set to 1 to enable local migration, the diffusion of people in and out of nearby nodes by foot travel. Set to 0 to disable local migration. |
|
Enable_Migration_Heterogeneity |
boolean |
0 |
1 |
1 |
Set to 1 to use migration rate distribution in the demographics file. Set to 0 to assume by default that the same migration rate applies for all individuals. |
|
Enable_Regional_Migration |
boolean |
0 |
1 |
0 |
Set to 1 to account for migration by road vehicle in and out of nodal cities in the road network. Set to 0 to assume by default that no migration occurs by road travel. |
|
Enable_Sea_Migration |
boolean |
0 |
1 |
0 |
Set to 1 to account for migration on ships in and out of coastal cities with seaports. Set to 0 to assume by default that no migration occurs by sea travel. |
|
Local_Migration_Roundtrip_Duration |
float |
0 |
10000 |
1 |
The average time spent at the destination node during a round-trip migration by foot travel. |
|
Local_Migration_Roundtrip_Probability |
float |
0 |
1 |
0.95 |
Likelihood that an individual who walks into a neighboring cell will return to the cell of origin during the next migration. Relevant only when Enable_Local_Migration is selected. |
|
Migration_Model |
enum |
NA |
NA |
NO_MIGRATION |
Model to use for migration. Possible values are: NO_MIGRATION, FIXED_RATE_MIGRATION, VARIABLE_RATE_MIGRATION and LEVY_FLIGHTS. Note: VARIABLE_RATE_MIGRATION and LEVY FLIGHTS are currently not supported. |
|
Migration_Pattern |
enum |
NA |
NA |
RANDOM_WALK_DIFFUSION |
The type of roundtrip. For example, a single roundtrip. Possible values are: RANDOM_WALK_DIFFUSION, SINGLE_ROUND_TRIPS and WAYPOINTS_HOME. |
|
Regional_Migration_Roundtrip_Duration |
float |
0 |
10000 |
1 |
The average time spent at the destination node during a round-trip migration by road network. |
|
Regional_Migration_Roundtrip_Probability |
float |
0 |
1 |
0.1 |
Likelihood that an individual who travels by vehicle to another cell will return to the cell of origin during the next migration. Relevant only when Enable_Regional_Migration is checked. |
|
Roundtrip_Waypoints |
integer |
0 |
1000 |
10 |
The maximum number of points reached during a trip before steps are retraced on the return trip home. |
|
Sea_Migration_Roundtrip_Duration |
float |
0 |
10000 |
1 |
The average time spent at the destination node during a round-trip migration by ship. |
|
Sea_Migration_Roundtrip_Probability |
float |
0 |
1 |
0.25 |
Likelihood that an individual who travels by ship into a neighboring cell will return to the cell of origin during the next migration. Relevant only when Enable_Sea_Migration is checked. |
Output options¶
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Enable_Default_Reporting |
boolean |
0 |
1 |
1 |
Enables and disables the default InsetChart.json report. |
|
Enable_Demographics_Reporting |
boolean |
0 |
1 |
1 |
Outputs demographic summary data and age-binned reports to file. |
|
Enable_Property_Output |
boolean |
0 |
1 |
0 |
Use value 0 to disable Individual_Property reports (separate file). Use value 1 to enable. |
|
Enable_Spatial_Output |
boolean |
0 |
1 |
0 |
Use value 0 to disable all spatial output channels. Use value 1 to enable spatial output of all channels listed in the array parameter: Spatial_Output_Channels. For full documentation of spatial output channels, please see the online documentation. |
|
Report_Event_Recorder |
boolean |
0 |
1 |
1 |
Enables or disables the ReportEventRecorder.csv report. |
Sampling¶
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Base_Individual_Sample_Rate |
float |
0 |
1 |
1 |
Base rate of sampling for individuals. This rate equals the fraction of individuals in each node being sampled. |
|
Individual_Sampling_Type |
enum |
NA |
NA |
TRACK_ALL |
Type of individual human sampling. Possible values are: TRACK_ALL, FIXED_SAMPLING, ADAPTED_SAMPLING_BY_POPULATION_SIZE, ADAPTED_SAMPLING_BY_AGE_GROUP, and ADAPTED_SAMPLING_BY_AGE_GROUP_AND_POP_SIZE. |
|
Max_Node_Population_Samples |
float |
1 |
3.40E+38 |
30 |
Number of individuals when the sampling rate starts dropping to the rate of adapted sampling by population size. |
Scaling factors¶
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
x_Air_Migration |
float |
0 |
3.40E+38 |
1 |
Multiplier for rate of migration by air. |
|
x_Birth |
float |
0 |
3.40E+38 |
1 |
Multiplier of birth rate from input demographics file. |
|
x_Local_Migration |
float |
0 |
3.40E+38 |
1 |
Multiplier for rate of migration by foot travel. |
|
x_Other_Mortality |
float |
0 |
3.40E+38 |
1 |
Multiplier for scaling mortality from causes other than disease being simulated. |
|
x_Regional_Migration |
float |
0 |
3.40E+38 |
1 |
Multiplier for rate of migration by road vehicle. |
|
x_Sea_Migration |
float |
0 |
3.40E+38 |
1 |
Multiplier for rate of migration by sea. |
|
x_Temporary_Larval_Habitat |
double |
NA |
NA |
1 |
Scales the habitat size for all mosquito populations. |
Simulation setup¶
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Config_Name |
string |
NA |
NA |
UNINITIALIZED STRING |
User-supplied title naming a configuration. |
|
Enable_Interventions |
boolean |
0 |
1 |
0 |
Set to 1 to load campaign interventions from file. Set to 0 to run a baseline simulation without campaigns. |
|
Run_Number |
integer |
0 |
2147480000 |
1 |
Sets the random number seed through a bit manipulation process for USE_PSEUDO_DES. When running a multi-core simulation, combines with processor rank to produce independent random number streams for each process. |
|
Simulation_Duration |
float |
0 |
1000000 |
1 |
Elapsed time noted in days from the start to the end of a simulation. |
|
Simulation_Timestep |
float |
0 |
1000000 |
1 |
Value indicating simulation time step in days. |
|
Simulation_Type |
enum |
NA |
NA |
GENERIC_SIM |
Type of disease being simulated. Supported values include GENERIC_SIM, VECTOR_SIM, MALARIA_SIM, TB_SIM, STI_SIM, HIV_SIM, and PY_SIM. To query the values supported for your specific build, use the –version option. |
|
Start_Time |
float |
0 |
1000000 |
1 |
Time noted in days when the simulation begins. This time influences the point in the temporal input data, such as where in the climate input the DTK starts running the simulation. Note: The Start_Day of campaign events is in absolute time. So, time relative to the beginning of the simulation depends on this parameter. |
Spatial output channels¶
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Spatial_Output_Channels |
Fixed String Set |
NA |
NA |
[] |
An array of channel names for spatial output by Node and time step. The data from each channel will be written to a separate binary file. Channel names include Air_Temperature, Births, Campaign_Cost, Disease_Deaths, Human_Infectious_Reservoir, Infection_Rate, Land_Temperature, New_Infections, New_Reported_Infections, Population, Rainfall and Relative_Humidity. |
Vector¶
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Age_Dependent_Biting_Risk_Type |
enum |
NA |
NA |
OFF |
Type of functional form for age-dependent biting-risk. Possible values are: OFF, LINEAR, SURFACE_AREA_DEPENDENT. |
|
Egg_Hatch_Delay_Distribution |
enum |
NA |
NA |
NO_DELAY |
Is there a delay in egg hatching from oviposition? If so, how is that delay distributed? Possible values are: NO_DELAY and EXPONENTIAL_DURATION. |
|
Egg_Saturation_At_Oviposition |
enum |
NA |
NA |
NO_SATURATION |
Is there saturation in the number of viable eggs at oviposition depending on available larval habitat? Possible values are: NO_SATURATION and SATURATION_AT_OVIPOSITION. |
|
Enable_Temperature_Dependent_Feeding_Cycle |
boolean |
0 |
1 |
0 |
Set to 1 to have the duration between blood-feeds depend on the temperature. Otherwise, if set to 0, days between feeds will be constant and specified by the parameter, Days_Between_Feeds, for each species. |
|
Enable_Vector_Aging |
boolean |
0 |
1 |
0 |
Set to 1 to simulate vector senescence. Set to 0 to assume that vector behavior does not change with age. |
|
Enable_Vector_Migration |
boolean |
0 |
1 |
0 |
Set to 1 to enable vector migration. Set to 0 to disable vector migration. Note: Only supported for individual-vector model with Vector_Sampling_Type value of TRACK_ALL_VECTORS or SAMPLE_IND_VECTORS. |
|
Enable_Vector_Migration_Local |
boolean |
0 |
1 |
0 |
Set to 1 to consider local vector migration by vectors to adjacent nodes. Set to 0 to not consider local vector migration. |
|
HEG_Fecundity_Limiting |
float |
0 |
1 |
0 |
The fractional reduction in the number of eggs laid by a female mosquito who is homozygous in the homing endonuclease gene (HEG). |
|
HEG_Homing_Rate |
float |
0 |
1 |
0 |
The fractional redistribution from heterozygous offspring to the homozygous homing-endonuclease-gene (HEG) behavior. |
|
HEG_Model |
enum |
NA |
NA |
OFF |
An enum governing the homing endonuclease gene (HEG) dynamics. Possible values are: OFF, GERMLINE_HOMING (early), EGG_HOMING (late). |
|
Human_Feeding_Mortality |
float |
0 |
1 |
0.1 |
Fraction (dimensionless) of mosquitoes that die while feeding on a human. As of version 1.5, this is now a one-value-per-simulation quantity. |
|
Larval_Density_Dependence |
enum |
NA |
NA |
UNIFORM_WHEN_OVERPOPULATION |
Set to UNIFORM_WHEN_OVERPOPULATION to apply mortality uniformly to all larvae when the population exceeds the specified carrying capacity for that habitat. Set to GRADUAL_INSTAR_SPECIFIC to apply mortality and delayed growth in an instar-specific fashion where the younger larvae are more susceptible to predation and competition from older larvae. The LARVAL_AGE_DENSITY_DEPENDENT_MORTALITY_ONLY and DENSITY_DELAYED_GROWTH_NOT_MORTALITY options configure only the age-dependent mortality component or only the delayed-growth component, respectively. For the NO_DENSITY_DEPENDENCE option, there is no additional larval density-dependent mortality factor. Possible values are: UNIFORM_WHEN_OVERPOPULATION, GRADUAL_INSTAR_SPECIFIC, LARVAL_AGE_DENSITY_DEPENDENT_MORTALITY_ONLY, DENSITY_DELAYED_GROWTH_NOT_MORTALITY, NO_DENSITY_DEPENDENCE. |
|
Larval_Density_Mortality_Offset |
float |
0.0001 |
1000 |
0.1 |
An offset factor in the formula determining the larval-age-dependent mortality for the GRADUAL_INSTAR_SPECIFIC and LARVAL_AGE_DENSITY_DEPENDENT_MORTALITY_ONLY models. |
|
Larval_Density_Mortality_Scalar |
float |
0.01 |
1000 |
10 |
A scale factor in the formula determining the larval-age-dependent mortality for the GRADUAL_INSTAR_SPECIFIC and LARVAL_AGE_DENSITY_DEPENDENT_MORTALITY_ONLY models. |
|
Larval_Rainfall_Mortality_Threshold |
float |
0.01 |
1000 |
100 |
The threshold value on daily rainfall in millimeters, above which larval mortality is applied when Vector_Larval_Rainfall_Mortality is set to either SIGMOID or SIGMOID_HABITAT_SHIFTING. |
|
Mean_Egg_Hatch_Delay |
float |
0 |
120 |
0 |
Mean delay in egg hatch time noted in days from the time of oviposition. |
|
Mosquito_Weight |
integer |
1 |
10000 |
1 |
Value indicating how many mosquitoes are represented by a sample mosquito. This parameter is only used for Vector_Sampling_Type value SAMPLE_IND_VECTORS. |
|
Rainfall_In_mm_To_Fill_Swamp |
float |
1 |
10000 |
1000 |
Millimeters of rain to fill larval habitat to capacity. Only used for vector species with Habitat_Type set to BRACKISH_SWAMP. |
|
Semipermanent_Habitat_Decay_Rate |
float |
0.0001 |
100 |
0.01 |
Daily rate of larval habitat loss for semi-permanent habitats with Habitat_Type parameter of WATER_VEGETATION or BRACKISH_SWAMP. |
|
Temporary_Habitat_Decay_Factor |
float |
0.001 |
100 |
0.05 |
Factor to convert raw evaporation rate (ignoring boundary layer effects) to the daily rate of larval habitat loss for temporary habitats with Habitat_Type parameter of TEMPORARY_RAINFALL. Units are (larval carrying capacity per day) / (kg per square meter per second). |
|
Vector_Larval_Rainfall_Mortality |
enum |
NA |
NA |
NONE |
Type of vector larval mortality function due to rainfall. Possible values are NONE, SIGMOID, where the mortality rate grows linearly from 0 at the threshold to 1 at twice the threshold value, and SIGMOID_HABITAT_SHIFTING, where the threshold value is reduced by a factor proportional to how full the larval habitat is. |
|
Vector_Sampling_Type |
enum |
NA |
NA |
TRACK_ALL_VECTORS |
Type of vector module used in a simulation. Possible values are: TRACK_ALL_VECTORS, SAMPLE_IND_VECTORS, VECTOR_COMPARTMENTS_NUMBER and VECTOR_COMPARTMENTS_PERCENT. |
|
Vector_Species_Names |
Dynamic String Set |
NA |
NA |
[] |
An array of vector species names for all species that are present in the simulation. For example, [“arabiensis”, “funestus”, “gambiae”]. Any vector species name may be used. However, you must specify the parameters as listed in Vector_Species_Params for each species. See Vector Species Parameters for a usage example. |
|
Vector_Sugar_Feeding_Frequency |
enum |
NA |
NA |
VECTOR_SUGAR_FEEDING_NONE |
How often does a female mosquito take a sugar feed? Possible values are: VECTOR_SUGAR_FEEDING_NONE - No sugar feeding. VECTOR_SUGAR_FEEDING_ON_EMERGENCE_ONLY - Sugar feeding once at emergence. VECTOR_SUGAR_FEEDING_EVERY_FEED - Sugar feeding occurs once per blood meal. VECTOR_SUGAR_FEEDING_EVERY_DAY - Sugar feeding occurs everyday. Note: This parameter is used in conjunction with the SugarTrap and OvipositionTrap interventions. These features require Vector_Sampling_Type to be set to TRACK_ALL_VECTORS or SAMPLE_IND_VECTORS. |
|
Wolbachia_Infection_Modification |
float |
0 |
100 |
1 |
The change in vector susceptibility to infection due to a Wolbachia infection. |
|
Wolbachia_Mortality_Modification |
float |
0 |
100 |
1 |
The change in vector mortality due to a Wolbachia infection. |
Vector species¶
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Acquire_Modifier |
float |
0 |
1 |
1 |
Modifier of the probability of successful infection of a mosquito by a malaria-infected individual, given the individual’s infectiousness. |
{"Acquire_Modifier": 0.2}
|
Adult_Life_Expectancy |
float |
0 |
730 |
10 |
Number of days an average mosquito survives. The daily adult mortality rate is 1 / (value of this parameter). |
|
Anthropophily |
float |
0 |
1 |
1 |
Propensity (dimensionless) of the mosquito to feed on humans instead of animals, measured as a fraction of blood meals from human hosts. This propensity is important in differentiating the mosquito species. |
|
Aquatic_Arrhenius_1 |
float |
0 |
1.00E+15 |
8.42E+10 |
The duration of aquatic development is a decreasing function of temperature. The daily rate of fractional progression from egg-hatching to emergence is parameterized by the Arrhenius equation, a1 * exp( -a2 / T ), with T in degrees Kelvin. This parameter, a1, is a temperature-independent scale factor on the development rate. Reference: Craig M.H., Snow R.W., et al. (1999). A climate-based distribution model of malaria transmission in sub-Saharan Africa. Parasitology Today 15(3) 105-111. |
|
Aquatic_Arrhenius_2 |
float |
0 |
1.00E+15 |
8328 |
The duration of aquatic development is a decreasing function of temperature. The daily rate of fractional progression from egg-hatching to emergence is parameterized by the Arrhenius equation, a1 * exp( -a2 / T ), with T in degrees Kelvin. This parameter, a2, governs how quickly the rate changes with temperature. Reference: Craig M.H., Snow R.W., et al. (1999). A climate-based distribution model of malaria transmission in sub-Saharan Africa. Parasitology Today 15(3) 105-111. |
|
Aquatic_Mortality_Rate |
float |
0 |
1 |
0.1 |
Base aquatic mortality per day for the species before adjustment for effects of overpopulation and drying. The rate is locally adjusted for other considerations, including overpopulation and drying out of the aquatic habitat. |
|
Days_Between_Feeds |
float |
0 |
730 |
2 |
Factor (in days) used to calculate the rate at which female mosquitoes attempt to feed. The rate equals 1 / (value of this parameter). The Days_Between_Feeds parameter represents a 3-day feeding cycle at normal tropical temperatures (30 C). If it is colder, the cycle is 4 days at 24 C, and if the temperature is warmer, the cycle is shorter, increasing the fraction of 2-day cycles from 30 C to 48 C. Whole temperature-dependent distribution can be shifted by changing Days_Between_Feeds which is always the value at 30 C. See Enable_Temperature_Dependent_Feeding_Cycle for weather-dependent feeding. |
|
Egg_Batch_Size |
float |
0 |
10000 |
100 |
Size of egg batch laid by one female mosquito that has fed successfully. |
|
Immature_Duration |
float |
0 |
730 |
2 |
Factor (in days) used to calculate the rate at which immature organisms develop into adult mosquitoes, which equals 1 / (value of this parameter). Note that development from immature individual to adult is not dependent on temperature. |
|
Indoor_Feeding_Fraction |
float |
0 |
1 |
1 |
Fraction (dimensionless) of feeds on humans that occur indoors. Thus the fraction of feeds on humans that occur outdoors equals 1 - (value of this parameter).” |
|
Infected_Arrhenius_1 |
float |
0 |
1.00E+15 |
1.17E+11 |
The duration of sporogony is a decreasing function of temperature. The daily rate of fractional progression of infected mosquitoes to an infectious state is parameterized by the Arrhenius equation, a1 * exp( -a2 / T ), with T in degrees Kelvin. This parameter, a1, is a temperature-independent scale factor on the development rate. Reference: Craig M.H., Snow R.W., et al. (1999). A climate-based distribution model of malaria transmission in sub-Saharan Africa. Parasitology Today 15(3) 105-111. |
|
Infected_Arrhenius_2 |
float |
0 |
1.00E+15 |
8340 |
The duration of sporogony is a decreasing function of temperature. The daily rate of fractional progression of infected mosquitoes to an infectious state is parameterized by the Arrhenius equation, a1 * exp( -a2 / T ), with T in degrees Kelvin. This parameter, a2, governs how quickly the rate changes with temperature. Reference: Craig M.H., Snow R.W., et al. (1999). A climate-based distribution model of malaria transmission in sub-Saharan Africa. Parasitology Today 15(3) 105-111. |
|
Infected_Egg_Batch_Factor |
float |
0 |
10 |
0.8 |
Factor (dimensionless) used to modify egg batch size to account for effects of infection in a female mosquito. Female mosquitoes undergoing sporogony have reduced fertility.” |
|
Infectious_Human_Feed_Mortality_Factor |
float |
0 |
1000 |
1.5 |
Correction factor (dimensionless) used to modify the death rate of mosquitoes when feeding on humans, to account for the increase caused by infections in mosquitoes. Infected mosquitoes have a higher mortality rate while feeding on humans than uninfected mosquitoes. References: Wekesa J. W., Copeland R. S. , et al. (1992). Effect of Plasmodium Falciparum on Blood Feeding Behavior of Naturally Infected Anopheles Mosquitoes in Western Kenya. Am J Trop Med Hyg 47(4): 484-488.Anderson R. A., Knols B. G. J., et al. (2000). Plasmodium falciparum sporozoites increase feeding-associated mortality of their mosquito hosts Anopheles gambiae s.l. Parasitology 120(04): 329-333. |
|
Larval_Habitat_Types |
nested JSON object |
0 |
1.00E+15 |
1.25E+10 |
A measure of the habitat type and scaling factors to use to estimate larval population. You can specify this with a dictionary specifying habitat type with a simple numeric scaling factor or, for LINEAR_SPLINE, with a more detailed configuration for scaling. Possible habitat values are:
The numeric scaling value represents larval density with the number of larvae in a 1x1-degree area. The factor multiplicatively scales the resulting weather or population dependent functional form. The LINEAR_SPLINE configuration specifies the day of year, larval value, and larval capacity scaling number. The model linearly interpolates the values to estimate the habitat availability for each vector species without requiring climatological data. |
The following example shows how to specify the larval habitat using one of the climatological habitat types. {"Larval_Habitat_Types": {"TEMPORARY_RAINFALL": 1000000}}
The following example shows how to specify larval habitat using LINEAR_SPLINE. {
"Larval_Habitat_Types":
{
"LINEAR_SPLINE":
{
"Capacity_Distribution_Per_Year":
{
"Times": [0.0, 60.833, 121.667, 182.5, 243.333, 204.167],
"Values": [0.0, 0.0, 0.2, 1.0, 0.5, 0.0]
},
"Max_Larval_Capacity": 10000000000
}
}
}
|
Vector_Species_Params |
JSON object |
NA |
NA |
NA |
For each of the species specified in Vector_Species_Name, there is a set of required species-parameters. The code example, using the arabiensis parameter value, configures the vector Anopheles arbiensis. The parameter name values for the other species are: farauti, fenestus and bambiae. |
{
"Vector_Species_Params": {
"arabiensis": {
"Acquire_Modifier": 1,
"Adult_Life_Expectancy": 10,
"Anthropophily": 0.95,
"Aquatic_Arrhenius_1": 84200000000,
"Aquatic_Arrhenius_2": 8328,
"Aquatic_Mortality_Rate": 0.1,
"Days_Between_Feeds": 3,
"Egg_Batch_Size": 100,
"Immature_Duration": 4,
"Indoor_Feeding_Fraction": 0.5,
"Infected_Arrhenius_1": 117000000000,
"Infected_Arrhenius_2": 8336,
"Infected_Egg_Batch_Factor": 0.8,
"Infectious_Human_Feed_Mortality_Factor": 1.5,
"Larval_Habitat_Types": {
"TEMPORARY_RAINFALL": 1e10,
"WATER_VEGETATION": 1e6
},
"Transmission_Rate": 0.5
}
}
}
|
Campaign parameters¶
The parameters described in this reference section determine the interventions used as part of a campaign to control the spread of a disease and the outbreak of the disease itself. Interventions are configured in the campaign file. They are enabled in the configuration file by setting the parameter, Enable_Interventions, to 1. The configuration file also indicates the name of the campaign file.
Warning
The Outbreak must be the last event in the campaign file or none of the interventions will take place.
EMOD supports a variety of supported campaign interventions. It is worth noting that each intervention contains multiple parameters used to configure that specific event. For example, the distribution of bednets behaves differently in the model than does the use of larvicides. It is therefore important to ensure that interventions are configured correctly and applied to the appropriate level target (individual versus node).
This framework enables rigorous testing of possible control strategies to determine which events or combination of events will best aid in the elimination of disease for specific geographic locations.
Like the configuration file, the campaign file is a JavaScript Object Notation (JSON) formatted file. For information on JSON, see EMOD parameter reference. However, the configuration file is mostly a flat list of JSON key-value pairs while the campaign file is structured as an array of campaign events that contain many nested JSON objects. For more information on how to set up the elements in a campaign file, see Create a campaign file.
Warning
Parameters are case-sensitive. For Boolean parameters, set to 1 for true or 0 for false.
The tables below contain only parameters available when using the malaria simulation type.
Contents
Events array¶
The campaign file contains an Events array with one or more CampaignEvent elements and a Use_Defaults parameter. If Use_Defaults is set to 1, the simulation will use the default values for required parameters that are not configured in the campaign file. If it is set to 0, you must configure all required parameters.
{
"Events": [{
"Event_Name": "Campaign Event 1",
"class": "CampaignEvent"
}, {
"Event_Name": "Campaign Event 2",
"class": "CampaignEvent"
}],
"Use_Defaults": 1
}
Campaign event¶
To distribute an intervention, you must configure a campaign event, an event coordinator, and the intervention. The campaign event configures when the event occurs (Start_Day) and where it is distributed (Nodeset_Config). The event coordinator is a nested JSON object in Event_Coordinator_Config. For more information on how to structure this file, see Create a campaign file.
{
"Event_Name": "Campaign Event Example",
"class": "CampaignEvent",
"Start_Day": 1,
"Nodeset_Config": {
},
"Event_Coordinator_Config": {
}
}
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Event_Coordinator_Config |
EventCoordinator |
0 |
3.40E+38 |
1 |
The configuration of the EventCoordinator to be instantiated. |
|
Nodeset_Config |
NodeSet |
0 |
3.40E+38 |
1 |
The NodeSet describing the nodes covered by the event. |
|
Start_Day |
float |
0 |
3.40E+38 |
1 |
Day of the simulation to activate the event’s event coordinator. |
The Event_Coordinator_Config configures who can receive the intervention (for example, Target_Demographic and Demographic_Coverage) and which intervention is distributed. The intervention is a nested JSON object in Intervention_Config.
{
"Event_Coordinator_Config": {
"class": "The event coordinator that will be used.",
"Target_Demographic": "Everyone",
"Demographic_Coverage": 0.0005,
"Intervention_Config": {}
}
}
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Demographic_Coverage |
float |
0 |
1 |
1 |
Fraction of individuals in the target demographic that will receive this intervention. |
|
Intervention_Config |
Intervention |
0 |
3.40E+38 |
3.40E+38 |
The nested json of the actual intervention to be distributed by this event coordinator. |
|
Number_Repetitions |
integer |
-1 |
1000 |
-1 |
Number of times intervention given, used with tsteps_between_reps. |
|
Property_Restrictions |
Dynamic String Set |
NA |
NA |
[] |
Fraction of individuals in the target demographic that will receive this intervention. |
|
Property_Restrictions_Within_Node |
PropertyRestrictions |
NA |
NA |
Everyone |
Fraction of individuals in the target demographic that will receive this intervention. |
|
Target_Age_Max |
float |
0 |
3.40E+38 |
3.40E+38 |
Upper end of age targeted for intervention, in years. |
|
Target_Age_Min |
float |
0 |
3.40E+38 |
0 |
Lower end of age targeted for intervention, in years. |
|
Target_Demographic |
enum |
NA |
NA |
Everyone |
The demographic group (from a list of possibles) targeted by this intervention (e.g., Infants). |
|
Target_Gender |
enum |
NA |
NA |
All |
Specify the gender restriction for the intervention. Defaults to All. |
|
Target_Residents_Only |
boolean |
0 |
1 |
0 |
If true, only distribute the intervention to those that claim the node as their residence - i.e. started in the node |
|
Timesteps_Between_Repetitions |
integer |
-1 |
10000 |
-1 |
Repetition interval. |
There are three configuration options: NodeSetAll, NodeSetNodeList and NodeSetPolygon.
If you are running a simulation that is in a single location (node) or you want the intervention to be distributed in all of the nodes, use NodeSetAll.
{
"Nodeset_Config": "NodeSetAll"
}
The intervention is distributed to nodes in the intervention list.
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Node_List |
NodeListConfig |
NA |
NA |
SHAPE |
Comma-separated list of node IDs. |
The intervention is distributed in nodes within the specified polygon.
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Polygon_Format |
enum |
NA |
NA |
SHAPE |
Right now just SHAPE. |
|
Vertices |
string |
NA |
NA |
SHAPE |
Comma-separated list of polygon points. |
Interventions¶
Name |
Intervention Level |
Target Life Stage |
Target Biting Preference |
Target Biting Location |
Effect |
Notes |
|---|---|---|---|---|---|---|
AnimalFeedKill |
node |
feeding cycle |
animal |
– |
killing |
|
ArtificialDiet |
node |
feeding cycle |
human |
all |
blocking |
Artificial_Diet_Target: AD_WithinVillage, AD_OutsideVillage |
HumanHostSeekingTrap |
individual |
feeding cycle |
human |
killing |
||
InsectKillingFence |
node |
feeding cycle |
all |
all |
killing |
Photonic fence. Two killing chances per feeding cycle: to and from oviposition site. |
InsectKillingFenceHousingModification |
individual |
killing |
||||
IRSHousingModification |
individual |
feeding cycle |
human |
indoor |
killing, blocking |
|
Ivermectin |
individual |
feeding cycle |
human |
all |
killing |
|
Larvicides |
node |
larva |
all |
all |
killing, reduction |
Specify targeted habitat. |
MosquitoRelease |
node |
|||||
OutdoorRestKill |
node |
feeding cycle |
human |
outdoor |
killing |
|
OvipositionTrap |
node |
feeding cycle |
all |
all |
killing |
Specify targeted habitat. Requires individual vectors. |
ScaleLarvalHabitat |
node |
larva |
all |
all |
reduction |
Specify targeted habitat. |
ScreeningHousingModification |
individual |
feeding cycle |
human |
indoor |
killing, blocking |
|
SimpleBednet |
individual |
feeding cycle |
human |
indoor |
killing, blocking |
|
SimpleIndividualRepellent |
individual |
feeding cycle |
human |
all |
blocking |
|
SpaceSpraying (SpaceSpray_Indoor) |
node |
feeding cycle |
human |
indoor |
killing |
Node-based IRS for household. |
SpaceSpraying |
node |
feeding cycle |
human |
outdoor |
killing |
SpaceSpray_FemalesOnly, MalesOnly, FemalesAndMales |
SpatialRepellent |
node |
feeding cycle |
all |
outdoor |
blocking |
|
SpatialRepellentHousingModification |
individual |
feeding cycle |
human |
indoor |
killing, blocking |
|
SugarTrap |
node |
adults |
all |
all |
killing |
Requires individual vectors. Killing parameter specifies effective fraction of sugar feeds that occur on bait stations. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Cost_To_Consumer |
float |
0 |
999999 |
10 |
Unit cost per vector control (unamortized). |
|
Intervention_Name |
string |
0 |
43800 |
365 |
The optional name used to refer to this intervention as a means to differentiate it from others that use the same class. |
|
Killing_Config |
WaningEffect |
0 |
999999 |
10 |
Killing of the targeted stage. A parameter that is decayed. |
When using antimalarial drugs in the model, the configuration file sets the main effects of the drug, while the intervention (configured in the campaign file) determines which drugs are used and to whom they are distributed.
The following example provides the syntax to configure the AntiMalarialDrug intervention:
{
"Intervention_Config": {
"class": "AntimalarialDrug",
"Cost_To_Consumer": 3.75,
"Drug_Type": "Chloroquine",
"Dosing_Type": "FullTreatmentParasiteDetect"
}
}
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Cost_To_Consumer |
float |
0 |
999999 |
10 |
Unit cost per drug (unamortized). |
|
Dont_Allow_Duplicates |
boolean |
0 |
1 |
0 |
If an individual’s container has an intervention, set to 1 to prevent them from receiving another copy of the intervention. Supported by all intervention types. |
|
Dosing_Type |
enum |
NA |
NA |
SingleDose |
Type of anti-malarial dosing to distribute in a drugs intervention. Possible values are: SingleDose, FullTreatmentCourse, Prophylaxis, SingleDoseWhenSymptom, FullTreatmentWhenSymptom, SingleDoseParasiteDetect, FullTreatmentParasiteDetect, SingleDoseNewDetectionTech, FullTreatmentNewDetectionTech. |
|
Drug_Type |
Constrained String |
0 |
3.40E+38 |
3.40E+38 |
Type of anti-malarial drug to distribute in a drugs intervention. Possible values are: Artemisinin (AR), Chloroquine (CQ), Quinine (QN), Sulfadoxine-Pyramethamine (SP), Primaquine (PQ), Artemether_Lumefantrine (AR_LM), Generic Transmission Blocking (TBD), Generic Pre-erythrocytic (PED), Tafenoquine (TQ). |
|
Intervention_Name |
string |
0 |
3.40E+38 |
3.40E+38 |
The optional name used to refer to this intervention as a means to differentiate it from others that use the same class. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Artificial_Diet_Target |
enum |
NA |
NA |
AD_WithinVillage |
The targeted deployment of artificial diet. Possible values are: AD_WithinVillage, AD_OutsideVillage. |
|
Attraction_Config |
WaningEffect |
NA |
NA |
AD_WithinVillage |
Fraction of feeds attracted to the artificial diet. A parameter that is decayed. |
|
Cost_To_Consumer |
float |
0 |
999999 |
10 |
Unit cost per vector control (unamortized). |
|
Intervention_Name |
string |
0 |
1 |
0 |
The optional name used to refer to this intervention as a means to differentiate it from others that use the same class. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Blocking_Config |
WaningEffect |
0 |
999999 |
8 |
The configuration of pre-feed mosquito repellency and waning for housing modification. |
|
Cost_To_Consumer |
float |
0 |
999999 |
8 |
Unit cost per housing modification (unamortized). |
|
Dont_Allow_Duplicates |
boolean |
0 |
1 |
0 |
If an individual’s container has an intervention, set to 1 to prevent them from receiving another copy of the intervention. Supported by all intervention types. |
|
Intervention_Name |
string |
0 |
1 |
0 |
The optional name used to refer to this intervention as a means to differentiate it from others that use the same class. |
|
Killing_Config |
WaningEffect |
0 |
999999 |
8 |
The configuration of killing efficacy and waning for housing modification. Killing is conditional on NOT blocking the mosquito before feeding. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Actual_IndividualIntervention_Config |
IndividualIntervention |
0 |
1 |
0 |
The configuration of the actual individual-based intervention sought. Selects a class for the intervention and configures the parameters specific for that intervention class. |
|
Demographic_Coverage |
float |
0 |
1 |
1 |
Fraction of individuals in the target demographic that will receive this intervention. |
|
Duration |
float |
-1 |
3.40E+38 |
-1 |
The number of days to continue this intervention. |
|
Intervention_Name |
string |
-1 |
1 |
0 |
The optional name used to refer to this intervention as a means to differentiate it from others that use the same class. |
|
Property_Restrictions |
Dynamic String Set |
NA |
NA |
[] |
Fraction of individuals in the target demographic that will receive this intervention. |
|
Property_Restrictions_Within_Node |
PropertyRestrictions |
NA |
NA |
Everyone |
Fraction of individuals in the target demographic that will receive this intervention. |
|
Target_Age_Max |
float |
0 |
3.40E+38 |
3.40E+38 |
Upper end of age targeted for intervention, in years. |
|
Target_Age_Min |
float |
0 |
3.40E+38 |
0 |
Lower end of age targeted for intervention, in years. |
|
Target_Demographic |
enum |
NA |
NA |
Everyone |
The demographic group (from a list of possibles) targeted by this intervention (e.g., Infants). |
|
Target_Gender |
enum |
NA |
NA |
All |
Specify the gender restriction for the intervention. Defaults to All. |
|
Target_Residents_Only |
boolean |
0 |
1 |
0 |
If true, only distribute the intervention to those that claim the node as their residence - i.e. started in the node |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Broadcast_Event |
Constrained String |
0 |
1 |
0 |
The event that should occur at the end of the delay period. |
|
Dont_Allow_Duplicates |
boolean |
0 |
1 |
0 |
If an individual’s container has an intervention, set to 1 to prevent them from receiving another copy of the intervention. Supported by all intervention types. |
|
Intervention_Name |
string |
0 |
1 |
0 |
The optional name used to refer to this intervention as a means to differentiate it from others that use the same class. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Actual_IndividualIntervention_Configs |
IndividualIntervention |
0 |
3.40E+38 |
6 |
Array of nested interventions to be distributed at end of delay period to covered fraction. |
|
Coverage |
float |
0 |
1 |
1 |
Fraction of individuals receiving delayed distribution of configured interventions. |
|
Delay_Distribution |
enum |
NA |
NA |
NOT_INITIALIZED |
Distribution of duration of delay period. Supported values are: FIXED_DURATION UNIFORM_DURATION GAUSSIAN_DURATION EXPONENTIAL_DURATION. |
|
Delay_Period |
float |
0 |
3.40E+38 |
6 |
If a FIXED_DURATION is specified as the Delay_Distribution, add this parameter Delay_Period to directly specify the time delay (in number of days). If EXPONENTIAL_DURATION is specified as the Delay_Distribution, Delay_Period, represents the exponential rate that describes the distribution of the time delay (in units of 1/days). |
|
Delay_Period_Max |
float |
0.6 |
3.40E+38 |
0 |
If a UNIFORM_DURATION is specified as the Delay_Distribution, add this parameter Delay_Period_Max to directly specify the maximum time delay (in number of days). |
|
Delay_Period_Mean |
float |
0 |
3.40E+38 |
6 |
If a GAUSSIAN_DURATION is specified as the Delay_Distribution, add this parameter Delay_Period_Mean to directly specify the mean time delay (in number of days). |
|
Delay_Period_Min |
float |
0 |
3.40E+38 |
0 |
If a UNIFORM_DURATION is specified as the Delay_Distribution, add this parameter Delay_Period_Min to directly specify the minimum time delay (in number of days). |
|
Delay_Period_Std_Dev |
float |
0 |
3.40E+38 |
1 |
If a GAUSSIAN_DURATION is specified as the Delay_Distribution, add this parameter Delay_Period_Std_Dev to specify the standard deviation describing the Gaussian distribution (in number of days). |
|
Dont_Allow_Duplicates |
boolean |
0 |
1 |
0 |
If an individual’s container has an intervention, set to 1 to prevent them from receiving another copy of the intervention. Supported by all intervention types. |
|
Intervention_Name |
string |
0 |
1 |
0 |
The optional name used to refer to this intervention as a means to differentiate it from others that use the same class. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Actual_IndividualIntervention_Config |
IndividualIntervention |
0 |
1 |
1 |
The configuration of an actual intervention sought. Selects a class for the intervention and configures the parameters specific for that intervention class. |
|
Actual_IndividualIntervention_Event |
Constrained String |
1900 |
2200 |
2000 |
The event of an actual intervention sought. Selects a class for the intervention and configures the parameters specific for that intervention class. |
|
Dont_Allow_Duplicates |
boolean |
0 |
1 |
0 |
If an individual’s container has an intervention, set to 1 to prevent them from receiving another copy of the intervention. Supported by all intervention types. |
|
Event_Or_Config |
enum |
NA |
NA |
Config |
Specifies whether the current intervention distributes a nested intervention (the Config option) or an event will be broadcast which may trigger other interventions in the campaign file (the Event option). See event list. |
|
Intervention_Name |
string |
1900 |
2200 |
2000 |
The optional name used to refer to this intervention as a means to differentiate it from others that use the same class. |
|
Single_Use |
boolean |
0 |
1 |
1 |
One-time, or persistent? |
|
Tendency |
float |
0 |
1 |
1 |
The probability of seeking healthcare. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Attract_Config |
WaningEffect |
0 |
999999 |
3.75 |
The configuration of killing efficacy and waning for human host-seeking trap. |
|
Cost_To_Consumer |
float |
0 |
999999 |
3.75 |
Unit cost per trap (unamortized). |
|
Dont_Allow_Duplicates |
boolean |
0 |
1 |
0 |
If an individual’s container has an intervention, set to 1 to prevent them from receiving another copy of the intervention. Supported by all intervention types. |
|
Intervention_Name |
string |
0 |
999999 |
10 |
The optional name used to refer to this intervention as a means to differentiate it from others that use the same class. |
|
Killing_Config |
WaningEffect |
0 |
999999 |
3.75 |
The configuration of killing efficacy and waning for human host-seeking trap. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Cost_To_Consumer |
float |
0 |
999999 |
10 |
Unit cost per vector control (unamortized). |
|
Intervention_Name |
string |
0 |
1 |
0 |
The optional name used to refer to this intervention as a means to differentiate it from others that use the same class. |
|
Killing_Config |
WaningEffect |
0 |
999999 |
10 |
Killing of the targeted stage. A parameter that is decayed. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Blocking_Config |
WaningEffect |
0 |
999999 |
8 |
The configuration of pre-feed mosquito repellency and waning for housing modification. |
|
Cost_To_Consumer |
float |
0 |
999999 |
8 |
Unit cost per housing modification (unamortized). |
|
Dont_Allow_Duplicates |
boolean |
0 |
1 |
0 |
If an individual’s container has an intervention, set to 1 to prevent them from receiving another copy of the intervention. Supported by all intervention types. |
|
Intervention_Name |
string |
NA |
NA |
Config |
The optional name used to refer to this intervention as a means to differentiate it from others that use the same class. |
|
Killing_Config |
WaningEffect |
0 |
999999 |
8 |
The configuration of killing efficacy and waning for housing modification. Killing is conditional on NOT blocking the mosquito before feeding. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Times |
float |
0 |
99999 |
NA |
An array of years. |
|
Values |
float |
0 |
3.40282e+38 |
NA |
An array of values to match the defined Times. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Blocking_Config |
WaningEffect |
0 |
999999 |
8 |
The configuration of pre-feed mosquito repellency and waning for housing modification. |
|
Cost_To_Consumer |
float |
0 |
999999 |
8 |
Unit cost per housing modification (unamortized). |
|
Dont_Allow_Duplicates |
boolean |
0 |
1 |
0 |
If an individual’s container has an intervention, set to 1 to prevent them from receiving another copy of the intervention. Supported by all intervention types. |
|
Intervention_Name |
string |
0 |
1 |
0 |
The optional name used to refer to this intervention as a means to differentiate it from others that use the same class. |
|
Killing_Config |
WaningEffect |
0 |
999999 |
8 |
The configuration of killing efficacy and waning for housing modification. Killing is conditional on NOT blocking the mosquito before feeding. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Age_Dependence |
enum |
NA |
NA |
OFF |
Dependence of InputEIR on age. |
|
Intervention_Name |
string |
0 |
3.40E+38 |
1 |
The optional name used to refer to this intervention as a means to differentiate it from others that use the same class. |
|
Monthly_EIR |
InputEIRConfig |
NA |
NA |
OFF |
An array of 12 elements that contain an entomological inoculation rate (EIR) for each month. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Actual_IndividualIntervention_Configs |
IndividualIntervention |
0 |
1 |
0 |
An array of interventions distributed as specified in the calendar. Selects a class for the intervention and configures the parameters specific for that intervention class. |
|
Calendar |
CalendarIV |
0 |
1 |
0 |
An array of ages days and probability of receiving an intervention. |
|
Dont_Allow_Duplicates |
boolean |
0 |
1 |
0 |
If an individual’s container has an intervention, set to 1 to prevent them from receiving another copy of the intervention. Supported by all intervention types. |
|
Dropout |
boolean |
0 |
1 |
0 |
If an intervention in the series is missed, all subsequent interventions are also missed. If false (0), all calendar dates/doses are applied independently of each other. If true (1), a missed dose (by applying the probability) will result in no further doses, that is, a miss means you have dropped out altogether. |
|
Intervention_Name |
string |
0 |
1 |
0 |
The optional name used to refer to this intervention as a means to differentiate it from others that use the same class. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Cost_To_Consumer |
float |
0 |
999999 |
8 |
Unit cost per Ivermectin dosing (unamortized). |
|
Dont_Allow_Duplicates |
boolean |
0 |
1 |
0 |
If an individual’s container has an intervention, set to 1 to prevent them from receiving another copy of the intervention. Supported by all intervention types. |
|
Intervention_Name |
string |
0 |
1 |
0 |
The optional name used to refer to this intervention as a means to differentiate it from others that use the same class. |
|
Killing_Config |
WaningEffect |
0 |
999999 |
8 |
The configuration of drug-killing efficacy and waning. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Blocking_Config |
WaningEffect |
NA |
NA |
TEMPORARY_RAINFALL |
The configuration of larval habitat reduction and waning for targeted stage. |
|
Cost_To_Consumer |
float |
0 |
999999 |
10 |
Unit cost per vector control (unamortized). |
|
Habitat_Target |
enum |
NA |
NA |
TEMPORARY_RAINFALL |
The larval habitat type targeted for larvicide. Possible values are: TEMPORARY_RAINFALL WATER_VEGETATION HUMAN_POPULATION CONSTANT BRACKISH_SWAMP ALL_HABITATS. |
|
Intervention_Name |
string |
0 |
1 |
0 |
The optional name used to refer to this intervention as a means to differentiate it from others that use the same class. |
|
Killing_Config |
WaningEffect |
NA |
NA |
TEMPORARY_RAINFALL |
The configuration of larval killing efficacy and waning for targeted stage. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Challenge_Type |
enum |
NA |
NA |
InfectiousBites |
The type of malaria challenge. Possible values are: InfectiousBites Sporozoites. |
|
Coverage |
float |
0 |
1 |
1 |
Fraction of individuals receiving distribution. |
|
Infectious_Bite_Count |
integer |
0 |
1000 |
1 |
Number of infectious bites. |
|
Intervention_Name |
string |
NA |
NA |
Config |
The optional name used to refer to this intervention as a means to differentiate it from others that use the same class. |
|
Sporozoite_Count |
integer |
0 |
1000 |
1 |
Number of sporozoites. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Cost_To_Consumer |
float |
0 |
999999 |
0 |
Cost of each mosquito release. |
|
Intervention_Name |
string |
NA |
NA |
WOLBACHIA_FREE |
The optional name used to refer to this intervention as a means to differentiate it from others that use the same class. |
|
Mated_Genetics |
ResistanceHegGenetics |
1 |
100000000 |
10000 |
The genetic properties of the mate if released mosquitoes have mated, e.g. HEG and pesticide resistance. |
|
Released_Gender |
enum |
NA |
NA |
VECTOR_FEMALE |
Gender of released mosquitoes. Possible values are: VECTOR_FEMALE VECTOR_MALE. |
|
Released_Genetics |
ResistanceHegGenetics |
1 |
100000000 |
10000 |
The genetic properties of the released mosquito, e.g. HEG and pesticide resistance. |
|
Released_Number |
integer |
1 |
100000000 |
10000 |
Number of mosquitos released in the intervention. |
|
Released_Species |
Constrained String |
1 |
100000000 |
10000 |
Name of released mosquito species. |
|
Released_Sterility |
enum |
NA |
NA |
VECTOR_FERTILE |
Sterility of released mosquitoes. Possible values are: VECTOR_FERTILE VECTOR_STERILE. |
|
Released_Wolbachia |
enum |
NA |
NA |
WOLBACHIA_FREE |
Wolbachia type of released mosquitoes. Possible values are: WOLBACHIA_FREE VECTOR_WOLBACHIA_A VECTOR_WOLBACHIA_B VECTOR_WOLBACHIA_AB. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Dont_Allow_Duplicates |
boolean |
0 |
1 |
0 |
If an individual’s container has an intervention, set to 1 to prevent them from receiving another copy of the intervention. Supported by all intervention types. |
|
Intervention_List |
IndividualIntervention |
0 |
1 |
0 |
The array of nested JSON parameters for the interventions to be distributed by this intervention. |
|
Intervention_Name |
string |
0 |
1 |
0 |
The optional name used to refer to this intervention as a means to differentiate it from others that use the same class. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Actual_IndividualIntervention_Config |
IndividualIntervention |
-1 |
3.40E+38 |
-1 |
The configuration of the actual individual-based intervention sought. Selects a class for the intervention and configures the parameters specific for that intervention class. |
|
Actual_NodeIntervention_Config |
NodeIntervention |
NA |
NA |
[] |
The configuration of the actual node-level intervention sought. Selects a class for the intervention and configures the parameters specific for that intervention class. |
|
Blackout_Event_Trigger |
Constrained String |
0 |
1 |
0 |
The event to broadcast if an intervention cannot be distributed due to the Blackout_Period. |
|
Blackout_Period |
float |
0 |
3.40E+38 |
0 |
After the initial intervention distribution, the time, in days, to wait before distributing the intervention again. If it cannot distribute due to the blackout period, it will broadcast the user-defined Blackout_Event_Trigger. |
|
Demographic_Coverage |
float |
0 |
1 |
1 |
Fraction of individuals in the target demographic that will receive this intervention. |
|
Duration |
float |
-1 |
3.40E+38 |
-1 |
The number of days to continue this intervention. |
|
Intervention_Name |
string |
0 |
1 |
0 |
The optional name used to refer to this intervention as a means to differentiate it from others that use the same class. |
|
Property_Restrictions |
Dynamic String Set |
NA |
NA |
[] |
Fraction of individuals in the target demographic that will receive this intervention. |
|
Property_Restrictions_Within_Node |
PropertyRestrictions |
0 |
3.40E+38 |
3.40E+38 |
Fraction of individuals in the target demographic that will receive this intervention. |
|
Target_Age_Max |
float |
0 |
3.40E+38 |
3.40E+38 |
Upper end of age targeted for intervention, in years. |
|
Target_Age_Min |
float |
0 |
3.40E+38 |
0 |
Lower end of age targeted for intervention, in years. |
|
Target_Demographic |
enum |
NA |
NA |
Everyone |
The demographic group (from a list of possibles) targeted by this intervention (e.g., Infants). |
|
Target_Gender |
enum |
NA |
NA |
All |
Specify the gender restriction for the intervention. Defaults to All. |
|
Target_Residents_Only |
boolean |
0 |
1 |
0 |
If true, only distribute the intervention to those that claim the node as their residence - i.e. started in the node |
|
Trigger_Condition |
Constrained String |
NA |
NA |
[] |
The condition for triggering a health seeking intervention. |
|
Trigger_Condition_List |
Vector String |
NA |
NA |
[] |
Included if Trigger_Condition is set to ‘TriggerList’. A list (JSON array) of triggers (strings) drawn from config.json ‘Listed_Events’ or from built-in events. |
|
Trigger_Condition_String |
Constrained String |
0 |
1 |
0 |
Included if Trigger_Condition is set to ‘TriggerString’. Normally Trigger_Condition must specify an event from the built-in list of pre-defined events or triggers from the IndividualEventTriggerType enum. Trigger_Condition_String can be any string provided it is listed in the config.json’s Listed_Events array. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Antigen |
integer |
0 |
10 |
0 |
The antigenic ID of the outbreak infection. |
|
Genome |
integer |
-1 |
16777200 |
0 |
The genetic ID of the outbreak infection. |
|
Import_Age |
float |
0 |
43800 |
365 |
Age (in days) of infected import cases |
|
Incubation_Period_Override |
boolean |
0 |
1 |
0 |
0 = outbreak will bypass incubation, 1 = outbreak infection will start from beginning. |
|
Intervention_Name |
string |
0 |
1 |
0 |
The optional name used to refer to this intervention as a means to differentiate it from others that use the same class. |
|
Number_Cases_Per_Node |
integer |
0 |
2147480000 |
1 |
Number of new cases of Outbreak to import (per node). This will increase population. There is no control over demographics of these individuals. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Antigen |
integer |
0 |
10 |
0 |
The antigenic ID of the outbreak infection. |
|
Dont_Allow_Duplicates |
boolean |
0 |
1 |
0 |
If an individual’s container has an intervention, set to 1 to prevent them from receiving another copy of the intervention. Supported by all intervention types. |
|
Genome |
integer |
-1 |
16777200 |
0 |
The genetic ID of the outbreak infection. |
|
Incubation_Period_Override |
boolean |
0 |
1 |
0 |
0 = outbreak will bypass incubation, 1 = outbreak infection will start from beginning. |
|
Intervention_Name |
string |
NA |
NA |
OFF |
The optional name used to refer to this intervention as a means to differentiate it from others that use the same class. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Cost_To_Consumer |
float |
0 |
999999 |
10 |
Unit cost per vector control (unamortized). |
|
Intervention_Name |
string |
0 |
1 |
0 |
The optional name used to refer to this intervention as a means to differentiate it from others that use the same class. |
|
Killing_Config |
WaningEffect |
0 |
999999 |
10 |
Killing of the targeted stage. A parameter that is decayed. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Cost_To_Consumer |
float |
0 |
999999 |
10 |
Unit cost per vector control (unamortized). |
|
Habitat_Target |
enum |
NA |
NA |
TEMPORARY_RAINFALL |
The oviposition habitat type targeted by oviposition traps. Possible values are: TEMPORARY_RAINFALL WATER_VEGETATION HUMAN_POPULATION CONSTANT BRACKISH_SWAMP ALL_HABITATS. |
|
Intervention_Name |
string |
0 |
1 |
0 |
The optional name used to refer to this intervention as a means to differentiate it from others that use the same class. |
|
Killing_Config |
WaningEffect |
NA |
NA |
TEMPORARY_RAINFALL |
Killing of the targeted stage. A parameter that is decayed. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Dont_Allow_Duplicates |
boolean |
0 |
1 |
0 |
If an individual’s container has an intervention, set to 1 to prevent them from receiving another copy of the intervention. Supported by all intervention types. |
|
Efficacy |
float |
0 |
1 |
0.5 |
Represents the efficacy of a PMTCT intervention, defined as the rate ratio of MTCT between women receiving the intervention and women not receiving the intervention. A setting of 1 is equivalent to 100% blocking efficacy, and 0 reverts to the default probability of transmission, configured through the config.json parameter Maternal_Transmission_Probability. |
|
Intervention_Name |
string |
0 |
3.40E+38 |
6 |
The optional name used to refer to this intervention as a means to differentiate it from others that use the same class. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Daily_Probability |
float |
0 |
1 |
1 |
The probability per day that an individual will move to the Target_Property_Value. |
|
Dont_Allow_Duplicates |
boolean |
0 |
1 |
0 |
If an individual’s container has an intervention, set to 1 to prevent them from receiving another copy of the intervention. Supported by all intervention types. |
|
Intervention_Name |
string |
NA |
NA |
Config |
The optional name used to refer to this intervention as a means to differentiate it from others that use the same class. |
|
Maximum_Duration |
float |
-1 |
3.40E+38 |
3.40E+38 |
Maximum duration of ‘delayed-release’ property change, in days. |
|
Revert |
float |
0 |
10000 |
0 |
Reversion rate, or the number of days until reversion, or a 2 if reversion is allowed. |
|
Target_Property_Key |
Constrained String |
0 |
10000 |
0 |
The name of the EMOD-defined property of the group. |
|
Target_Property_Value |
Constrained String |
0 |
10000 |
0 |
The value of the user-defined group where the individuals will be transitioned. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Antibody_Type |
enum |
NA |
NA |
CSP |
Antibody type. Possible values are: CSP MSP1 PfEMP1_minor PfEMP1_major. |
|
Antibody_Variant |
integer |
0 |
100000 |
0 |
Antibody variant index. |
|
Boosted_Antibody_Concentration |
float |
0 |
3.40E+38 |
1 |
Boosted antibody concentration (where unity equals maximum from natural exposure). |
|
Cost_To_Consumer |
float |
0 |
999999 |
3.75 |
Unit cost of RTS,S vaccination (unamortized). |
|
Dont_Allow_Duplicates |
boolean |
0 |
1 |
0 |
If an individual’s container has an intervention, set to 1 to prevent them from receiving another copy of the intervention. Supported by all intervention types. |
|
Intervention_Name |
string |
0 |
1 |
0 |
The optional name used to refer to this intervention as a means to differentiate it from others that use the same class. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Intervention_Name |
string |
0 |
1 |
0 |
The optional name used to refer to this intervention as a means to differentiate it from others that use the same class. |
|
Larval_Habitat_Multiplier |
LarvalHabitatMultiplier |
0 |
1 |
0 |
The value by which to scale the larval habitat availability specified in Larval_Habitat_Types across all habitat types, for specific habitat types, or for specific mosquito species within each habitat type. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Blocking_Config |
WaningEffect |
0 |
999999 |
8 |
The configuration of pre-feed mosquito repellency and waning for housing modification. |
|
Cost_To_Consumer |
float |
0 |
999999 |
8 |
Unit cost per housing modification (unamortized). |
|
Dont_Allow_Duplicates |
boolean |
0 |
1 |
0 |
If an individual’s container has an intervention, set to 1 to prevent them from receiving another copy of the intervention. Supported by all intervention types. |
|
Intervention_Name |
string |
-1 |
3.40E+38 |
-1 |
The optional name used to refer to this intervention as a means to differentiate it from others that use the same class. |
|
Killing_Config |
WaningEffect |
0 |
999999 |
8 |
The configuration of killing efficacy and waning for housing modification. Killing is conditional on NOT blocking the mosquito before feeding. |
The following example provides the syntax for configuring a SimpleBednet campaign.
{
"Events": [
{
"Event_Coordinator_Config": {
"Demographic_Coverage": 0.7,
"Intervention_Config": {
"Bednet_Type": "ITN",
"Cost_To_Consumer": 3.75,
"Killing_Config": {
"Initial_Effect": 0.6,
"Decay_Time_Constant": 1460,
"class": "WaningEffectExponential"
},
"Blocking_Config": {
"Initial_Effect": 0.9,
"Decay_Time_Constant": 730,
"class": "WaningEffectExponential"
},
"class": "SimpleBednet"
},
"Target_Demographic": "Everyone",
"class": "StandardInterventionDistributionEventCoordinator"
},
"Nodeset_Config": {
"class": "NodeSetAll"
},
"Start_Day": 1460,
"class": "CampaignEvent"
}
],
"Use_Defaults": 1
}
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Bednet_Type |
enum |
NA |
NA |
Barrier |
Type of bednet. Possible values are: Barrier, ITN, LLIN, Retreatment. Note: This enum does not affect the code at all. |
|
Blocking_Config |
WaningEffect |
0 |
999999 |
3.75 |
A fraction of feed attempts that have been blocked. The rate of blocking indoor feeds on individuals with an ITN. A parameter that is decayed. |
|
Cost_To_Consumer |
float |
0 |
999999 |
3.75 |
Unit cost per bednet (unamortized). |
|
Dont_Allow_Duplicates |
boolean |
0 |
1 |
0 |
If an individual’s container has an intervention, set to 1 to prevent them from receiving another copy of the intervention. Supported by all intervention types. |
|
Intervention_Name |
string |
0 |
1 |
1 |
The optional name used to refer to this intervention as a means to differentiate it from others that use the same class. |
|
Killing_Config |
WaningEffect |
0 |
999999 |
3.75 |
A fraction of feed attempts that kill the vector. The rate at which the mosquito dies conditional on a successfully blocked feed. A parameter that is decayed. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Base_Sensitivity |
float |
0 |
1 |
1 |
The sensitivity of the diagnostic. This sets the proportion of the time that individuals with the condition being tested receive a positive diagnostic test. When set to 1, the diagnostic always accurately reflects the condition. When set to zero, then individuals who have the condition always receive a false-negative diagnostic test. |
|
Base_Specificity |
float |
0 |
1 |
1 |
The specificity of the diagnostic. This sets the proportion of the time that individuals without the condition being tested receive a negative diagnostic test. When set to 1, the diagnostic always accurately reflects the lack of having the condition. When set to zero, then individuals who do not have the condition always receive a false-positive diagnostic test. |
|
Cost_To_Consumer |
float |
0 |
3.40E+38 |
1 |
The unit ‘cost’ assigned to the intervention. Setting Cost_To_Consumer to zero for all other interventions, and to a non-zero amount for one intervention, provides a convenient way to track the number of times the intervention has been applied in a simulation. |
|
Days_To_Diagnosis |
float |
0 |
3.40E+38 |
0 |
The number of days from test until diagnosis. |
|
Dont_Allow_Duplicates |
boolean |
0 |
1 |
0 |
If an individual’s container has an intervention, set to 1 to prevent them from receiving another copy of the intervention. Supported by all intervention types. |
|
Event_Or_Config |
enum |
NA |
NA |
Config |
Specifies whether the current intervention distributes a nested intervention (the Config option) or an event will be broadcast which may trigger other interventions in the campaign file (the Event option). See event list. |
|
Intervention_Name |
string |
0 |
1 |
1 |
The optional name used to refer to this intervention as a means to differentiate it from others that use the same class. |
|
Positive_Diagnosis_Config |
IndividualIntervention |
0 |
1 |
0 |
The intervention distributed to individuals if they test positive. This is only valid if Event_Or_Config is set to ‘Config’. |
|
Positive_Diagnosis_Event |
Constrained String |
0 |
3.40E+38 |
0 |
The next health care event that occurs for individuals who test negative. This is only valid if Event_Or_Config is set to ‘Event’. |
|
Treatment_Fraction |
float |
0 |
1 |
1 |
The fraction of positive diagnoses that are treated. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Actual_IndividualIntervention_Config |
IndividualIntervention |
0 |
1 |
1 |
The configuration of an actual intervention sought. Selects a class for the intervention and configures the parameters specific for that intervention class. |
|
Actual_IndividualIntervention_Event |
Constrained String |
0 |
999999 |
10 |
The event of an actual intervention sought. Selects a class for the intervention and configures the parameters specific for that intervention class. |
|
Dont_Allow_Duplicates |
boolean |
0 |
1 |
0 |
If an individual’s container has an intervention, set to 1 to prevent them from receiving another copy of the intervention. Supported by all intervention types. |
|
Event_Or_Config |
enum |
NA |
NA |
Config |
Specifies whether the current intervention distributes a nested intervention (the Config option) or an event will be broadcast which may trigger other interventions in the campaign file (the Event option). See event list. |
|
Intervention_Name |
string |
0 |
999999 |
10 |
The optional name used to refer to this intervention as a means to differentiate it from others that use the same class. |
|
Single_Use |
boolean |
0 |
1 |
1 |
One-time, or persistent? |
|
Tendency |
float |
0 |
1 |
1 |
The probability of seeking healthcare. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Blocking_Config |
WaningEffect |
0 |
999999 |
8 |
The configuration of efficacy and waning for individual repellent. |
|
Cost_To_Consumer |
float |
0 |
999999 |
8 |
Unit cost per repellent (unamortized). |
|
Dont_Allow_Duplicates |
boolean |
0 |
1 |
0 |
If an individual’s container has an intervention, set to 1 to prevent them from receiving another copy of the intervention. Supported by all intervention types. |
|
Intervention_Name |
string |
NA |
NA |
Config |
The optional name used to refer to this intervention as a means to differentiate it from others that use the same class. |
The following example provides the syntax for configuring a SimpleVaccine campaign.
{
"Events": [
{
"Event_Coordinator_Config": {
"Demographic_Coverage": 0.8,
"Intervention_Config": {
".Reduced_Acquire": 0.9,
"Vaccine_Take": 1,
"Vaccine_Type": "AcquisitionBlocking",
"Cost_To_Consumer": 10.0,
"class": "SimpleVaccine",
"Waning_Config": {
"Initial_Effect": 0.9,
"Box_Duration": 730,
"class": "WaningEffectBox"
}
},
"Target_Demographic": "Everyone",
"class": "StandardInterventionDistributionEventCoordinator"
},
"Nodeset_Config": {
"class": "NodeSetAll"
},
"Start_Day": 1460,
"class": "CampaignEvent"
}
],
"Use_Defaults": 1
}
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Cost_To_Consumer |
float |
0 |
999999 |
10 |
Unit cost per vaccine (unamortized). The unit ‘cost’ assigned to the intervention. Setting Cost_To_Consumer to zero for all other interventions, and to a non-zero amount for one intervention, provides a convenient way to track the number of times the intervention has been applied in a simulation. |
|
Dont_Allow_Duplicates |
boolean |
0 |
1 |
0 |
If an individual’s container has an intervention, set to 1 to prevent them from receiving another copy of the intervention. Supported by all intervention types. |
|
Intervention_Name |
string |
0 |
1 |
0 |
The optional name used to refer to this intervention as a means to differentiate it from others that use the same class. |
|
Vaccine_Take |
float |
0 |
1 |
1 |
This specifies the rate at which delivered vaccines will successfully stimulate an immune response and achieve the desired efficacy. If vaccine_take is set to 0.9, then with a 90 percent chance, the vaccine will start with the specified efficacy, and with a 10 percent chance it will have no efficacy at all. |
|
Vaccine_Type |
enum |
NA |
NA |
Generic |
Type of vaccine to distribute in a vaccine intervention. Possible values are: - Generic - The vaccine can have all of reduced_transmit, reduced_acquire, and reduced_mortality fields and the vaccine will have both effects. - TransmissionBlocking - Specifies the fraction by which the infectivity of the infected and vaccinated Individual is reduced (reduced_transmit). - AcquisitionBlocking - Specifies the fraction by which the force of infection experienced by the vaccinated Individual is reduced (reduced_acqu |
|
Waning_Config |
WaningEffect |
0 |
1 |
1 |
The configuration of drug-killing efficacy and waning. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Cost_To_Consumer |
float |
0 |
999999 |
10 |
Unit cost per vector control (unamortized). |
|
Habitat_Target |
enum |
NA |
NA |
TEMPORARY_RAINFALL |
The larval habitat type targeted for habitat reduction. Possible values are: TEMPORARY_RAINFALL WATER_VEGETATION HUMAN_POPULATION CONSTANT BRACKISH_SWAMP ALL_HABITATS. |
|
Intervention_Name |
string |
0 |
1 |
0 |
The optional name used to refer to this intervention as a means to differentiate it from others that use the same class. |
|
Killing_Config |
WaningEffect |
NA |
NA |
NA |
The configuration of killing efficacy and waning for space spaying. |
|
Reduction_Config |
WaningEffect |
NA |
NA |
NA |
The configuration of larval habitat reduction and waning for space spraying. |
|
Spray_Kill_Target |
enum |
NA |
NA |
SpaceSpray_FemalesOnly |
The gender kill target of space spraying. Possible values are: SpaceSpray_FemalesOnly SpaceSpray_MalesOnly SpaceSpray_FemalesAndMales. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Cost_To_Consumer |
float |
0 |
999999 |
10 |
Unit cost per vector control (unamortized). |
|
Intervention_Name |
string |
0 |
1 |
0 |
The optional name used to refer to this intervention as a means to differentiate it from others that use the same class. |
|
Repellency_Config |
WaningEffect |
0 |
999999 |
10 |
The configuration of efficacy and waning for spatial repellent. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Blocking_Config |
WaningEffect |
0 |
999999 |
8 |
The configuration of pre-feed mosquito repellency and waning for housing modification. |
|
Cost_To_Consumer |
float |
0 |
999999 |
8 |
Unit cost per housing modification (unamortized). |
|
Dont_Allow_Duplicates |
boolean |
0 |
1 |
0 |
If an individual’s container has an intervention, set to 1 to prevent them from receiving another copy of the intervention. Supported by all intervention types. |
|
Intervention_Name |
string |
0 |
1 |
0 |
The optional name used to refer to this intervention as a means to differentiate it from others that use the same class. |
|
Killing_Config |
WaningEffect |
0 |
999999 |
8 |
The configuration of killing efficacy and waning for housing modification. Killing is conditional on NOT blocking the mosquito before feeding. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Cost_To_Consumer |
float |
0 |
999999 |
10 |
Unit cost per vector control (unamortized). |
|
Intervention_Name |
string |
0 |
100000 |
0 |
The optional name used to refer to this intervention as a means to differentiate it from others that use the same class. |
|
Killing_Config |
WaningEffect |
0 |
999999 |
10 |
Killing of the targeted stage. A parameter that is decayed. |
Waning effect parameters are used with interventions such as vaccines, drugs, and bednets. It has six different classes. Each class specifies a different waning effect and uses a different combination of the waning effect parameters.
The waning effect parameters are a nested JSON object and can be configured using several different parameters such as Killing_Config, Waning_Config, Blocking_Config, and Mortality_Config. The following example uses Killing_Config. See the specific intervention sections for the supported parameters.
{
"Intervention_Config": {
"class": "Ivermectin",
"Cost_To_Consumer": 1,
"Killing_Config": {
"class": "WaningEffectBox",
"Box_Duration": 3,
"Initial_Effect": 1
}
}
}
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Box_Duration |
float |
0 |
100000 |
100 |
Box duration of effect in days. |
|
Initial_Effect |
float |
0 |
1 |
1 |
Initial strength of effect. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Box_Duration |
float |
0 |
100000 |
100 |
Box duration of effect in days. |
|
Decay_Time_Constant |
float |
0 |
100000 |
100 |
Exponential decay length in days. |
|
Initial_Effect |
float |
0 |
1 |
1 |
Initial strength of effect. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Initial_Effect |
float |
0 |
1 |
1 |
Initial strength of effect. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Decay_Time_Constant |
float |
0 |
100000 |
100 |
Exponential decay length in days. |
|
Initial_Effect |
float |
0 |
1 |
1 |
Initial strength of effect. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Expire_At_Durability_Map_End |
boolean |
0 |
1 |
0 |
Set to 1 to let the intervention expire when the end of the map is reached. |
|
Initial_Effect |
float |
0 |
1 |
1 |
Initial strength of effect. |
Parameter |
Data type |
Minimum |
Maximum |
Default |
Description |
Example |
|---|---|---|---|---|---|---|
Expire_At_Durability_Map_End |
boolean |
0 |
1 |
0 |
Set to 1 to let the intervention expire when the end of the map is reached. |
|
Initial_Effect |
float |
0 |
1 |
1 |
Initial strength of effect. |
Input and output file structure¶
The section describes the file structure used for input data files, each of the built-in output report types, and error logs to help debug issues you may encounter.
Input data file structure¶
The input data files described in this section determine the demographics, migration, climate, and other relatively fixed information about the population within each geographic node. These files are in contrast to the demographic, geographic, and migration parameters in the configuration file that control simulation-wide qualities, such as enabling air migration across all nodes in the simulation.
Generally, you will download and use input data files without modification. For instructions, see Use input data files. However, demographics files can include many user-defined parameter values and will likely require modification. Only the demographics file is required for a simulation.
Each type of input data generally requires both a metadata file that contains provenance information and a binary file that contains the actual data for each node. Some input files included with EMOD were prepared using CIESIN Gridded Population of the World (GPW) population distribution and a corresponding spatial resolution grid (for example, 2.5 arc minutes) to define the initial population and extent of the nodes for country-wide input files. Therefore, the naming convention for this files usually leads with the geographic location, followed by the spatial resolution, and input file type.
IdReference types and NodeID generation¶
All input files include the parameter IdReference in the metadata, which is used to generate the NodeID associated with each node in a simulation. There are currently four IdReference values that indicate the associated algorithm used for generating the NodeID for each node in a simulation. EMOD uses these values to compare different input data files in a simulation in order to match NodeID values. However, there is no reason you cannot assign any string you want to IdReference, provided all the input data files used in a simulation, such as demographics, climate, or migration, have the same IDReference value.
Generally, nodes are defined using a geographic grid laid over the world, forming approximately square nodes of a certain size. The NodeID is then calculated based on latitude and longitude of the node, where the location of the node is considered the upper left corner of the square it represents. NodeID values are numbered from south to north for latitude and west to east for longitude, with 0,0 being the intersection of the International Date Line and the South Pole. Note that NodeID values are <longitude><latitude> values (not <latitude><longitude> values as normally provided).
The three possible values for IdReference that use a grid to define and identify nodes differ based on grid resolution. EMOD is limited to a minimum node size of 30 arc-seconds (approximately 1 km at the equator). The three gridded values are:
Gridded world grump30arcsec
Gridded world grump2.5arcmin
Gridded world grump1degree
The algorithm for encoding latitude and longitude into a NodeID is as follows:
unsigned int xpix = math.floor((lon + 180.0) / resolution)
unsigned int ypix = math.floor((lat + 90.0) / resolution)
unsigned int **NodeID** = (xpix << 16) + ypix + 1
This generates a NodeID that is a 4-byte unsigned integer (like all NodeID values); the first two bytes represent the longitude of the node and the second two bytes represent the latitude. Note that as part of the final calculation, we add one to the NodeID. This is to reserve the NodeID of zero, since that is used as a null-value in a number of places (for example, in the migration files to signal that no migration value is present).
In addition to these gridded values, you can assign a value of “Legacy” to indicate that NodeID values are defined in and copied directly from the demographics files. This is used when the nodes are not demarcated by a grid, but rather by some social construct, such as voting precincts or other administrative boundary, that don’t necessarily have an obvious heuristic for enumeration. Some types of data may only be available or easier to obtain when in such a form. The name “Legacy” does not imply anything derogatory or time-related. It is simply a term that was chosen to represent data that isn’t related to the gridded information.
When using nodes that are not using the gridded format, you cannot assume or easily determine resolution, shape, or actual location of the node point within/along that shape. In other words, the format of such files depends on the creator of the file and you should refer to any information provided in the metadata of the file or by the creator when parsing the file.
Climate file structure¶
There are two general types of climate files usable by EMOD, namely, climate files generated through actual data, referred to as “climate by data,” and climate files generated from the Koppen classification system, referred to as “climate by Koppen”. Climate data (both types) is contained in a set of two files, a metadata file with header information (<name>.bin.json) and a binary data file (<name>.bin). Both files are required.
The metadata file is a JSON-formatted file that includes a metadata section and a node offsets section. The Metadata parameter contains a JSON object with parameters, some of which are strictly informational and some of which are used by Eradication.exe. However, the informational ones may still be important to understand the provenance and meaning of the data.
In a second section, the NodeOffsets parameter contains a list of hex-encoded 16-byte values used to find the data for each given node (the NodeID).They are not 16-byte offsets, but instead, two 8-byte hex-encoded character strings. This encoding includes the source NodeID. You can map the binary data to its corresponding source NodeID by using the NodeOffset information.
The binary file contains the climate data in a sequential stream. In other words, it presents all the data for the first node, then all the data for the second node, all the way through to the last node.
To use the climate files, you must set Climate_Model to either “CLIMATE_BY_DATA” or “CLIMATE_BY_KOPPEN”, as appropriate, in the configuration file. There are also additional parameters in the configuration file you can use to scale or otherwise modify the data included in the climate files.
At this time, the EMOD executable (Eradication.exe) reads land temperature data, but does not use the data in any calculations. IDM clones the air temperature and uses that as the land temperature in the climate data files. If you are going to be constructing your own climate files, we advise you to do the same.
The following parameters in the metadata section are informational:
Parameter |
Data type |
Description |
|---|---|---|
DateCreated |
string |
The day the file was created. |
Author |
string |
The author of the file. |
OriginalDataYears |
string |
The years from which the original data was derived. |
StartDayOfYear |
string |
The day of the year representing the first day in the climate file. |
DataProvenance |
string |
The source of the data. |
The following parameters in the metadata section are used by Eradication.exe:
Parameter |
Data type |
Description |
|---|---|---|
IdReference |
string |
A unique, user-selected string that indicates the method used for generating NodeID values in the input file. For more information, see Input data file structure. |
NodeCount |
integer |
The number of nodes to expect in this file. |
DatavalueCount |
integer |
The number of data values per node. The number must be the same across every node in the binary file. |
UpdateResolution |
enum |
The time resolution of the climate file. Available values are:
|
An example of climate by data metadata is as follows:
{
"Metadata": {
"DateCreated": "Sun Sep 25 19:02:09 2011",
"Tool": "createclimateheader.py",
"Author": "authorName",
"IdReference": "Gridded world grump2.5arcmin",
"NodeCount": 1,
"DatavalueCount": 3650,
"UpdateResolution": "CLIMATE_UPDATE_DAY",
"OriginalDataYears": "1990-1993",
"StartDayOfYear": "January 1",
"DataProvenance": "47 consecutive months of data were used to generate one average year of data that is repeated for 10 years"
},
"NodeOffsets": "144B07A400000000"
}
The binary file is a stream of 4-byte floating point values that contain the data value at the data count position for a given node, running from 1 to the maximum data count value.
The binary format is as follows:
The Koppen classification system is one of the most widely used climate classification systems. The Koppen classification system makes the assumption that native vegetation is the best expression of climate.
The following parameters in the metadata section are informational:
Parameter |
Data type |
Description |
|---|---|---|
DateCreated |
string |
The day the file was created. |
Author |
string |
The author of the file. |
DataProvenance |
string |
The source of the data. |
Tool |
string |
The script used to create the file. |
The following parameters in the metadata section are used by Eradication.exe:
Parameter |
Data type |
Description |
|---|---|---|
IdReference |
string |
A unique, user-selected string that indicates the method used for generating NodeID values in the input file. For more information, see Input data file structure. |
NodeCount |
integer |
The number of nodes to expect in this file. |
An example of climate by Koppen metadata is as follows:
{
"Metadata": {
"DateCreated": "Sun Sep 25 19:08:52 2011",
"Tool": "createclimateheader.py",
"Author": "authorName",
"IdReference": "Gridded world grump2.5arcmin",
"NodeCount": 2,
"DataProvenance": "Köppen-Geiger Classification System from http://koeppen-geiger.vu-wien.ac.at/"
},
"NodeOffsets": "157D075200000000157E07520000000"
}
The binary file parameters use the naming convention below to store the data.
Parameter |
Data type |
Description |
|---|---|---|
KoppenIndexX |
integer, 4 bytes |
The Koppen Index value, with X running from 1 to the maximum number of nodes. |
The binary format is as follows:
Migration file structure¶
There are four types of migration files that can be used by EMOD, namely, local migration, regional migration, air migration and sea migration. For all types, migration data is contained in a set of two files, a metadata file with header information and a binary data file. Both files are required. The structure of each of these files is nearly the same for all migration types, with the only exception being the number of columns in the binary file.
The metadata file is a JSON-formatted file that includes a metadata section and a node offsets section. The Metadata parameter contains a JSON object with parameters, some of which are strictly informational and some of which are used by Eradication.exe. However, the informational ones may still be important to understand the provenance and meaning of the data.
The following parameters in the metadata section are informational:
Parameter |
Data type |
Description |
|---|---|---|
DateCreated |
string |
The day the file was created. |
Author |
string |
The author of the file. |
Tool |
string |
The script used to create the file. |
The following parameters in the metadata section are used by Eradication.exe:
Parameter |
Data type |
Description |
|---|---|---|
IdReference |
string |
A unique, user-selected string that indicates the method used for generating NodeID values in the input file. For more information, see Input data file structure. |
NodeCount |
integer |
The number of nodes to expect in this file. |
DatavalueCount |
integer |
The number of data values per node. The number must be the same across every node in the binary file. |
In a second section, the NodeOffsets parameter contains a list of hex-encoded 16-byte values used to find the data for each given node (the NodeID).They are not 16-byte offsets, but instead, two 8-byte hex-encoded character strings. This encoding includes the source NodeID. You can map the binary data to its corresponding source NodeID by using the NodeOffset information.
The binary file contains the migration rate data in a sequential stream. In other words, it presents all the data for the first node, then all the data for the second node, all the way through to the last node. The data is laid out in rows and columns, with each row corresponding to a node and the number of columns varying based on the migration type.
The following diagram shows how migration data is laid out in rows, with each row corresponding to a source node and each column corresponding to the destination node.
Migration rate is that fraction of the node’s population that is migrating per day. Units are per person per day, meaning the number of people migrating per day divided by the total population of the node. For example, if a node had 1,000 people and a migration rate of .01, there would be .01 * 1,000 = 10 people migrating per day, but if the node had 10,000 people in the node and a migration rate of .01, it would .01 * 10,000 = 100 people migrating per day. You can adjust this rate using configuration parameters for scaling (these begin with x).
The binary file contains a stream of 4-byte unsigned integers that identify the destination node, followed by a stream of 8-byte double floating point values that contain the rate associated with the destination node (per person per day), running from 1 to the total number of nodes.
To use the migration files, you must set Migration_Model in the configuration file to a valid migration type except “NO_MIGRATION”. Each migration types also requires you to set another parameter to enable the particular type of migration selected. There are also additional parameters in the configuration file you can use to scale or otherwise modify the data included in the climate files.
Local migration describes the foot travel movement of people into and out of adjacent nodes. A local migration file is required for simulations that support more than one node. You must also set the Enable_Local_Migration parameter in the configuration file to 1. For each location, the local migration file represents up to eight adjacent destination nodes, with 0 used as the value for unused nodes.
The following diagram shows the format for the local migration binary file data:
Regional migration describes migration that occurs on a road or rail network for a simulation. If a node is not part of the network, the regional migration of individuals to and from that node considers the closest road hub city. A Voronoi tiling based on road hubs is constructed of the region, with each non-hub connected to the hub of its tile. These connections are created when the migration file is constructed. They are not performed at runtime.
A regional migration file is required for simulations that cover a geography that is large enough that road/rail migration is relevant. You must also set the Enable_Regional_Migration parameter in the configuration file to 1. For each location, the regional migration file represents up to 30 adjacent destination nodes, with 0 used as the value for unused nodes.
The following diagram shows the format for the regional migration binary file data:
Air migration describes migration that occurs by airplane travel. An air migration file is usually required for simulations of an entire country or larger geographies. You must also set the Enable_Air_Migration parameter in the configuration file to 1. For each location, the air migration file represents up to 60 adjacent destination nodes, with 0 used as the value for unused nodes.
The following diagram shows the format for the air migration binary file data:
Sea migration describes migration that occurs by ship travel. Unlike the other migration files, the sea migration file only contains information for the nodes that are seaports. A sea migration file does not contain every node like local, regional or air migration files do. You must also set the Enable_Sea_Migration parameter in the configuration file to 1. For each location, the sea migration file represents up to five adjacent destination nodes, with 0 used as the value for unused nodes.
The following diagram shows the format for the sea migration binary file data:
Load-balancing file structure¶
If you are running a large simulation with multiple nodes, you may want to use a load balancing file to distribute the computing load among multiple cores. This can be especially helpful if the nodes vary considerably in size and, therefore, processing time. If no load balancing file is submitted, the Eradication.exe allocates simulation nodes to cores according to a checkerboard pattern.
For each simulation, the load balancing file contains information about the relative level of processing time required for each geographical node. The Eradication.exe can then allocate nodes to processors in such a way that the total processing required is evenly distributed across all processors.
To use a load balancing file, you must set Load_Balance_Filename to the path to the load balancing file, relative to the input file directory. Load-balancing files can be JavaScript Object Notation (JSON) files, which are preferred, or binary files.
A JSON load-balancing file contains the Load_Balance_Scheme_Nodes_On_Core_Matrix parameter, assigned an array of arrays, each of which represents a core and lists the NodeID of each node to be processed on that core. If any node contained in the demographics file is not listed in the load-balancing file, it will be processed on the first core.
For example, the load-balancing file shown below distributes the processing for 100 nodes across 4 cores, assigning more nodes to particular cores when those nodes require less processing time.
{
"Load_Balance_Scheme_Nodes_On_Core_Matrix": [
[1, 2, 3, 4, 5, 11, 12, 13, 14, 15, 21, 22, 23, 24, 25, 31, 32, 33, 34, 35, 45],
[6, 7, 8, 9, 10, 16, 17, 18, 19, 20, 26, 27, 28, 29, 30, 36, 37, 38, 39, 40, 46, 47, 48, 49, 50, 56, 57, 58, 59, 60, 67, 68, 69, 70],
[61, 62, 63, 64, 65, 66, 71, 72, 73, 74, 75, 76, 77, 81, 82, 83, 84, 85, 86, 87, 91, 92, 93, 94, 95, 96, 97],
[41, 42, 43, 44, 51, 52, 53, 54, 55, 78, 79, 80, 88, 89, 90, 98, 99, 100]
]
}
A binary load-balancing file starts with an initial unsigned 4-byte integer that indicates the number of nodes. Following that value is a series of 4-byte unsigned integers representing the NodeID values. The number of values will be equal to the previously read number of nodes. Following that, another series of 4-byte floating point values, with each value representing the relative processing time required for each geographic node. Again, the number of values will be equal to the previously read number of nodes. The series of values are set up such that the \(i^{th}\) entry in the series is equal to the cumulative proportion of the processing load for all the previous nodes 0 through \(i\).
The cumulative nature of each node’s value does not mean each node is assigned that amount for its processing load. Rather, it is a way to make the internal calculations more efficient. For example, if you had ten nodes and each node was assigned 10% of the load, the values assigned from node0 to node9 would be the following: 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9 and 1.0.
The following diagram shows the format for the binary load-balancing file data:
Output report structure¶
This topic defines various data output reports that the EMOD executable (Eradication.exe) can produce. These reports are built in to the Eradication.exe and are either always output with each simulation, such as inset chart, or reports that you can configure to be included with the output. All output reports can be found in the output directory of the working directory, for example, 1_Generic_Seattle_MultiNodeoutput.
If the available output reports do not meet your needs, you can use custom reporters that are EMODules separate from the Eradication.exe code. There are several provided in the EMOD GitHub repository or you can even write your own custom reporter to post- process and format data as you desire into a new output report. For more information, see Custom reporters.
Inset chart output report¶
The inset chart output report is output with every simulation. It is a JSON-formatted file with the channel output results of the simulation, consisting of simulation-wide averages by time step. The channels in an inset chart are fully specified by the simulation type and cannot be altered without making changes to the EMOD source code. The file name is InsetChart.json. You can use the information contained in the file to create a chart, but EMOD does not automatically output a chart.
The file contains a header and a channels section.
Header¶
The header section contains the following parameters.
Parameter |
Data type |
Description |
|---|---|---|
DateTime |
string |
The time stamp indicating when the report was generated. |
DTK_Version |
string |
The version of EMOD used. |
Report_Type |
string |
The type of output report. |
Report_Version |
string |
The format version of the report. |
Start_Time |
integer |
The time noted in days when the simulation begins. |
Simulation_Timestep |
integer |
The number of days in each time step. |
Timesteps |
integer |
The number of time steps in this simulation. |
Channels |
integer |
The number of channels in the simulation. |
Channels¶
The channels section contains the following parameters.
Parameter |
Data type |
Description |
|---|---|---|
<Channel_Title> |
string |
The title of the particular channel. |
Units |
string |
The units used for this channel. |
Data |
array |
A list of the channel data at each time step. |
Example¶
The following is a sample of an InsetChart.json file.
{
"Header": {
"DateTime": "Thu Dec 03 11:28:34 2015",
"DTK_Version": "5738 HIV-Ongoing 2015/11/18 13:00:39",
"Report_Type": "InsetChart",
"Report_Version": "3.2",
"Start_Time": 0,
"Simulation_Timestep": 1,
"Timesteps": 5,
"Channels": 17
},
"Channels": {
"Births": {
"Units": "Births",
"Data": [
0,
0,
0,
0,
0
]
},
"Campaign Cost": {
"Units": "USD",
"Data": [
0,
0,
0,
0,
0
]
},
"Cumulative Infections": {
"Units": "",
"Data": [
0,
0,
0,
0,
0
]
},
"Cumulative Reported Infections": {
"Units": "",
"Data": [
0,
0,
0,
0,
0
]
},
"Disease Deaths": {
"Units": "",
"Data": [
0,
0,
0,
0,
0
]
},
"Exposed Population": {
"Units": "Exposed Fraction",
"Data": [
0,
0,
0,
0,
0
]
},
"Human Infectious Reservoir": {
"Units": "Total Infectivity",
"Data": [
0,
0,
0,
0,
0
]
},
"Infected": {
"Units": "Infected Fraction",
"Data": [
0,
0,
0,
0,
0
]
},
"Infectious Population": {
"Units": "Infectious Fraction",
"Data": [
0,
0,
0,
0,
0
]
},
"Log Prevalence": {
"Units": "Log Prevalence",
"Data": [-10, -10, -10, -10, -10]
},
"New Infections": {
"Units": "",
"Data": [
0,
0,
0,
0,
0
]
},
"New Reported Infections": {
"Units": "",
"Data": [
0,
0,
0,
0,
0
]
},
"Daily (Human) Infection Rate": {
"Units": "Infection Rate",
"Data": [
0,
0,
0,
0,
0
]
},
"Recovered Population": {
"Units": "Recovered (Immune) Fraction",
"Data": [
0,
0,
0,
0,
0
]
},
"Statistical Population": {
"Units": "Population",
"Data": [
900,
900,
900,
900,
900
]
},
"Susceptible Population": {
"Units": "Susceptible Fraction",
"Data": [
1,
1,
1,
1,
1
]
},
"Waning Population": {
"Units": "Waning Immunity Fraction",
"Data": [
0,
0,
0,
0,
0
]
}
}
}
Binned output report¶
The binned output report is a JSON-formatted file where the channel data has been sorted into bins. It is very similar to an inset chart. However, with the binned report, channels are broken down into sub-channels (bins) based on various criteria. For example, instead of having a single prevalence channel, you might have prevalence in the “0-3 years old bin” and the “4-6 years old bin, and so forth. The file name is BinnedReport.json.
To generate the binned report, set the Enable_Demographics_Reporting configuration parameter to 1. The demographics summary output report will also be generated.
The file contains a header and a channels section.
Header¶
The header section contains the following parameters.
Parameter |
Data type |
Description |
|||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
DateTime |
string |
The time stamp indicating when the report was generated. |
|||||||||||||||
DTK_Version |
string |
The version of EMOD used. |
|||||||||||||||
Report_Type |
string |
The type of output report. |
|||||||||||||||
Report_Version |
string |
The format version of the report. |
|||||||||||||||
Timesteps |
integer |
The number of time steps in this simulation. |
|||||||||||||||
Channels |
integer |
The number of channels in the simulation. |
|||||||||||||||
Subchannel_Metadata |
nested JSON object |
Metadata that describes the bins and axis information. The metadata includes the following parameters:
|
Channels¶
The channels section contains the following parameters.
Parameter |
Data type |
Description |
|---|---|---|
<Channel_Title> |
string |
The title of the particular channel. |
Units |
string |
The units used for this channel. |
Data |
array |
A list of the channel data at each time step. |
Example¶
The following is a sample of an BinnedReport.json file.
{
"Header": {
"DateTime": "Wed Oct 01 14:49:22 2014",
"DTK_Version": "4284 trunk 2014/09/23 11:50:52",
"Report_Version": "2.1",
"Timesteps": 150,
"Subchannel_Metadata": {
"AxisLabels": [
["Age"]
],
"NumBinsPerAxis": [
[21]
],
"ValuesPerAxis": [
[
[1825, 3650, 5475, 7300, 9125, 10950, 12775, 14600, 16425, 18250, 20075, 21900, 23725, 25550, 27375, 29200, 31025, 32850, 34675, 36500, 999999]
]
],
"MeaningPerAxis": [
[
["<5", "5-9", "10-14", "15-19", "20-24", "25-29", "30-34", "35-39", "40-44", "45-49", "50-54", "55-59", "60-64", "65-69", "70-74", "75-79", "80-84", "85-89", "90-94", "95-99", ">100"]
]
]
},
"Channels": 4
},
"Channels": {
"Disease Deaths": {
"Units": "",
"Data": [
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]
]
},
"Infected": {
"Units": "",
"Data": [
[0, 2, 1, 1, 1, 1, 2, 6, 7, 5, 3, 5, 4, 3, 2, 3, 5, 6, 14, 21, 22, 22, 31, 52, 63, 101, 108, 119, 121, 139, 132, 133, 130, 115, 97, 95, 93, 98, 102, 108, 105, 110, 98, 90, 93, 94, 92, 81, 62, 57, 63, 56, 58, 61, 67, 64, 52, 54, 45, 33, 34, 36, 27, 32, 29, 21, 17, 7, 13, 15, 17, 13, 14, 11, 14, 15, 11, 12, 7, 6, 6, 3, 3, 3, 4, 4, 3, 3, 2, 2, 3, 3, 4, 2, 2, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 3, 4, 2, 2, 2, 1, 5, 5, 7, 7, 6, 13, 14, 12, 8, 8, 15, 15, 23, 39, 43, 61, 82, 110, 133, 142, 133, 113, 105, 105, 100, 86, 81, 81, 75, 73, 68, 74, 74, 73, 70, 75, 67, 63, 72, 61, 63, 60, 55, 52, 47, 44, 36, 39, 36, 39, 36, 28, 27, 22, 20, 17, 20, 22, 17, 13, 13, 10, 10, 6, 7, 7, 3, 4, 5, 5, 4, 2, 1, 2, 2, 2, 3, 2, 2, 4, 2, 1, 2, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 1, 1, 2, 3, 3, 4, 3, 3, 3, 2, 2, 3, 5, 5, 7, 7, 12, 15, 21, 24, 22, 23, 33, 36, 42, 58, 93, 95, 104, 105, 102, 83, 75, 72, 67, 66, 63, 68, 61, 61, 63, 62, 60, 67, 67, 55, 52, 55, 42, 43, 40, 43, 36, 37, 30, 33, 33, 29, 23, 20, 15, 15, 17, 16, 14, 13, 13, 12, 15, 13, 11, 7, 13, 12, 11, 10, 9, 6, 5, 7, 4, 7, 4, 2, 2, 2, 1, 0, 0, 0, 1, 0, 1, 2, 2, 2, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 1, 0, 0, 0, 0, 0, 1, 2, 3, 4, 2, 3, 2, 4, 6, 7, 10, 12, 13, 14, 16, 18, 27, 31, 50, 50, 48, 58, 68, 86, 81, 82, 76, 67, 58, 58, 59, 63, 57, 47, 40, 48, 43, 46, 45, 46, 46, 44, 40, 39, 40, 40, 37, 39, 36, 29, 26, 20, 22, 23, 20, 15, 13, 16, 16, 14, 12, 10, 7, 9, 8, 6, 6, 5, 3, 3, 3, 2, 1, 1, 1, 1, 1, 2, 3, 3, 2, 2, 1, 1, 0, 1, 2, 3, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 2, 1, 2, 2, 2, 2, 4, 5, 4, 5, 4, 2, 4, 9, 11, 12, 16, 15, 20, 22, 28, 40, 44, 44, 48, 54, 54, 50, 53, 52, 51, 48, 47, 43, 44, 40, 40, 43, 36, 34, 30, 34, 32, 29, 26, 27, 24, 16, 22, 21, 18, 18, 17, 17, 15, 12, 14, 9, 12, 5, 6, 3, 6, 7, 6, 8, 9, 8, 7, 7, 6, 5, 4, 3, 5, 5, 4, 3, 3, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 1, 1, 1, 2, 2, 2, 1, 1, 1, 0, 0, 1, 2, 3, 1, 2, 2, 7, 7, 9, 10, 10, 19, 22, 34, 44, 42, 45, 62, 63, 48, 48, 48, 46, 42, 37, 33, 33, 33, 35, 34, 37, 34, 33, 30, 28, 32, 33, 24, 27, 27, 23, 25, 18, 12, 13, 12, 12, 11, 9, 5, 6, 7, 8, 10, 7, 7, 5, 6, 5, 5, 5, 4, 6, 5, 2, 3, 5, 4, 4, 1, 1, 1, 2, 2, 2, 2, 2, 1, 2, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 1, 2, 3, 2, 5, 11, 16, 20, 25, 25, 26, 36, 53, 49, 48, 51, 54, 47, 44, 35, 30, 21, 20, 22, 24, 26, 28, 22, 16, 19, 31, 26, 30, 25, 30, 26, 19, 22, 21, 18, 21, 23, 15, 11, 11, 11, 9, 6, 4, 2, 4, 4, 3, 2, 2, 4, 3, 3, 4, 3, 2, 1, 1, 0, 0, 3, 3, 2, 4, 3, 2, 1, 1, 1, 1, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 2, 1, 3, 2, 1, 1, 2, 3, 10, 13, 12, 12, 17, 26, 24, 23, 25, 30, 33, 28, 35, 34, 27, 22, 22, 21, 22, 19, 18, 16, 13, 17, 13, 14, 11, 14, 11, 11, 10, 12, 14, 13, 13, 15, 12, 9, 7, 6, 5, 5, 5, 6, 6, 7, 8, 6, 6, 5, 4, 4, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 2, 3, 3, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 2, 2, 1, 1, 2, 2, 0, 0, 0, 0, 2, 3, 5, 7, 5, 4, 10, 8, 10, 11, 18, 18, 21, 27, 27, 26, 24, 20, 24, 24, 21, 20, 21, 20, 19, 21, 18, 21, 13, 12, 14, 15, 13, 12, 14, 15, 13, 9, 8, 7, 6, 4, 8, 6, 4, 3, 4, 5, 5, 3, 1, 1, 1, 1, 0, 1, 1, 1, 3, 2, 3, 2, 1, 2, 0, 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 2, 2, 2, 3, 5, 7, 7, 9, 13, 13, 16, 20, 18, 22, 26, 26, 23, 23, 24, 20, 13, 7, 7, 11, 9, 12, 13, 16, 19, 20, 14, 11, 12, 11, 10, 11, 8, 6, 6, 4, 4, 2, 6, 6, 4, 4, 3, 3, 3, 6, 6, 6, 4, 4, 3, 4, 2, 3, 5, 3, 4, 2, 2, 4, 4, 5, 3, 3, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 1, 1, 0, 0, 1, 2, 0, 0, 0, 1, 1, 2, 2, 2, 2, 1, 0, 1, 2, 3, 3, 5, 5, 4, 6, 15, 16, 18, 22, 22, 22, 23, 12, 8, 5, 6, 4, 5, 6, 9, 10, 11, 13, 17, 14, 11, 13, 13, 9, 7, 6, 11, 9, 7, 4, 4, 3, 2, 2, 3, 1, 4, 3, 2, 1, 2, 4, 2, 2, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 1, 1, 1, 3, 2, 3, 2, 2, 4, 3, 8, 11, 10, 12, 10, 6, 6, 12, 9, 5, 6, 6, 9, 8, 6, 8, 7, 8, 12, 11, 11, 7, 14, 13, 12, 11, 12, 6, 8, 10, 12, 8, 4, 4, 2, 3, 3, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 1, 1, 2, 1, 0, 0, 0, 1, 2, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 0, 1, 3, 4, 6, 9, 9, 11, 12, 12, 11, 12, 12, 9, 9, 6, 7, 8, 10, 10, 11, 4, 8, 7, 10, 9, 9, 10, 8, 6, 4, 3, 2, 2, 3, 3, 5, 7, 4, 4, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 2, 2, 2, 4, 4, 4, 4, 6, 4, 8, 9, 7, 6, 7, 6, 6, 5, 5, 4, 2, 4, 4, 4, 4, 4, 6, 5, 3, 2, 3, 4, 2, 1, 2, 4, 6, 6, 6, 5, 2, 3, 4, 2, 1, 3, 3, 3, 3, 2, 2, 0, 0, 0, 0, 0, 1, 1, 2, 2, 2, 2, 1, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 2, 1, 2, 2, 4, 5, 10, 9, 10, 10, 10, 8, 10, 9, 9, 8, 7, 7, 4, 5, 5, 6, 4, 3, 4, 2, 4, 4, 4, 5, 4, 2, 3, 2, 4, 2, 2, 2, 0, 0, 1, 2, 1, 1, 1, 1, 0, 0, 1, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 2, 2, 1, 2, 2, 1, 1, 2, 3, 4, 3, 4, 6, 6, 3, 4, 2, 1, 2, 3, 3, 4, 2, 2, 4, 6, 4, 3, 2, 2, 2, 3, 4, 5, 4, 5, 4, 4, 3, 5, 5, 4, 4, 3, 3, 3, 2, 3, 3, 3, 3, 3, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 1, 3, 6, 8, 8, 5, 5, 5, 6, 3, 4, 4, 5, 6, 6, 6, 4, 2, 2, 2, 2, 4, 2, 3, 3, 5, 6, 6, 6, 4, 3, 2, 2, 0, 2, 1, 1, 0, 0, 0, 0, 2, 2, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 4, 6, 7, 7, 7, 6, 5, 5, 4, 4, 4, 1, 1, 1, 1, 2, 1, 1, 0, 0, 1, 2, 1, 2, 3, 2, 1, 2, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 2, 1, 1, 1, 2, 5, 4, 5, 3, 3, 1, 1, 1, 1, 1, 0, 1, 1, 0, 0, 2, 2, 1, 2, 2, 2, 3, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 2, 2, 2, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 3, 3, 4, 3, 4, 4, 5, 5, 3, 1, 1, 1, 1, 3, 2, 2, 2, 1, 1, 2, 2, 2, 2, 2, 2, 1, 2, 2, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 0, 1, 2, 2, 2, 3, 9, 9, 10, 15, 12, 8, 8, 11, 12, 12, 9, 6, 5, 9, 8, 10, 11, 11, 7, 5, 4, 5, 5, 5, 3, 0, 1, 3, 6, 7, 4, 4, 2, 1, 1, 2, 2, 2, 1, 1, 0, 1, 1, 1, 2, 2, 2, 2, 0, 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]
]
},
"New Infections": {
"Units": "",
"Data": [
[0, 2, 0, 0, 0, 0, 1, 4, 1, 0, 0, 3, 1, 1, 1, 2, 3, 1, 11, 14, 5, 9, 15, 30, 28, 50, 34, 47, 42, 48, 46, 36, 37, 17, 25, 33, 22, 31, 33, 37, 28, 36, 21, 21, 27, 26, 28, 15, 14, 20, 21, 15, 17, 19, 23, 17, 7, 15, 10, 10, 7, 14, 2, 13, 10, 1, 2, 0, 8, 4, 6, 1, 4, 3, 4, 4, 1, 3, 2, 1, 3, 0, 1, 1, 1, 1, 0, 1, 1, 0, 2, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 3, 1, 0, 0, 1, 0, 5, 0, 3, 1, 2, 7, 6, 4, 3, 2, 11, 9, 15, 19, 16, 30, 34, 46, 51, 49, 39, 27, 32, 37, 22, 23, 19, 25, 17, 28, 15, 25, 28, 25, 22, 24, 17, 15, 25, 13, 24, 17, 14, 17, 12, 14, 11, 19, 9, 11, 10, 5, 8, 5, 6, 3, 6, 5, 4, 2, 6, 1, 2, 1, 2, 2, 0, 2, 2, 2, 1, 0, 1, 1, 0, 1, 1, 0, 1, 2, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 0, 3, 1, 5, 6, 8, 6, 9, 7, 19, 12, 19, 30, 46, 25, 35, 35, 33, 21, 23, 19, 20, 12, 21, 19, 15, 19, 16, 14, 20, 17, 17, 6, 15, 14, 14, 14, 12, 11, 7, 10, 6, 9, 8, 7, 3, 4, 3, 5, 5, 4, 6, 7, 3, 3, 6, 2, 2, 2, 7, 0, 2, 2, 2, 2, 1, 2, 0, 4, 0, 0, 2, 0, 0, 0, 0, 0, 1, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 1, 0, 0, 0, 0, 0, 1, 1, 2, 1, 0, 1, 1, 2, 2, 2, 4, 3, 5, 4, 5, 8, 13, 10, 23, 18, 19, 23, 26, 32, 26, 21, 15, 9, 16, 17, 17, 16, 17, 13, 9, 19, 10, 15, 11, 15, 12, 13, 10, 16, 13, 9, 13, 13, 4, 6, 7, 6, 8, 2, 4, 2, 5, 5, 4, 3, 5, 3, 1, 5, 0, 0, 2, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 1, 2, 1, 0, 1, 0, 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 2, 0, 1, 1, 0, 0, 2, 1, 1, 2, 0, 0, 2, 6, 2, 4, 6, 6, 6, 10, 12, 20, 16, 15, 20, 16, 18, 14, 15, 14, 12, 13, 12, 9, 13, 10, 13, 11, 9, 9, 6, 15, 9, 9, 7, 8, 7, 6, 8, 6, 6, 3, 5, 5, 4, 4, 4, 0, 4, 0, 3, 0, 3, 2, 1, 3, 4, 0, 0, 3, 1, 2, 1, 1, 2, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 0, 1, 1, 5, 2, 2, 3, 5, 10, 9, 17, 15, 14, 14, 26, 20, 11, 14, 10, 14, 10, 8, 11, 9, 12, 13, 7, 12, 11, 9, 11, 8, 12, 7, 7, 9, 6, 6, 5, 2, 4, 3, 3, 7, 2, 0, 2, 2, 3, 4, 3, 0, 1, 0, 1, 1, 2, 2, 0, 3, 1, 1, 1, 2, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 1, 1, 3, 6, 6, 6, 6, 9, 12, 17, 25, 14, 14, 17, 12, 8, 12, 10, 9, 2, 8, 5, 11, 9, 7, 4, 2, 11, 15, 4, 11, 4, 9, 6, 3, 12, 4, 2, 6, 6, 3, 2, 4, 2, 1, 1, 0, 0, 3, 0, 0, 0, 0, 2, 1, 1, 1, 0, 0, 1, 0, 0, 0, 3, 0, 0, 2, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 0, 0, 2, 0, 0, 0, 1, 1, 8, 3, 2, 5, 7, 13, 6, 9, 8, 10, 10, 4, 13, 8, 8, 5, 6, 9, 6, 5, 5, 4, 5, 7, 2, 6, 3, 7, 2, 5, 2, 5, 5, 5, 4, 5, 2, 1, 2, 1, 2, 1, 1, 1, 1, 2, 3, 0, 1, 2, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 2, 0, 0, 0, 1, 0, 0, 0, 0, 0, 2, 2, 2, 2, 1, 1, 7, 0, 5, 3, 10, 6, 8, 12, 3, 10, 5, 6, 10, 8, 4, 4, 5, 6, 2, 4, 3, 7, 2, 1, 6, 5, 5, 4, 6, 3, 3, 3, 3, 1, 1, 1, 4, 1, 1, 1, 2, 1, 0, 0, 0, 1, 0, 0, 0, 1, 1, 0, 2, 0, 2, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 1, 1, 2, 3, 3, 2, 4, 4, 6, 7, 8, 10, 8, 6, 7, 4, 4, 1, 3, 0, 7, 4, 5, 6, 5, 4, 6, 7, 3, 4, 4, 2, 2, 3, 0, 1, 1, 0, 2, 0, 4, 2, 1, 0, 1, 1, 1, 3, 0, 1, 1, 0, 1, 1, 0, 1, 2, 0, 2, 0, 1, 2, 1, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 1, 0, 0, 0, 1, 1, 0, 0, 0, 1, 0, 1, 0, 1, 0, 0, 0, 1, 1, 2, 1, 2, 1, 3, 4, 9, 5, 7, 7, 3, 6, 5, 1, 1, 2, 2, 2, 2, 2, 3, 3, 4, 5, 7, 0, 2, 4, 2, 0, 1, 1, 7, 1, 1, 1, 2, 0, 0, 1, 1, 0, 3, 0, 0, 0, 1, 2, 1, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 1, 0, 1, 2, 0, 1, 1, 0, 3, 2, 5, 4, 3, 5, 2, 2, 4, 6, 1, 1, 1, 2, 3, 3, 1, 3, 1, 3, 5, 4, 3, 1, 8, 1, 4, 1, 3, 0, 4, 3, 3, 3, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 2, 1, 2, 5, 3, 6, 3, 4, 4, 5, 3, 2, 2, 1, 2, 2, 4, 2, 4, 0, 5, 4, 4, 1, 3, 4, 1, 1, 0, 0, 0, 1, 2, 1, 2, 3, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 1, 0, 1, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 1, 3, 0, 2, 0, 4, 1, 6, 1, 1, 1, 3, 1, 1, 2, 2, 2, 0, 2, 1, 1, 1, 1, 2, 0, 0, 1, 1, 2, 0, 0, 1, 2, 2, 1, 1, 1, 0, 1, 1, 0, 0, 2, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 2, 2, 7, 2, 2, 3, 3, 1, 2, 3, 2, 3, 0, 2, 1, 1, 2, 1, 1, 1, 2, 0, 2, 0, 1, 2, 0, 0, 2, 0, 2, 1, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 1, 1, 0, 1, 0, 0, 1, 1, 2, 1, 1, 3, 2, 2, 0, 2, 0, 0, 2, 1, 1, 2, 0, 0, 3, 2, 0, 1, 0, 1, 1, 1, 2, 1, 1, 1, 0, 1, 0, 2, 1, 1, 0, 0, 1, 0, 1, 1, 1, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 2, 3, 2, 2, 0, 2, 0, 2, 0, 1, 0, 2, 1, 0, 2, 0, 0, 1, 0, 0, 4, 0, 1, 1, 2, 3, 2, 1, 2, 0, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 1, 2, 2, 2, 1, 1, 0, 2, 0, 1, 1, 0, 0, 1, 1, 1, 0, 0, 0, 0, 1, 2, 0, 1, 2, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 1, 1, 3, 0, 1, 1, 0, 0, 0, 1, 0, 0, 0, 1, 0, 0, 0, 2, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 3, 0, 1, 1, 1, 1, 2, 0, 0, 1, 0, 0, 1, 2, 0, 0, 1, 0, 0, 1, 0, 1, 0, 0, 1, 0, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 1, 1, 0, 1, 2, 6, 3, 3, 8, 3, 0, 3, 4, 3, 4, 0, 2, 0, 4, 1, 3, 1, 3, 1, 0, 1, 1, 1, 1, 0, 0, 1, 2, 3, 1, 1, 1, 0, 0, 0, 1, 1, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]
]
},
"Population": {
"Units": "",
"Data": [
[1908, 1908, 1905, 1905, 1904, 1900, 1899, 1897, 1896, 1895, 1894, 1894, 1893, 1892, 1892, 1891, 1890, 1890, 1888, 1888, 1886, 1886, 1886, 1885, 1884, 1884, 1884, 1883, 1881, 1881, 1881, 1881, 1881, 1881, 1881, 1880, 1878, 1877, 1876, 1876, 1875, 1874, 1871, 1870, 1867, 1866, 1865, 1864, 1863, 1863, 1863, 1861, 1861, 1861, 1861, 1861, 1859, 1859, 1859, 1859, 1859, 1858, 1857, 1855, 1855, 1853, 1852, 1852, 1852, 1851, 1851, 1850, 1849, 1849, 1848, 1848, 1847, 1846, 1846, 1846, 1846, 1845, 1844, 1842, 1842, 1841, 1840, 1838, 1836, 1835, 1832, 1830, 1827, 1826, 1826, 1826, 1825, 1825, 1823, 1821, 1821, 1820, 1820, 1819, 1819, 1818, 1818, 1818, 1817, 1816, 1816, 1815, 1815, 1813, 1811, 1811, 1811, 1811, 1810, 1810, 1809, 1807, 1805, 1803, 1803, 1803, 1799, 1797, 1794, 1794, 1792, 1792, 1792, 1791, 1790, 1789, 1787, 1787, 1786, 1786, 1786, 1785, 1784, 1783, 1782, 1779, 1777, 1776, 1775, 1774],
[1592, 1591, 1592, 1592, 1592, 1596, 1596, 1597, 1597, 1598, 1599, 1597, 1598, 1598, 1597, 1598, 1599, 1599, 1601, 1600, 1601, 1600, 1599, 1599, 1600, 1600, 1599, 1600, 1602, 1602, 1602, 1602, 1601, 1599, 1599, 1599, 1601, 1601, 1601, 1600, 1601, 1602, 1604, 1604, 1607, 1607, 1608, 1607, 1606, 1606, 1605, 1607, 1607, 1605, 1605, 1605, 1605, 1604, 1604, 1602, 1599, 1599, 1599, 1601, 1600, 1600, 1601, 1600, 1598, 1598, 1597, 1597, 1597, 1596, 1595, 1593, 1592, 1593, 1592, 1591, 1591, 1590, 1591, 1592, 1592, 1592, 1593, 1594, 1596, 1597, 1598, 1598, 1601, 1602, 1602, 1601, 1602, 1601, 1602, 1604, 1603, 1604, 1603, 1603, 1603, 1602, 1601, 1600, 1600, 1600, 1597, 1598, 1597, 1598, 1599, 1598, 1598, 1598, 1598, 1597, 1598, 1599, 1599, 1600, 1599, 1595, 1599, 1601, 1604, 1602, 1604, 1604, 1602, 1602, 1602, 1602, 1603, 1601, 1602, 1602, 1602, 1602, 1602, 1602, 1602, 1605, 1606, 1606, 1605, 1605],
[1252, 1252, 1254, 1252, 1252, 1252, 1253, 1253, 1254, 1254, 1253, 1255, 1253, 1253, 1253, 1253, 1253, 1253, 1252, 1251, 1252, 1252, 1253, 1252, 1252, 1252, 1253, 1253, 1253, 1253, 1252, 1252, 1253, 1254, 1253, 1253, 1251, 1250, 1251, 1252, 1250, 1248, 1248, 1247, 1247, 1248, 1247, 1247, 1247, 1247, 1248, 1247, 1246, 1248, 1248, 1248, 1249, 1250, 1250, 1252, 1254, 1253, 1254, 1254, 1255, 1256, 1255, 1256, 1258, 1259, 1260, 1261, 1262, 1262, 1264, 1266, 1268, 1268, 1269, 1269, 1269, 1271, 1269, 1269, 1269, 1270, 1270, 1270, 1269, 1268, 1270, 1272, 1272, 1272, 1272, 1273, 1273, 1272, 1273, 1271, 1271, 1271, 1272, 1272, 1272, 1273, 1271, 1270, 1271, 1272, 1275, 1275, 1274, 1274, 1274, 1272, 1271, 1271, 1271, 1271, 1271, 1271, 1271, 1272, 1273, 1277, 1276, 1276, 1274, 1276, 1275, 1275, 1277, 1278, 1279, 1279, 1278, 1280, 1278, 1277, 1276, 1277, 1278, 1279, 1280, 1279, 1280, 1280, 1282, 1282],
[1034, 1035, 1034, 1036, 1035, 1034, 1034, 1033, 1033, 1031, 1032, 1032, 1034, 1033, 1034, 1034, 1034, 1033, 1033, 1035, 1035, 1035, 1034, 1035, 1035, 1035, 1034, 1033, 1031, 1031, 1032, 1032, 1031, 1032, 1033, 1034, 1036, 1038, 1036, 1035, 1037, 1039, 1039, 1041, 1041, 1041, 1042, 1044, 1046, 1046, 1046, 1047, 1048, 1048, 1048, 1046, 1046, 1046, 1046, 1046, 1047, 1049, 1049, 1049, 1049, 1049, 1050, 1050, 1050, 1050, 1049, 1048, 1048, 1049, 1049, 1049, 1047, 1047, 1047, 1047, 1045, 1045, 1047, 1048, 1048, 1047, 1047, 1046, 1047, 1047, 1047, 1046, 1045, 1045, 1044, 1044, 1044, 1046, 1045, 1047, 1048, 1048, 1048, 1049, 1049, 1049, 1052, 1053, 1053, 1053, 1053, 1052, 1052, 1051, 1051, 1054, 1054, 1054, 1054, 1055, 1054, 1055, 1056, 1056, 1054, 1053, 1052, 1052, 1054, 1054, 1052, 1050, 1049, 1048, 1048, 1049, 1051, 1051, 1053, 1053, 1054, 1053, 1053, 1051, 1050, 1051, 1051, 1051, 1050, 1051],
[793, 793, 794, 794, 796, 796, 796, 798, 798, 800, 799, 797, 797, 798, 798, 798, 798, 799, 800, 800, 800, 800, 801, 801, 801, 800, 801, 802, 804, 803, 802, 801, 802, 802, 802, 802, 801, 800, 802, 802, 802, 802, 800, 800, 800, 799, 799, 799, 799, 799, 798, 798, 797, 796, 796, 798, 799, 799, 798, 798, 798, 798, 797, 797, 797, 798, 797, 797, 797, 797, 798, 799, 799, 798, 798, 797, 798, 798, 797, 798, 800, 800, 800, 800, 800, 799, 799, 801, 801, 800, 800, 801, 801, 801, 802, 802, 801, 800, 801, 801, 799, 799, 798, 798, 797, 798, 797, 798, 798, 798, 797, 798, 800, 802, 803, 803, 803, 803, 804, 804, 805, 805, 805, 805, 806, 807, 807, 807, 807, 807, 810, 812, 813, 813, 813, 813, 813, 812, 812, 813, 813, 814, 813, 815, 815, 815, 815, 816, 817, 817],
[665, 665, 665, 665, 662, 663, 663, 663, 663, 662, 663, 664, 664, 664, 664, 664, 664, 662, 662, 662, 662, 662, 661, 661, 660, 661, 661, 661, 660, 661, 661, 661, 661, 661, 661, 661, 661, 661, 661, 662, 661, 661, 664, 664, 664, 665, 665, 665, 664, 663, 664, 664, 665, 666, 666, 666, 666, 666, 667, 667, 666, 666, 666, 666, 666, 666, 666, 665, 665, 665, 665, 665, 665, 665, 665, 666, 667, 667, 667, 667, 667, 667, 667, 667, 667, 669, 669, 668, 668, 670, 670, 669, 669, 668, 668, 667, 668, 667, 666, 666, 668, 668, 669, 669, 670, 670, 671, 670, 669, 669, 669, 669, 669, 669, 669, 668, 667, 666, 666, 664, 664, 662, 663, 662, 663, 661, 662, 661, 661, 661, 661, 661, 661, 662, 662, 662, 662, 662, 662, 662, 662, 662, 663, 663, 663, 663, 663, 663, 662, 662],
[567, 567, 565, 565, 568, 567, 567, 567, 567, 567, 566, 565, 565, 566, 565, 565, 565, 566, 566, 566, 566, 567, 568, 568, 569, 567, 566, 566, 567, 567, 568, 569, 569, 569, 569, 569, 570, 571, 571, 571, 572, 572, 572, 572, 572, 570, 570, 570, 571, 572, 572, 571, 571, 570, 570, 570, 570, 569, 568, 568, 569, 569, 569, 569, 569, 569, 569, 569, 568, 568, 568, 568, 568, 569, 569, 569, 568, 567, 568, 568, 567, 567, 567, 566, 566, 565, 565, 564, 564, 564, 564, 564, 565, 566, 566, 567, 567, 569, 570, 570, 569, 569, 568, 568, 567, 567, 566, 567, 567, 567, 567, 567, 567, 567, 567, 568, 570, 571, 571, 573, 573, 575, 575, 576, 576, 578, 578, 579, 577, 577, 577, 577, 577, 576, 574, 573, 573, 574, 573, 573, 573, 573, 573, 573, 574, 574, 573, 573, 574, 574],
[398, 398, 400, 399, 399, 399, 399, 399, 399, 400, 401, 402, 402, 402, 403, 402, 402, 403, 403, 403, 403, 403, 403, 404, 404, 406, 407, 407, 407, 407, 407, 407, 407, 407, 407, 406, 406, 406, 406, 406, 406, 405, 403, 403, 403, 405, 404, 404, 403, 403, 403, 404, 403, 404, 404, 404, 403, 403, 404, 404, 404, 404, 405, 405, 405, 405, 405, 406, 407, 406, 405, 405, 405, 405, 405, 404, 405, 406, 406, 406, 407, 407, 407, 408, 408, 409, 409, 411, 411, 411, 411, 411, 411, 410, 410, 410, 409, 409, 409, 408, 409, 409, 410, 410, 411, 411, 412, 412, 413, 413, 414, 413, 413, 413, 413, 413, 413, 412, 412, 412, 411, 410, 410, 410, 410, 408, 409, 409, 411, 411, 410, 410, 410, 410, 412, 412, 411, 411, 411, 411, 411, 411, 411, 411, 411, 411, 412, 412, 412, 412],
[349, 349, 349, 350, 350, 351, 351, 351, 350, 350, 350, 351, 351, 350, 350, 351, 351, 351, 351, 351, 350, 350, 350, 350, 350, 349, 349, 349, 349, 349, 349, 349, 349, 349, 349, 350, 350, 350, 349, 349, 349, 350, 351, 351, 351, 350, 351, 351, 352, 352, 352, 351, 352, 352, 352, 351, 352, 353, 352, 352, 352, 352, 352, 352, 352, 352, 353, 353, 353, 354, 355, 355, 354, 354, 354, 355, 355, 355, 355, 355, 355, 355, 355, 355, 354, 354, 353, 353, 353, 353, 353, 353, 353, 354, 354, 354, 355, 355, 355, 356, 356, 356, 356, 356, 356, 356, 355, 355, 354, 354, 354, 355, 355, 355, 355, 354, 354, 355, 355, 355, 356, 357, 357, 357, 357, 359, 359, 359, 359, 358, 359, 359, 359, 360, 360, 361, 362, 362, 363, 363, 363, 361, 361, 360, 359, 359, 359, 359, 359, 359],
[281, 281, 281, 281, 281, 281, 280, 280, 281, 281, 281, 281, 281, 282, 282, 282, 282, 282, 282, 282, 283, 283, 282, 282, 282, 283, 283, 283, 283, 283, 283, 283, 283, 283, 283, 282, 281, 281, 282, 282, 281, 281, 282, 282, 282, 283, 283, 283, 283, 283, 283, 284, 284, 284, 284, 285, 285, 285, 286, 286, 286, 286, 286, 286, 286, 286, 286, 286, 286, 286, 286, 286, 287, 287, 287, 287, 286, 286, 286, 286, 285, 285, 285, 285, 285, 285, 286, 286, 286, 286, 286, 287, 287, 287, 287, 287, 287, 287, 287, 287, 286, 286, 286, 285, 285, 285, 285, 285, 286, 284, 284, 284, 284, 283, 282, 283, 283, 283, 283, 283, 283, 282, 282, 282, 282, 282, 282, 282, 281, 282, 282, 282, 282, 282, 282, 281, 281, 281, 281, 281, 281, 283, 283, 283, 284, 283, 283, 282, 282, 282],
[218, 218, 218, 218, 218, 218, 219, 219, 218, 218, 218, 218, 218, 218, 218, 218, 218, 218, 218, 218, 218, 218, 219, 219, 219, 219, 219, 219, 219, 219, 218, 216, 216, 215, 215, 215, 215, 214, 214, 214, 215, 215, 215, 215, 215, 215, 215, 215, 215, 215, 215, 215, 214, 214, 214, 214, 214, 214, 214, 214, 213, 213, 213, 213, 213, 213, 213, 213, 213, 213, 213, 213, 213, 213, 213, 213, 214, 214, 214, 214, 215, 215, 215, 215, 215, 215, 215, 215, 215, 215, 215, 215, 215, 215, 215, 215, 215, 215, 215, 215, 215, 215, 215, 216, 216, 216, 217, 217, 216, 218, 218, 218, 218, 219, 220, 220, 220, 220, 220, 220, 220, 221, 221, 221, 221, 221, 220, 220, 221, 221, 220, 220, 220, 220, 220, 221, 221, 221, 221, 221, 221, 221, 221, 222, 222, 223, 223, 224, 224, 224],
[188, 188, 188, 188, 188, 188, 188, 188, 189, 189, 188, 188, 188, 188, 188, 187, 187, 187, 187, 186, 186, 184, 182, 182, 182, 182, 182, 182, 181, 181, 182, 184, 184, 185, 185, 186, 187, 188, 188, 188, 188, 188, 188, 188, 188, 188, 188, 188, 188, 188, 188, 188, 189, 189, 189, 188, 188, 188, 188, 187, 188, 188, 188, 188, 188, 188, 188, 188, 188, 188, 188, 187, 187, 187, 187, 187, 187, 187, 187, 187, 187, 186, 186, 186, 187, 186, 186, 186, 186, 186, 185, 184, 184, 184, 184, 184, 184, 184, 183, 183, 184, 184, 184, 184, 184, 184, 184, 184, 185, 185, 185, 185, 185, 185, 185, 185, 184, 184, 184, 184, 184, 184, 184, 184, 184, 184, 185, 185, 185, 185, 186, 186, 186, 186, 186, 186, 186, 186, 186, 186, 186, 186, 186, 186, 186, 186, 186, 186, 186, 186],
[158, 158, 158, 158, 158, 158, 158, 158, 158, 158, 159, 159, 159, 159, 159, 160, 160, 160, 160, 161, 161, 163, 165, 165, 165, 165, 165, 165, 166, 166, 166, 166, 166, 166, 166, 166, 166, 166, 165, 165, 165, 165, 165, 165, 165, 165, 165, 165, 165, 165, 165, 165, 165, 164, 164, 165, 165, 165, 165, 166, 166, 166, 166, 166, 166, 166, 166, 166, 166, 166, 166, 167, 167, 167, 167, 167, 167, 167, 167, 167, 167, 168, 168, 168, 168, 169, 169, 169, 169, 169, 170, 171, 171, 171, 171, 171, 171, 171, 170, 170, 170, 170, 168, 168, 168, 168, 168, 168, 168, 167, 166, 166, 165, 165, 165, 165, 166, 166, 166, 166, 166, 166, 166, 166, 166, 166, 165, 165, 165, 165, 165, 164, 164, 164, 164, 164, 164, 164, 164, 164, 164, 164, 164, 163, 163, 163, 163, 163, 163, 163],
[115, 115, 115, 114, 114, 114, 114, 114, 114, 114, 114, 114, 113, 113, 113, 113, 112, 112, 112, 111, 111, 111, 111, 111, 111, 111, 110, 110, 109, 109, 109, 109, 109, 109, 109, 109, 109, 109, 110, 110, 110, 110, 109, 109, 109, 109, 109, 109, 109, 109, 109, 109, 109, 110, 110, 110, 110, 110, 110, 110, 110, 110, 110, 110, 110, 109, 109, 109, 109, 109, 109, 108, 108, 108, 108, 108, 108, 108, 108, 108, 108, 108, 108, 108, 107, 107, 107, 107, 107, 107, 107, 107, 107, 107, 106, 106, 106, 106, 108, 108, 108, 108, 110, 109, 109, 109, 109, 109, 109, 110, 111, 111, 112, 112, 112, 112, 112, 112, 112, 112, 112, 112, 112, 112, 112, 112, 113, 113, 113, 113, 113, 114, 113, 113, 113, 113, 113, 113, 113, 113, 113, 113, 113, 114, 114, 114, 114, 114, 114, 114],
[95, 95, 95, 96, 96, 96, 96, 96, 96, 96, 96, 96, 97, 97, 97, 97, 98, 98, 98, 99, 99, 99, 99, 99, 99, 99, 100, 100, 101, 101, 101, 101, 101, 101, 101, 101, 101, 101, 101, 101, 101, 101, 102, 102, 102, 102, 102, 102, 102, 102, 102, 102, 102, 102, 102, 102, 102, 102, 102, 102, 102, 102, 102, 102, 102, 103, 103, 103, 103, 103, 103, 104, 104, 104, 104, 104, 104, 103, 103, 103, 103, 103, 103, 103, 104, 104, 104, 103, 103, 103, 103, 103, 103, 103, 104, 104, 104, 104, 104, 104, 104, 104, 104, 105, 105, 105, 105, 105, 105, 105, 105, 105, 105, 105, 105, 105, 104, 104, 104, 104, 104, 104, 104, 104, 104, 104, 104, 104, 104, 104, 104, 104, 105, 105, 105, 105, 105, 105, 105, 105, 105, 105, 105, 105, 105, 105, 105, 105, 105, 105],
[78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 78, 77, 77, 77, 77, 77, 77, 77, 76, 76, 76, 76, 76, 76, 76, 76, 76, 76, 76, 76, 76, 76, 76, 76, 76, 76, 76, 76, 76, 76, 76, 76, 76, 76, 76, 76, 76, 76, 76, 76, 76, 76, 76, 76, 76, 76, 76, 76, 76, 76, 75, 75, 75, 75, 75, 75, 75, 75, 75, 76, 76, 75, 75, 75, 75, 75, 75, 75, 75, 76, 76, 76, 76, 76, 76, 76, 76, 76, 76, 76, 76, 76, 75, 75, 75, 75, 75, 75, 75, 75, 75, 75, 75, 75, 75, 75, 75, 74, 75, 75, 75, 75, 75, 75, 75, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74],
[72, 72, 72, 72, 72, 72, 72, 72, 72, 72, 72, 72, 71, 71, 71, 71, 71, 71, 71, 72, 72, 72, 72, 72, 72, 72, 73, 73, 73, 73, 73, 73, 73, 73, 73, 73, 73, 73, 73, 73, 73, 73, 73, 73, 73, 73, 73, 73, 73, 73, 73, 73, 73, 73, 73, 73, 73, 73, 73, 73, 73, 73, 73, 73, 73, 73, 72, 72, 73, 73, 73, 73, 73, 73, 73, 73, 73, 73, 73, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 75, 75, 75, 75, 75, 75, 75, 75, 75, 75, 74, 74, 73, 73, 73, 73, 73, 73, 73, 73, 73, 73, 73, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74, 74],
[49, 49, 49, 49, 49, 49, 49, 49, 49, 49, 49, 49, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 51, 51, 51, 51, 51, 51, 51, 51, 51, 51, 51, 51, 51, 51, 51, 51, 51, 51, 51, 51, 51, 49, 49, 49, 49, 49, 49, 49, 49, 49, 49, 49, 49, 49, 49, 49, 49, 49, 49, 49, 49, 49, 49, 49, 50, 50, 51, 51, 51, 52, 52, 52, 52, 51, 51, 51, 51, 51, 51, 51, 51, 51, 51, 51, 51, 51, 51, 51, 51, 51, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50, 50],
[39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 39, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 41, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 42, 43, 43, 43, 43, 43, 43, 43, 43, 43, 42, 42, 42, 42, 42],
[30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 30, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 28, 29, 29, 29, 29, 29],
[119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 119, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 120, 121, 121, 121, 121, 121, 121, 121, 121, 121, 121, 121, 121, 121, 121, 121, 121, 121]
]
}
}
}
Demographic summary output report¶
The demographic summary output report is a JSON-formatted file with the demographic channel output results of the simulation, consisting of simulation-wide averages by time step. The format is identical to the inset chart output report, except the channels reflect demographic categories, such as gender ratio. The file name is DemographicsSummary.json.
To generate the demographics summary report, set the Enable_Demographics_Reporting configuration parameter to 1. The binned output report will also be generated.
The file contains a header and a channels section.
Header¶
The header section contains the following parameters.
Parameter |
Data type |
Description |
|---|---|---|
DateTime |
string |
The time stamp indicating when the report was generated. |
DTK_Version |
string |
The version of EMOD used. |
Report_Type |
string |
The type of output report. |
Report_Version |
string |
The format version of the report. |
Start_Time |
integer |
The time noted in days when the simulation begins. |
Simulation_Timestep |
integer |
The number of days in each time step. |
Timesteps |
integer |
The number of time steps in this simulation. |
Channels |
integer |
The number of channels in the simulation. |
Channels¶
The channels section contains the following parameters.
Parameter |
Data type |
Description |
|---|---|---|
<Channel_Title> |
string |
The title of the particular demographic channel. |
Units |
string |
The units used for this channel. |
Data |
array |
A list of the channel data at each time step. |
Example¶
The following is a sample of a DemographicsSummary.json file.
{
"Header": {
"DateTime": "Mon Mar 16 07:45:10 2015",
"DTK_Version": "4777 v2.0-HIV 2015/02/26 10:51:25",
"Report_Type": "InsetChart",
"Report_Version": "3.2",
"Start_Time": 0,
"Simulation_Timestep": 1,
"Timesteps": 5,
"Channels": 28
},
"Channels": {
"Average Age": {
"Units": "",
"Data": [
8592.415039063,
8593.427734375,
8594.439453125,
8595.41796875,
8596.412109375
]
},
"Gender Ratio (fraction male)": {
"Units": "",
"Data": [
0.5350999832153,
0.5350999832153,
0.5350999832153,
0.5350999832153,
0.5350999832153
]
},
"New Births": {
"Units": "",
"Data": [
0,
0,
0,
0,
0
]
},
"New Natural Deaths": {
"Units": "",
"Data": [
0,
0,
0,
0,
0
]
},
"Population Age 10-14": {
"Units": "",
"Data": [
1203,
1204,
1202,
1203,
1203
]
},
"Population Age 15-19": {
"Units": "",
"Data": [
1056,
1057,
1059,
1059,
1059
]
},
"Population Age 20-24": {
"Units": "",
"Data": [
810,
810,
810,
809,
809
]
},
"Population Age 25-29": {
"Units": "",
"Data": [
732,
732,
732,
732,
732
]
},
"Population Age 30-34": {
"Units": "",
"Data": [
540,
539,
539,
539,
539
]
},
"Population Age 35-39": {
"Units": "",
"Data": [
410,
411,
411,
411,
411
]
},
"Population Age 40-44": {
"Units": "",
"Data": [
351,
351,
351,
352,
352
]
},
"Population Age 45-49": {
"Units": "",
"Data": [
294,
294,
294,
294,
294
]
},
"Population Age 5-9": {
"Units": "",
"Data": [
1599,
1597,
1598,
1597,
1599
]
},
"Population Age 50-54": {
"Units": "",
"Data": [
201,
201,
201,
201,
201
]
},
"Population Age 55-59": {
"Units": "",
"Data": [
194,
194,
194,
193,
193
]
},
"Population Age 60-64": {
"Units": "",
"Data": [
163,
163,
163,
164,
164
]
},
"Population Age 65-69": {
"Units": "",
"Data": [
111,
111,
111,
111,
111
]
},
"Population Age 70-74": {
"Units": "",
"Data": [
104,
104,
104,
103,
103
]
},
"Population Age 75-79": {
"Units": "",
"Data": [
86,
86,
86,
87,
87
]
},
"Population Age 80-84": {
"Units": "",
"Data": [
55,
55,
55,
55,
55
]
},
"Population Age 85-89": {
"Units": "",
"Data": [
61,
61,
61,
61,
61
]
},
"Population Age 90-94": {
"Units": "",
"Data": [
49,
49,
49,
49,
49
]
},
"Population Age 95-99": {
"Units": "",
"Data": [
30,
30,
30,
30,
30
]
},
"Population Age <5": {
"Units": "",
"Data": [
1829,
1829,
1828,
1828,
1826
]
},
"Population Age >100": {
"Units": "",
"Data": [
122,
122,
122,
122,
122
]
},
"Possible Mothers": {
"Units": "",
"Data": [
1912,
1912,
1912,
1913,
1913
]
},
"Pseudo-Population": {
"Units": "",
"Data": [
10000,
10000,
10000,
10000,
10000
]
},
"Statistical Population": {
"Units": "",
"Data": [
10000,
10000,
10000,
10000,
10000
]
}
}
}
Property output report¶
The property output report is a JSON-formatted file with the channel output results of the simulation, defined by the groups set up using IndividualProperties in the demographics file. See Configure heterogeneous populations. For example, it allows you to compare disease deaths for people in the high risk group versus the low risk group. If you want to aggregate the data, you must create a script for aggregating information. The file name is PropertyReport.json.
To generate the property report, set the Enable_Property_Output configuration parameter to 1.
The file contains a header and a channels section.
Header¶
The header section contains the following parameters.
Parameter |
Data type |
Description |
|---|---|---|
DateTime |
string |
The time stamp indicating when the report was generated. |
DTK_Version |
string |
The version of EMOD used. |
Report_Type |
string |
The type of output report. |
Report_Version |
string |
The format version of the report. |
Start_Time |
integer |
The time noted in days when the simulation begins. |
Simulation_Timestep |
integer |
The number of days in each time step. |
Timesteps |
integer |
The number of time steps in this simulation. |
Channels |
integer |
The number of channels in the simulation. |
Channels¶
The channels section contains the following parameters.
Parameter |
Data type |
Description |
|---|---|---|
<Channel_Title> |
string |
The title of the particular property channel. The channel titles use the following conventions: for a single property, <channel type>:<property>:<value>, and for multiple properties, <channel type>:<property 1>:<value>,<property 2>:<value>. For example, Infected:Accessibility:Easy or New Infections:Accessibility:Difficult,Risk:High. |
Units |
string |
The units used for this channel. |
Data |
array |
A list of the channel data at each time step. |
Example¶
The following is a sample of a PropertyReport.json file.
{
"Header": {
"DateTime": "Mon Feb 15 21:49:24 2016",
"DTK_Version": "5538 trunk 2015/08/07 14:40:43",
"Report_Type": "InsetChart",
"Report_Version": "3.2",
"Start_Time": 0,
"Simulation_Timestep": 1,
"Timesteps": 10,
"Channels": 8
},
"Channels": {
"Disease Deaths:Accessibility:Easy": {
"Units": "",
"Data": [
0,
0,
0,
0,
0,
0,
0,
0,
0,
0
]
},
"Disease Deaths:Accessibility:Hard": {
"Units": "",
"Data": [
0,
0,
0,
0,
0,
0,
0,
0,
0,
0
]
},
"Infected:Accessibility:Easy": {
"Units": "",
"Data": [
0,
0,
0,
0,
0,
0,
0,
0,
0,
0
]
},
"Infected:Accessibility:Hard": {
"Units": "",
"Data": [
0,
0,
0,
0,
0,
0,
0,
0,
0,
0
]
},
"New Infections:Accessibility:Easy": {
"Units": "",
"Data": [
0,
0,
0,
0,
0,
0,
0,
0,
0,
0
]
},
"New Infections:Accessibility:Hard": {
"Units": "",
"Data": [
0,
0,
0,
0,
0,
0,
0,
0,
0,
0
]
},
"Statistical Population:Accessibility:Easy": {
"Units": "",
"Data": [
6946,
6946,
6946,
6946,
6946,
6946,
6946,
6946,
6946,
6946
]
},
"Statistical Population:Accessibility:Hard": {
"Units": "",
"Data": [
3054,
3054,
3054,
3054,
3054,
3054,
3054,
3054,
3054,
3054
]
}
}
}
Spatial output report¶
The spatial output report breaks the channel data down per node, rather than across the entire simulation. It is a set of binary files, consisting of one file per channel. For each value set in the Spatial_Output_Channels configuration parameter array, a binary file with the name convention SpatialReport_<channel>.bin is generated. In addition, Enable_Spatial_Output must be set to 1.
The binary format of the file consists of a stream of 4-byte integers followed by a stream of 4-byte floating point values. The first value is a 4-byte integer representing the number of nodes in the file and the second is a 4-byte integer that contains the number of time steps in the file. Following these two values is a stream of 4-byte integers that contain the node ID values in the order they will appear in the rest of the file. Following the node IDs is an array of 4-byte floating point values that represent the output values at the first time step for each node. The next array contains the values at the second time step, and so on.
The following diagram shows the format for data in the spatial output report file:
Vector species output report¶
The vector species output report is a JSON-formatted file where the channel data has been sorted into bins. It is identical to the Binned output report, with the exception that the bins are based on vector species. The vector species report is generated for all malaria or vector simulations that contain values in the Vector_Species_Names configuration parameter. For example, if Vector_Species_Names was set to [“funestus”, “gambiae”], you would be able to see how many infectious vectors there were for A. funestus and A. gambiae. The file name is VectorSpeciesReport.json.
Example¶
The following is a sample of an VectorSpeciesReport.json file.
{
"Header": {
"DateTime": "Thu Jul 23 13:06:54 2015",
"DTK_Version": "5004 trunk Jun 25 2015",
"Report_Version": "2.1",
"Timesteps": 100,
"Subchannel_Metadata": {
"AxisLabels": [
["Vector Species"]
],
"NumBinsPerAxis": [
[1]
],
"ValuesPerAxis": [
[
[0]
]
],
"MeaningPerAxis": [
[
["arabiensis"]
]
]
},
"Channels": 4
},
"Channels": {
"Adult Vectors": {
"Units": "",
"Data": [
[8739, 7627, 6712, 5882, 5174, 4577, 4003, 3500, 3086, 2691, 2340, 2038, 1790, 1562, 1367, 1191, 1045, 924, 801, 704, 620, 540, 472, 412, 353, 314, 1188, 2198, 2152, 3738, 3706, 3918, 6105, 6197, 6897, 6632, 6656, 6858, 6748, 6390, 6300, 5540, 5284, 5230, 5744, 5487, 5689, 5563, 5935, 6077, 6091, 6123, 5970, 5879, 5956, 6220, 6315, 6361, 6144, 5943, 5193, 6508, 6692, 8009, 8139, 7800, 8575, 8431, 8126, 9099, 8682, 8691, 8499, 8901, 8603, 8359, 8135, 8068, 7890, 6884, 6667, 7782, 6809, 7698, 7520, 7447, 6513, 8771, 7674, 7577, 8048, 8304, 8448, 8710, 8881, 10137, 9867, 9320, 9019, 9754]
]
},
"Daily EIR": {
"Units": "",
"Data": [
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.00400000018999, 0, 0.006000000052154, 0.01200000010431, 0.01200000010431, 0.006000000052154, 0.01200000010431, 0.01400000043213, 0.01999999955297, 0.04399999976158, 0.0380000025034, 0.03400000184774, 0.03400000184774, 0.02400000020862, 0.02600000053644, 0.03400000184774, 0.0379999987781, 0.03999999910593, 0.04600000008941, 0.04600000008941, 0.04399999976158, 0.05599999800324, 0.03200000151992, 0.02400000020862, 0.0380000025034, 0.02000000141561, 0.02799999900162, 0.01999999955297, 0.01800000108778, 0.01400000043213, 0.009999999776483, 0.00800000037998, 0.00800000037998, 0.009999999776483, 0.006000000052154, 0.002000000094995, 0.002000000094995, 0.002000000094995, 0.002000000094995, 0.002000000094995, 0.002000000094995, 0, 0.002000000094995, 0.00400000018999, 0.002000000094995, 0.00800000037998, 0.006000000052154, 0.01799999922514, 0.01799999922514, 0.02400000020862, 0.02999999932945, 0.02799999900162, 0.02400000020862, 0.06000000238419]
]
},
"Daily HBR": {
"Units": "",
"Data": [
[5.640000343323, 4.909999847412, 4.430000305176, 3.762000083923, 3.458000183105, 2.966000080109, 2.68799996376, 2.226000070572, 2.041999816895, 1.78200006485, 1.50200009346, 1.351999998093, 1.195999979973, 1.089999914169, 0.8100000023842, 0.7920000553131, 0.6539999842644, 0.6039999723434, 0.539999961853, 0.4399999976158, 0.4039999842644, 0.3779999911785, 0.2840000092983, 0.2580000162125, 0.2320000082254, 0.2260000109673, 0.1800000071526, 0.6340000033379, 1.213999986649, 1.22000002861, 2.179999828339, 2.012000083923, 2.232000112534, 3.400000095367, 3.461999893188, 4.038000106812, 3.802000045776, 3.886000156403, 3.869999885559, 3.860000133514, 3.598000049591, 3.567999839783, 3.078000068665, 3.067999839783, 2.969999790192, 3.317999839783, 3.110000133514, 3.234000205994, 3.339999914169, 3.555999994278, 3.575999975204, 3.644000053406, 3.519999980927, 3.434000015259, 3.369999885559, 3.398000001907, 3.438000202179, 3.661999940872, 3.786000013351, 3.584000110626, 3.3140001297, 3.067999839783, 3.817999839783, 3.6859998703, 4.696000099182, 4.748000144958, 4.466000080109, 4.932000160217, 4.828000068665, 4.760000228882, 5.244000434875, 5.083999633789, 4.946000099182, 4.944000244141, 5.139999866486, 4.871999740601, 4.868000030518, 4.656000137329, 4.456000328064, 4.487999916077, 4.003999710083, 3.636000156403, 4.421999931335, 3.906000137329, 4.355999946594, 4.354000091553, 4.319999694824, 3.710000038147, 5.109999656677, 4.28200006485, 4.512000083923, 4.640000343323, 4.753999710083, 4.725999832153, 5.130000114441, 5.031999588013, 5.926000118256, 5.78600025177, 5.498000144958, 5.203999996185]
]
},
"Infectious Vectors": {
"Units": "",
"Data": [
[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.0001740947045619, 0.0007289958302863, 0.0007031112909317, 0.001617832109332, 0.001853412017226, 0.002468323102221, 0.003447709837928, 0.004246284719557, 0.005192629992962, 0.006973975338042, 0.009402283467352, 0.009967845864594, 0.009184481576085, 0.009432478807867, 0.0087890625, 0.009086320176721, 0.01251684967428, 0.00983405020088, 0.01046025101095, 0.008989886380732, 0.0088462959975, 0.009615384973586, 0.0103790089488, 0.008421302773058, 0.007506768219173, 0.005385207012296, 0.004952775780112, 0.00460246251896, 0.003529827110469, 0.003033366985619, 0.002673486014828, 0.002272999146953, 0.002212661318481, 0.001735250349157, 0.001520912512206, 0.001597908209078, 0.001199939986691, 0.001028013415635, 0.001028051134199, 0.0007794232224114, 0.0006648935959674, 0.0006714112823829, 0.0006141563062556, 0.0004560483503155, 0.0006515507120639, 0.0009238484781235, 0.0008697813027538, 0.001806358341128, 0.001775568234734, 0.002525832271203, 0.004391397349536, 0.004735128954053, 0.005472788121551, 0.006759656593204, 0.009646302089095, 0.007996719330549]
]
}
}
}
Error and logging files¶
When you run a simulation, EMOD will output basic error and logging information to help track the progress and help debug any issues that may occur. If you run the simulation on an HPC cluster, the cluster will generate additional logging and error files. See Troubleshooting EMOD simulations if you need help resolving an error.
Status¶
A status.txt file will be saved to the working directory that provides one output line per time step and includes the total run time of the simulation. A simulation with 50 time steps will look something like this:
Beginning Simulation...
1 of 50 steps complete.
2 of 50 steps complete.
3 of 50 steps complete.
4 of 50 steps complete.
5 of 50 steps complete.
6 of 50 steps complete.
7 of 50 steps complete.
8 of 50 steps complete.
9 of 50 steps complete.
10 of 50 steps complete.
11 of 50 steps complete.
12 of 50 steps complete.
13 of 50 steps complete.
14 of 50 steps complete.
15 of 50 steps complete.
16 of 50 steps complete.
17 of 50 steps complete.
18 of 50 steps complete.
19 of 50 steps complete.
20 of 50 steps complete.
21 of 50 steps complete.
22 of 50 steps complete.
23 of 50 steps complete.
24 of 50 steps complete.
25 of 50 steps complete.
26 of 50 steps complete.
27 of 50 steps complete.
28 of 50 steps complete.
29 of 50 steps complete.
30 of 50 steps complete.
31 of 50 steps complete.
32 of 50 steps complete.
33 of 50 steps complete.
34 of 50 steps complete.
35 of 50 steps complete.
36 of 50 steps complete.
37 of 50 steps complete.
38 of 50 steps complete.
39 of 50 steps complete.
40 of 50 steps complete.
41 of 50 steps complete.
42 of 50 steps complete.
43 of 50 steps complete.
44 of 50 steps complete.
45 of 50 steps complete.
46 of 50 steps complete.
47 of 50 steps complete.
48 of 50 steps complete.
49 of 50 steps complete.
50 of 50 steps complete.
Done - 0:00:02
Standard output¶
When you run a simulation on an HPC cluster, it will generate a standard output logging file
(StdOut.txt) in the working directory that captures all standard output messages. When you run a
simulation locally, the Command Prompt window will display the same information. If you want to save
this information to a text file instead, you can append > stdout.txt to the command used to
run the local EMOD simulation to redirect the console output to the file specified.
The file contains information about a particular simulation, such as the EMOD version used and the files in use, as well as other initialization information, including the default logging level and the logging levels set for particular modules. The file follows that information with log output using the following format: <timestep><HPC rank><log level><module><message>.
By default, the logging level is set to “INFO”. If you want to change the logging level, see Set log levels.
For example:
C:\EMOD\CampaignTesting\00_DefaultDemographics>..\Eradication.exe -C config.json -I ../input
Intellectual Ventures(R)/EMOD Disease Transmission Kernel 2.8.1331.0
Built on Sep 30 2016 08:39:43 by cwiswell from master (482fdae) checked in on 2016-09-28 14:47:20 -0700
Supports sim_types: GENERIC, VECTOR, MALARIA, AIRBORNE, POLIO, TB, TBHIV, STI, HIV, PY.
Using config file: config.json
Using input path: ../input
Using output path: output
using dll path:
--python-script-path (-P) not on command line - not using embedded python
Initializing environment...
Log-levels:
Default -> INFO
Eradication -> INFO
00:00:00 [0] [I] [Eradication] Loaded Configuration...
00:00:00 [0] [I] [Eradication] 99 parameters found.
00:00:00 [0] [I] [Eradication] Initializing Controller...
00:00:00 [0] [I] [Controller] DefaultController::execute_internal()...
00:00:00 [0] [I] [Simulation] Using PSEUDO_DES random number generator.
00:00:00 [0] [I] [Controller] DefaultController::execute_internal() populate simulation...
00:00:00 [0] [I] [Simulation] Campaign file name identified as: campaign.json
00:00:00 [0] [I] [NodeDemographics] Using default 3x3 torus geography
00:00:00 [0] [I] [Climate] Initialize
00:00:00 [0] [I] [LoadBalanceScheme] Using Checkerboard Load Balance Scheme.
00:00:01 [0] [I] [Simulation] Looking for campaign file campaign.json
00:00:01 [0] [I] [Simulation] Found campaign file successfully.
00:00:01 [0] [I] [DllLoader] ReadEmodulesJson: no file, returning.
00:00:01 [0] [I] [DllLoader] dllPath not passed in, getting from EnvPtr
00:00:01 [0] [I] [DllLoader] Trying to copy from string to wstring.
00:00:01 [0] [I] [DllLoader] DLL ws root path:
00:00:01 [0] [W] [Simulation] Failed to load intervention emodules for SimType: GENERIC_SIM from path: interventions
00:00:01 [0] [I] [JsonConfigurable] Using the default value ( "Number_Repetitions" : -1 ) for unspecified parameter.
00:00:01 [0] [I] [JsonConfigurable] Using the default value ( "Timesteps_Between_Repetitions" : -1 ) for unspecified parameter.
00:00:01 [0] [I] [JsonConfigurable] Using the default value ( "Intervention_Name" : "struct Kernel::BaseIntervention" ) for unspecified parameter.
00:00:01 [0] [I] [JsonConfigurable] Using the default value ( "Intervention_Name" : "struct Kernel::BaseIntervention" ) for unspecified parameter.
00:00:01 [0] [I] [JsonConfigurable] Using the default value ( "Intervention_Name" : "struct Kernel::BaseIntervention" ) for unspecified parameter.
00:00:01 [0] [I] [Simulation] populateFromDemographics() created 9 nodes
00:00:01 [0] [I] [Simulation] populateFromDemographics() generated 9 nodes.
00:00:01 [0] [I] [Simulation] Rank 0 contributes 9 nodes...
00:00:01 [0] [I] [Simulation] Merging node rank maps...
00:00:01 [0] [I] [Simulation] Merged rank 0 map now has 9 nodes.
00:00:01 [0] [I] [Simulation] Rank 0 map contents:
{ NodeRankMap:
[1,000000000C10A960]
[2,000000000BF4EAF0]
[3,000000000BFA2250]
[4,000000000BFF1C10]
[5,000000000BE37B60]
[6,000000000BE87380]
[7,000000000BEE74F0]
[8,000000000B721980]
[9,000000000B7AE280]
}
00:00:01 [0] [I] [Simulation] Initialized 'InsetChart.json' reporter
00:00:01 [0] [I] [Simulation] Initialized 'BinnedReport.json' reporter
00:00:01 [0] [I] [Simulation] Initialized 'DemographicsSummary.json' reporter
00:00:01 [0] [I] [Simulation] Update(): Time: 1.0 Rank: 0 StatPop: 900 Infected: 0
00:00:01 [0] [I] [SimulationEventContext] Time for campaign event. Calling Dispatch...
00:00:01 [0] [I] [SimulationEventContext] 9 node(s) visited.
00:00:01 [0] [I] [JsonConfigurable] Using the default value ( "Intervention_Name" : "struct Kernel::BaseIntervention" ) for unspecified parameter.
00:00:01 [0] [I] [JsonConfigurable] Using the default value ( "Intervention_Name" : "struct Kernel::BaseIntervention" ) for unspecified parameter.
00:00:01 [0] [I] [StandardEventCoordinator] UpdateNodes() gave out 0 'OutbreakIndividual' interventions at node 1
00:00:01 [0] [I] [StandardEventCoordinator] UpdateNodes() gave out 2 'OutbreakIndividual' interventions at node 2
00:00:01 [0] [I] [StandardEventCoordinator] UpdateNodes() gave out 0 'OutbreakIndividual' interventions at node 3
00:00:01 [0] [I] [StandardEventCoordinator] UpdateNodes() gave out 0 'OutbreakIndividual' interventions at node 4
00:00:01 [0] [I] [StandardEventCoordinator] UpdateNodes() gave out 0 'OutbreakIndividual' interventions at node 5
00:00:01 [0] [I] [StandardEventCoordinator] UpdateNodes() gave out 0 'OutbreakIndividual' interventions at node 6
00:00:01 [0] [I] [StandardEventCoordinator] UpdateNodes() gave out 0 'OutbreakIndividual' interventions at node 7
00:00:01 [0] [I] [StandardEventCoordinator] UpdateNodes() gave out 1 'OutbreakIndividual' interventions at node 8
00:00:01 [0] [I] [StandardEventCoordinator] UpdateNodes() gave out 1 'OutbreakIndividual' interventions at node 9
00:00:01 [0] [I] [Simulation] Update(): Time: 2.0 Rank: 0 StatPop: 900 Infected: 4
00:00:01 [0] [I] [Simulation] Update(): Time: 3.0 Rank: 0 StatPop: 900 Infected: 18
00:00:01 [0] [I] [Simulation] Update(): Time: 4.0 Rank: 0 StatPop: 900 Infected: 63
00:00:01 [0] [I] [Simulation] Update(): Time: 5.0 Rank: 0 StatPop: 900 Infected: 161
00:00:01 [0] [I] [Simulation] Update(): Time: 6.0 Rank: 0 StatPop: 900 Infected: 232
00:00:01 [0] [I] [Simulation] Update(): Time: 7.0 Rank: 0 StatPop: 900 Infected: 203
00:00:01 [0] [I] [Simulation] Update(): Time: 8.0 Rank: 0 StatPop: 900 Infected: 165
00:00:01 [0] [I] [Simulation] Update(): Time: 9.0 Rank: 0 StatPop: 900 Infected: 132
00:00:01 [0] [I] [Simulation] Update(): Time: 10.0 Rank: 0 StatPop: 900 Infected: 110
00:00:01 [0] [I] [Simulation] Update(): Time: 11.0 Rank: 0 StatPop: 900 Infected: 83
00:00:01 [0] [I] [Simulation] Update(): Time: 12.0 Rank: 0 StatPop: 900 Infected: 69
00:00:01 [0] [I] [Simulation] Update(): Time: 13.0 Rank: 0 StatPop: 900 Infected: 54
00:00:01 [0] [I] [Simulation] Update(): Time: 14.0 Rank: 0 StatPop: 900 Infected: 40
00:00:01 [0] [I] [Simulation] Update(): Time: 15.0 Rank: 0 StatPop: 900 Infected: 30
00:00:01 [0] [I] [Simulation] Update(): Time: 16.0 Rank: 0 StatPop: 900 Infected: 22
00:00:01 [0] [I] [Simulation] Update(): Time: 17.0 Rank: 0 StatPop: 900 Infected: 17
00:00:01 [0] [I] [Simulation] Update(): Time: 18.0 Rank: 0 StatPop: 900 Infected: 14
00:00:01 [0] [I] [Simulation] Update(): Time: 19.0 Rank: 0 StatPop: 900 Infected: 12
00:00:01 [0] [I] [Simulation] Update(): Time: 20.0 Rank: 0 StatPop: 900 Infected: 9
00:00:01 [0] [I] [Simulation] Update(): Time: 21.0 Rank: 0 StatPop: 900 Infected: 7
00:00:01 [0] [I] [Simulation] Update(): Time: 22.0 Rank: 0 StatPop: 900 Infected: 7
00:00:01 [0] [I] [Simulation] Update(): Time: 23.0 Rank: 0 StatPop: 900 Infected: 7
00:00:01 [0] [I] [Simulation] Update(): Time: 24.0 Rank: 0 StatPop: 900 Infected: 5
00:00:01 [0] [I] [Simulation] Update(): Time: 25.0 Rank: 0 StatPop: 900 Infected: 3
00:00:01 [0] [I] [Simulation] Update(): Time: 26.0 Rank: 0 StatPop: 900 Infected: 3
00:00:01 [0] [I] [Simulation] Update(): Time: 27.0 Rank: 0 StatPop: 900 Infected: 1
00:00:02 [0] [I] [Simulation] Update(): Time: 28.0 Rank: 0 StatPop: 900 Infected: 1
00:00:02 [0] [I] [Simulation] Update(): Time: 29.0 Rank: 0 StatPop: 900 Infected: 1
00:00:02 [0] [I] [Simulation] Update(): Time: 30.0 Rank: 0 StatPop: 900 Infected: 1
00:00:02 [0] [I] [Simulation] Update(): Time: 31.0 Rank: 0 StatPop: 900 Infected: 1
00:00:02 [0] [I] [Simulation] Update(): Time: 32.0 Rank: 0 StatPop: 900 Infected: 1
00:00:02 [0] [I] [Simulation] Update(): Time: 33.0 Rank: 0 StatPop: 900 Infected: 1
00:00:02 [0] [I] [Simulation] Update(): Time: 34.0 Rank: 0 StatPop: 900 Infected: 1
00:00:02 [0] [I] [Simulation] Update(): Time: 35.0 Rank: 0 StatPop: 900 Infected: 1
00:00:02 [0] [I] [Simulation] Update(): Time: 36.0 Rank: 0 StatPop: 900 Infected: 1
00:00:02 [0] [I] [Simulation] Update(): Time: 37.0 Rank: 0 StatPop: 900 Infected: 1
00:00:02 [0] [I] [Simulation] Update(): Time: 38.0 Rank: 0 StatPop: 900 Infected: 1
00:00:02 [0] [I] [Simulation] Update(): Time: 39.0 Rank: 0 StatPop: 900 Infected: 1
00:00:02 [0] [I] [Simulation] Update(): Time: 40.0 Rank: 0 StatPop: 900 Infected: 1
00:00:02 [0] [I] [Simulation] Update(): Time: 41.0 Rank: 0 StatPop: 900 Infected: 1
00:00:02 [0] [I] [Simulation] Update(): Time: 42.0 Rank: 0 StatPop: 900 Infected: 0
00:00:02 [0] [I] [Simulation] Update(): Time: 43.0 Rank: 0 StatPop: 900 Infected: 0
00:00:02 [0] [I] [Simulation] Update(): Time: 44.0 Rank: 0 StatPop: 900 Infected: 0
00:00:02 [0] [I] [Simulation] Update(): Time: 45.0 Rank: 0 StatPop: 900 Infected: 0
00:00:02 [0] [I] [Simulation] Update(): Time: 46.0 Rank: 0 StatPop: 900 Infected: 0
00:00:02 [0] [I] [Simulation] Update(): Time: 47.0 Rank: 0 StatPop: 900 Infected: 0
00:00:02 [0] [I] [Simulation] Update(): Time: 48.0 Rank: 0 StatPop: 900 Infected: 0
00:00:02 [0] [I] [Simulation] Update(): Time: 49.0 Rank: 0 StatPop: 900 Infected: 0
00:00:02 [0] [I] [Simulation] Update(): Time: 50.0 Rank: 0 StatPop: 900 Infected: 0
00:00:02 [0] [I] [Simulation] Finalizing 'InsetChart.json' reporter.
00:00:02 [0] [I] [Simulation] Finalizing 'BinnedReport.json' reporter.
00:00:02 [0] [I] [Simulation] Finalizing 'DemographicsSummary.json' reporter.
00:00:02 [0] [I] [Eradication] Controller executed successfully.
Standard error¶
When you run a simulation on an HPC cluster, it will also generate a standard error logging file (StdErr.txt) in the working directory that captures all standard error messages.
Troubleshooting EMOD simulations¶
If you encounter any of the following problems when attempting to run EMOD simulations, see the information below to resolve the issue.
If you need assistance, you can contact support for help with solving issues. You can contact Institute for Disease Modeling (IDM) support at support@idmod.org. When submitting the issue, please include any error information. See Debugging and testing for troubleshooting issues when attempting to build Eradication.exe or Eradication binary.
Contents
Exceptions¶
Whenever EMOD encounters an error condition while running a simulation, it should throw an exception. These exceptions are designed to help you diagnose any problems so that you can quickly resolve the issue. You can find the exceptions code in the utils directory of the EMOD source code within the files Exceptions.h and Exceptions.cpp.
Each exception will return, at a minimum, the following information:
Exception type caught
The filename where the exception occurred
The line number in the file where the exception was thrown (which may not be exactly where the error actually occurred in the code)
The function name where the exception was thrown
Specific exceptions may also return additional information in a message format. This message (msg) may contain variable names or the values of those variables, the name of a file that wasn’t found, and other informational text regarding the nature of the problem. For example, a file not found exception (FileNotFoundException) might look similar to the following:
00:00:00 [0] [E] [Eradication] FileNotFoundException caught: msg = Could not find
file ./Namawala_single_node_demographics.compiled.json, filename = NodeDemographics.cpp,
lineNumber = 227
BadEnumInSwitchStatementException¶
This exception is thrown when an enumeration value is not handled in the switch statement. In other words, this exception signals that there is a problem with an enumeration value, typically stemming from one of the configuration files, though this is not always the case.
It is possible the enumeration value is not a valid value (out of range of the numeric range of the enumeration), or perhaps the string value is currently not implemented in the code (and should not be used in configuration settings yet).
BadMapKeyException¶
This exception is thrown when there is a standard template library (STL) mapping error. Usually, this occurs where spatial output channel names are specified in the configuration file if an unrecognized channel name is used.
If you have not modified the EMOD source code or used an unrecognized channel name, this error could signal an internal problem with the code. Contact support@idmod.org.
CalculatedValueOutOfRangeException¶
EMOD performs a large number of mathematical operations on parameter values. Therefore, it is possible that, despite original parameter values being with range, the values resulting from these multiple calculations may end up outside its valid range. For example, a probability value (range: 0 to 1.0) that after multiple calculations during a simulation now exceeds 1.0.
This exception is thrown when such a situation is detected. This exception only applies to numeric or Boolean values, not enumeration values.
ConfigurationRangeException¶
This exception is thrown if a parameter value read from a configuration file is detected to be outside its valid range. This exception is similar to the more general case OutOfRangeException. However, this exception is only thrown if the out of range exception comes from a numeric or Boolean value, not enumeration value, read from a configuration file.
DllLoadingException¶
The EMOD architecture is modularized and can be built as a core Eradication.exe along with a series of custom reporter EMODules, as opposed to a single monolithic Eradication.exe. This exception indicates that EMOD couldn’t load on of the EMODule DLLs.
This situation could occur for several reasons:
EMOD couldn’t find the EMODule
Unresolved symbols were found (Windows)
EMOD could not find the necessary symbol during the GetProcAddress call
The custom reporter EMODules were built using a version of Visual Studio that is no longer supported. Rebuild the EMODules using Visual Studio 2015 (Professional, Premium, or Ultimate).
FactoryCreateFromJsonException¶
EMOD implements class factories that instantiate objects at run time and use information from JSON- based configuration information in the creation of these objects. The exception indicates something is incorrect with the JSON information.
In particular, in some cases, the JSON information is nested into a hierarchy of information. Therefore, as the factories are called to create the objects described by the outer layers of one of these nested hierarchies, the factories do not have any knowledge yet of the inner layers of the hierarchies. This inner information contains information the factory needs to complete the object instantiation, but this information might not be correct. If that happens, then the factory will throw this exception.
Campaign files often have this kind of nested hierarchical structure, so it’s important to t verify that the hierarchy is set up correctly. For example, if the class name were mistyped and EMOD had no implementation of that class, this exception will be thrown.
FileIOException¶
This exception is generated if there is an unrecoverable problem loading data from a file. The data might be corrupted or there may be a mismatch. For example, if the file format or configuration information indicates that there should be ten values of some array and there are only nine included in the file, then this exception would be thrown.
This exception is not the same as the exception thrown for a file that is not found. In this case, the file is found and loaded, but there is a problem with the data in the file.
FileNotFoundException¶
This exception is thrown if a file cannot be found. Possible causes might include a incorrectly typed filename in the configuration file, a wrong path to the file, or even the path not being set in the system environment leading to the system not finding a relative path to the file. One of the most likely causes is that quotes are missing around the file name.
GeneralConfigurationException¶
This exception is only thrown if a more specific exception cannot be used for the configuration problem detected. This exception is likely thrown when there is very little information available about the root problem.
For example, this exception might be thrown if a parameter name is invalid, such as using an older, deprecated version of a parameter name.
IllegalOperationException¶
This exception is thrown if an illegal operation was detected. In most cases, a more specific exception will be thrown rather than this more general one. This exception is likely thrown when there is very little information available about the root problem. For example, when a utility function error is detected, there is very sparse information available as to what may have led to the error. As a result, calling a more specific exception with more context is not an option.
IncoherentConfigurationException¶
This exception is thrown if mutually contradictory or incompatible configuration settings have been detected. For example, if mutually exclusive parameters are set, the minimum parameter value is greater than the maximum value, or two distribution axes are specified in a demographics file but there is a mismatch with the number of axes scale factors included. The exception can also occur if there isn’t a corresponding mapping between an reference ID in the metadata of a demographics file and its corresponding data file.
InitializationException¶
This exception is thrown if a problem with initialization was detected. In most cases, a more specific exception will be thrown rather than this more general one. This exception is likely thrown when there is very little information available about the root problem.
For example, if the very first part of a JSON file has corrupted or badly formatted data, this exception may be thrown instead of the more expected file input/output exception, FileIOException.
InvalidInputDataException¶
This exception is thrown when a problem with an input data file is detected. For example, if the wrong data type was detected, such as a float being detected when a string is expected you would see this exception thrown, or even, if a parameter has an invalid value even if the value is of the correct type. As the input data file most likely to have significant modifications, verify that the demographics file is set up correctly.
MissingParameterFromConfigurationException¶
This exception occurs when required parameters are missing. Verify that you are not using deprecated parameters and that all required parameters are specified (or set Use_Defaults to 1).
MPIException¶
This exception is thrown if there is an MPI error. As such, these types of issues are related to interfacing with MPI (and/or networking issues) and do not necessary imply something wrong with the EMOD code or JSON files.
NotYetImplementedException¶
This exception is thrown if an attempt is made to execute code that is not yet implemented. For example, there are areas of EMOD where placeholder enumeration values are defined but not yet implemented. If you specify a value like this, it is considered within a valid range, but this exception will be thrown in response. Verify that any enumeration values use one of the available values as described in the documentation and do not contain any typos.
NullPointerException¶
This exception is thrown when a NULL pointer is detected in the code, or rather when a NULL pointer - that should NOT be NULL - is used. When thrown at the application level, a NULL pointer exception is usually caused by some sort of initialization error, for example, a file not being found.
As a result, in most cases, a more specific exception will be thrown before the code execution reaches a point where this exception would occur. Therefore, this exception is uncommon and likely thrown only when there is very little information available about the root problem.
OutOfRangeException¶
This exception is thrown when a numeric or Boolean value is out of range. For example, if you index an array outside of its valid range, this exception will be thrown. There are other situations where more specific exceptions are thrown instead of this more general one. For example, when the numeric or Boolean values are from a configuration file, but are detected to be out of range, the ConfigurationRangeException is thrown. Likewise, if the value goes out of range as the result of a calculation, the CalculatedValueOutOfRangeException is thrown instead.
QueryInterfaceException¶
The EMOD architecture is modularized and many components now implement QueryInterface. This exception is thrown when a required interface is queried on an object and the object that returns that the interface is not supported.
If you have not modified the EMOD source code and receive this error, it could signal an internal problem with the code. Contact support@idmod.org.
SerializationException¶
This exception is thrown when there is a serialization (or de-serialization) issue. For example, if data is being passed over the network (MPI) and the connection drops, then the serialization fails and this exception is thrown.
CentOS on Azure environment¶
The following problems are specific to running simulations using the Eradication binary on CentOS 7.1 on Azure.
Regression test graphs differ when run on CentOS on Azure¶
After you run regression simulations on CentOS on Azure using runemod.sh in the Scripts directory, it plots graphs from the simulation output data with a red line for the reference output and a blue line for the new output. The reference output was created by running the simulation on Windows, which in some cases may be slightly different than the output from CentOS on Azure.
For simulations that plot a baseline, you can override the Windows reference output by modifying runemod.sh to use output/InsetChart.linux.json as the output location. In that case, the red reference plots should always be completely covered by the blue plots.
Eradication binary cannot locate the input data files¶
If you chose not to have the PrepareLinuxEnvironment.sh script download the EMOD source code and input data files, you need to set up the environment variable, path and symlink that are needed to run simulations on CentOS on Azure. See Install EMOD on CentOS on Azure.
Glossary¶
The Epidemiological MODeling software (EMOD) glossary is divided into the following subsections that define terms related to software usage, general epidemiology, and the particular disease being modeled.
Software terms¶
The following terms are used to describe both general computing processes and concepts and the files, features, and functionality related to running simulations with Epidemiological MODeling software (EMOD).
- agent-based model
A type of simulation that models the actions and interactions of autonomous agents (both individual and collective entities such as organizations or groups).
- Boost
Free, peer-reviewed, portable C++ source libraries aimed at a wide range of uses including parallel processing applications (Boost.MPI). For more information, please see the Boost website, http://www.boost.org.
- campaign
A collection of events that use interventions to modify a simulation.
- campaign event
A JSON object that determines when and where an intervention is distributed during a campaign.
- campaign file
A JavaScript Object Notation (JSON) formatted file that contains the parameters that specify the distribution instructions for all interventions used in a campaign, such as diagnostic tests, the target demographic, and the timing and cost of interventions. The location of this file is specified in the configuration file with the Campaign_Filename parameter. Typically, the file name is campaign.json.
- channel
A property of the simulation (for example, “Parasite Prevalence”) that is accumulated once per simulated time step and written to file, typically as an array of the accumulated values.
- class factory
A function that instantiate objects at run-time and use information from JSON-based configuration information in the creation of these objects.
- configuration file
A JavaScript Object Notation (JSON) formatted file that contains the parameters sufficient for initiating a simulation. It controls many different aspects of the simulation, such as population size, disease dynamics, and length of the simulation. Typically, the file name is config.json.
- core
In computing, a core refers to an independent central processing unit (CPU) in the computer. Multi-core computers have more than one CPU. However, through technologies such as Hyper- Threading Technology (HTT or HT), a single physical core can actually act like two virtual or logical cores, and appear to the operating system as two processors.
- demographics file
A JavaScript Object Notation (JSON) formatted file that contains the parameters that specify the demographics of a population, such as age distribution, risk, birthrate, and more. IDM provides demographics files for many geographic regions. This file is considered part of the input data files and is typically named <region>_demographics.json.
- disease-specific build
A build of the EMOD executable (Eradication.exe) built using SCons without any dynamic link libraries (DLLs).
- dynamic link library (DLL)
Microsoft’s implementation of a shared library, separate from the EMOD executable (Eradication.exe), that can be dynamically loaded (and unloaded when unneeded) at runtime. This loading can be explicit or implicit.
- EMODule
A modular component of EMOD that are consumed and used by the EMOD executable (Eradication.exe). Under Windows, a EMODule is implemented as a dynamic link library (DLL) and, under CentOS on Azure, EMODules are currently not supported. EMODules are primarily custom reporters.
- Epidemiological MODeling software (EMOD)
The modeling software from the Institute for Disease Modeling (IDM) for disease researchers and developers to investigate disease dynamics, and to assist in combating a host of infectious diseases. You may see this referred to as Disease Transmission Kernel (DTK) in the source code.
- Eradication.exe
Typical (default) name for the EMOD executable (Eradication.exe), whether built using monolithic build or modular (EMODule-enabled) build.
- event coordinator
A JSON object that determines who will receive a particular intervention during a campaign.
- flattened file
A single campaign or configuration file created by combining a default file with one or more overlay files. Multiple files must be flattened prior to running a simulation. Configuration files are flattened to a single-depth JSON file without nesting, the format required for consumption by the EMOD executable (Eradication.exe). Separating the parameters into multiple files is primarily used for testing and experimentation.
- Heterogeneous Intra-Node Transmission (HINT)
A feature for modeling person-to-person transmission of diseases in heterogeneous population segments within a node for generic simulations.
- high-performance computing (HPC)
The use of parallel processing for running advanced applications efficiently, reliably, and quickly.
- input data files
Fixed information used as input to a simulation. Examples of such data are population, altitude, demographics, and transportation migrations. The configurable config.json and campaign.json files do not fall into this category of static fixed input.
- intervention
An object aimed at reducing the spread of a disease that is distributed either to an individual; such as a vaccine, drug, or bednet; or to a node; such as a larvicide. Additionally, initial disease outbreaks and intermediate interventions that schedule another intervention are implemented as interventions in the campaign file.
- JavaScript Object Notation (JSON)
A human-readable, open standard, text-based file format for data interchange. It is typically used to represent simple data structures and associative arrays, and is language- independent.
- Keyhole Markup Language (KML)
A file format used to display geographic data in an Earth browser, for example, Google Maps. The format uses an XML-based structure (tag-based structure with nested elements and attributes). Tags are case-sensitive.
- Link-Time Code Generation (LTCG)
A method for the linker to optimize code (for size and/or speed) after compilation has occurred. The compiled code is turned not into actual code, but instead into an intermediate language form (IL, but not to be confused with .NET IL which has a different purpose). The LTCG then, unlike the compiler, can see the whole body of code in all object files and be able to optimize the result more effectively.
- Message Passing Interface (MPI)
An interface used to pass information between computing cores in parallel simulations. One example is the migration of individuals from one geographic location to another within EMOD simulations.
- microsolver
A microsolver is a type of “miniature model” that operates within the framework of EMOD to compute a particular set of parameters. Each microsolver, in effect, is creating a microsimulation in order to accurately capture the dynamics of that particular aspect of the model.
- Monte Carlo
Monte Carlo methods, or Monte Carlo experiments, are a class of computational algorithms utilizing repeated random sampling to obtain numerical results. Monte Carlo simulations create probability distributions for possible outcome values, which provides a more realistic way of describing uncertainty in variables.
- monolithic build
A single EMOD executable (Eradication.exe) with no DLLs that includes all components as part of Eradication.exe itself. You can still use EMODules with the monolithic build; for example, a custom reporter is a common type of EMODule. View the documentation on EMODules and emodules_map.json for more information about creation and use of EMODules.
- node
A grid size that is used for modeling geographies. Within EMOD, a node is a geographic region containing simulated individuals. Individuals migrate between nodes either temporarily or permanently using mobility patterns driven by local, regional, and long- distance transportation.
- node-targeted intervention
An intervention that is distributed to a geographical node rather than to a single individual. One example is larvicides, which affect all mosquitoes living and feeding within a given node.
- output report
A file that is the output from an EMOD simulation. Output reports are in JSON, CSV, or binary file format. You must pass the data from an output report to graphing software if you want to visualize the output of a simulation.
- overlay file
An additional configuration, campaign, or demographic file that overrides the default parameter values in the primary file. Separating the parameters into multiple files is primarily used for testing a nd experimentation. In the case of configuration and campaign files, the files can use an arbitrary hierarchical structure to organize parameters into logical groups. Configuration and campaign files must be flattened into a single file before running a simulation.
- preview
Software that undergoes a shorter testing cycle in order to make it available more quickly. Previews may contain software defects that could result in unexpected behavior. Use EMOD previews at your own discretion.
- regression test
A test to verify that existing EMOD functionality works with new updates, typically used to refer to one of a set of scenarios included in the EMOD bundle and located in the Regression subdirectory of the bundle. Directory names of each subdirectory in Regression describe the main regression attributes, for example, “1_Generic_Seattle_MultiNode”. Also can refer to the process of regression testing of software.
- release
Software that includes new functionality, scientific tutorials leveraging new or existing functionality, and/or bug fixes that have been thoroughly tested so that any defects have been fixed before release. EMOD releases undergo full regression testing.
- reporter
Functionality that extracts simulation data, aggregates it, and saves it as an output report. EMOD provides several built-in reporters for outputting data from simulations and you also have the ability to create a custom reporter.
- scenario
A collection of input, configuration, and campaign files that describes a real-world example of a disease outbreak and interventions. Many scenarios are included with both the Quick Start and EMOD source installations. The Quick Start documentation includes tutorials that describe how to run simulations using these scenarios to learn more about epidemiology and disease modeling.
- schema
A text or JSON file that can be generated from the EMOD executable (Eradication.exe) that defines all configuration and campaign parameters.
- simulation
An EMOD software execution performed by the kernel. Each simulation has an associated set of files that control the inputs, configuration, and campaign.
- simulation type
The disease or disease class to model.
EMOD supports the following simulation types for modeling a variety of diseases:
Generic disease (GENERIC_SIM)
Vector-borne diseases (VECTOR_SIM)
Malaria (MALARIA_SIM)
Tuberculosis (TB_SIM)
Sexually transmitted infections (STI_SIM)
HIV (HIV_SIM)
- solvers
Solvers are used to find computational solutions to problems. In simulations, they can be used, for example, to determine the time of the next simulation step, or to compute the states of a model at particular time steps.
- Standard Template Library (STL)
A library that contains a set of common C++ classes (including generic algorithms and data structures) that are independent of container and implemented as templates, which enables compile-time polymorphism (often more efficient than run-time polymorphism). For more information and discussion of STL, see Wikipedia.
- state transition event
A change in state (e.g. healthy to infected, undiagnosed to positive diagnosis, or birth) that may trigger a subsequent action, often an intervention. “Campaign events” should not be confused with state transition events.
- time step
A discrete number of hours or days in which the “simulation states” of all “simulation objects” (interventions, infections, immune systems, or individuals) are updated in a simulation. Each time step will complete processing before launching the next one. For example, a time step would process the migration data for populations moving between nodes via rail, airline, and road. The migration of individuals between nodes is the last step of the time step after updating states.
- tutorial
A set of instructions in the Quick Start documentation to learn more about epidemiology and disease modeling. Tutorials are based on real-world scenarios and demonstrate the mechanics of the the model. Each tutorial consists of one or more scenarios.
- working directory
The directory that contains the configuration and campaign files for a simulation. You must be in this directory when you invoke Eradication.exe at the command line to run a simulation.
Epidemiology terms¶
The following terms are used to describe general concepts and processes in the field of epidemiology and disease modeling.
- antigen
A substance that is capable of inducing a specific immune response and that evokes the production of one or more antibodies.
- Clausius-Clayperon relation
A way of characterizing a transition between two phases of matter; provides a method to find a relationship between temperature and pressure along phase boundaries. Frequently used in meteorology and climatology to describe the behavior of water vapor. See Wikipedia - Clausius-Clayperon relation for more information.
- deterministic
Characterized by the output being fully determined by the parameter values and the initial conditions. Given the same inputs, a deterministic model will always produce the same output.
- diffusive migration
The diffusion of people in and out of nearby nodes by foot travel.
- epitope
The portion of an antigen that the immune system recognizes. An epitope is also called an antigenic determinant.
- Euler method
Used in mathematics and computational science, this method is a first-order numerical procedure for solving ordinary differential equations with a given initial value.
- exp(
The exponential function, \(e^x\), where \(e\) is the number (approximately 2.718281828) such that the function \(e^x\) is its own derivative. The exponential function is used to model a relationship in which a constant change in the independent variable gives the same proportional change (i.e. percentage increase or decrease) in the dependent variable. The function is often written as \(exp(x)\). The graph of \(y = exp(x)\) is upward-sloping and increases faster as \(x\) increases.
- herd immunity
The resistance to the spread of a contagious disease within a population that results if a sufficiently high proportion of individuals are immune to the disease, especially through vaccination.
- incidence
The rate of new cases of a disease during a specified time period. This is a measure of the risk of contracting a disease.
- Koppen-Geiger Climate Classification System
A system based on the concept that native vegetation is a good expression of climate. Thus, climate zone boundaries have been selected with vegetation distribution in mind. It combines average annual and monthly temperatures and precipitation, and the seasonality of precipitation. The DTK has several options for configuring the climate, namely air temperature, rainfall, and humidity. One option utilizes Input Files that associate geographic nodes with Koppen climate indices. The modified Koppen classification uses three letters to divide the world into five major climate regions (A, B, C, D, and E) based on average annual precipitation, average monthly precipitation, and average monthly temperature. Each category is further divided into sub-categories based on temperature and precipitation. While the Koppen system does not take such things as temperature extremes, average cloud cover, number of days with sunshine, or wind into account, it is a good representation of our earth’s climate.
- loss to follow-up (LTFU)
Patients who at one point were actively participating in disease treatment or clinical research, but have become lost either by error or by becoming unreachable at the point of follow-up.
- ordinary differential equation (ODE)
A differential equation containing one or more functions of one independent variable and its derivatives.
- prevalence
The rate of all cases of a disease during a specified time period. This is a measure of how widespread a disease is.
- reproductive number
In a fully susceptible population, the basic reproductive number R:sub:0is the number of new cases generated by one infectious case over the course of the infectious period. The effective reproductive number takes into account non-susceptible individuals.
- routine immunization (RI)
The standard practice of vaccinating the majority of susceptible people in a population against vaccine-preventable diseases.
- stochastic
Characterized by having a random probability distribution that may be analyzed statistically but not predicted precisely.
- subpatent
When an individual is infected but asymptomatic, so the infection is not readily detectable.
- SEIR model
A generic epidemiological model that provides a simplified means of describing the transmission of an infectious disease through individuals where those individuals can pass through the following five states: susceptible, exposed, infectious, and recovered.
- SEIRS model
A generic epidemiological model that provides a simplified means of describing the transmission of an infectious disease through individuals where those individuals can pass through the following five states: susceptible, exposed, infectious, recovered, and susceptible.
- SI model
A generic epidemiological model that provides a simplified means of describing the transmission of an infectious disease through individuals where those individuals can pass through the following five states: susceptible and infectious.
- SIR model
A generic epidemiological model that provides a simplified means of describing the transmission of an infectious disease through individuals where those individuals can pass through the following five states: susceptible, infectious, and recovered.
- SIRS model
A generic epidemiological model that provides a simplified means of describing the transmission of an infectious disease through individuals where those individuals can pass through the following five states: susceptible, infectious, recovered, and susceptible.
- SIS model
A generic epidemiological model that provides a simplified means of describing the transmission of an infectious disease through individuals where those individuals can pass through the following five states: susceptible, infectious, and susceptible.
- supplemental immunization activity (SIA)
In contrast to routine immunization (RI), SIAs are large-scale operations with a goal of delivering vaccines to every household.
- WAIFW matrix
A matrix of values that describes the rate of transmission between different population groups. WAIFW is an abbrevation for Who Acquires Infection From Whom.
Malaria and vector terms¶
The following terms are used to describe the biological processes involved in malaria and other vector-borne diseases.
- circumsporozoite protein (CSP)
One of two proteins (the other being thrombospondin-related adhesive protein) involved in sporozoite recognition of host cells in malaria.
- cohort model
A vector model that tracks properties of mosquito “cohorts”, for example, population, age, gender, fertility, etc.
- cytokine
Cytokines are small secreted proteins that are released by cells, and have specific effects on interactions and communications between cells. They are especially important in the immune system, and can function in inflammatory (and anti-inflammatory) response.
- CRISPR
A genome-editing technique that utilizes the CRISPR/Cas9 system (a prokaryotic immune system) that can be used to splice DNA sequences in very specific target locations, enabling the precise insertion or removal of genes of interest.
- entomological inoculation rate (EIR)
The commonly-used measure of the intensity of malaria transmission. EIR is equal to the number of mosquito bites per night times the proportion of those bites that test positive for sporozoites.
- gametocyte
The male and female sexual forms of the blood-stage malaria parasite. During blood-feeding, gametocytes are taken up into the mosquito mid-gut where they commence the mosquito-phase of the parasite life cycle.
- gene drive mosquito
Gene drive is a genetic technique that promotes the inheritance of particular genes or specific genetic elements. In mosquitoes, approaches are being tested that work to suppress the spread of malaria by creating mosquito lines through fertility suppression, driving-Y chromosomes, or express genes that limit malaria transmission.
- gonotrophic
A female mosquito’s feeding and egg-laying cycle, where the duration is defined as the average number of days that gravid mosquitoes take to oviposit after taking a blood meal.
- hepatocyte
The cell type that comprises liver tissue.
- homing endonuclease genes (HEGs)
“Selfish” genetic elements that can spread through the mosquito population even at a cost to the host. The HEG allele status is a tracked property of mosquitoes. Large numbers of HEG- positive mosquitoes may be introduced into the simulation by a MosquitoRelease intervention.
- individual mosquito model
A vector model simulation where every individual mosquito in the population is modeled, as well as a simulation of a sampled subset of mosquitoes to represent the population as the whole.
- indoor residual spraying (IRS)
The process of spraying insecticide on the interior surfaces of dwellings in order to repel or kill mosquitoes.
- insecticide-treated nets (ITN)
Bednets hung over sleeping areas that are treated with insecticide to repel mosquitoes and kill those that land on them.
- larval habitat
Mosquito larval ecology includes several habitat types including brackish swamp, constant, human population, temporary rainfall, and semi-permanent water vegetation.
- larvicide
A node-targeted intervention that may be configured to have a direct killing effect on the larval population and a temporary reduction on available larval habitat.
- mass drug administration (MDA)
The treatment of an entire population in a geographic area with a curative dose of an antimalarial drug without first testing for infection, and regardless of the presence of symptoms.
- merozoite
The malaria parasite stage that invades red blood cells. After each cycle of asexual reproduction within the infected red blood cell, merozoites are released into the bloodstream to invade new red blood cells.
- merozoites
see “merozoite”
- merozoite surface protein (MSP)
A type of protein integral to red blood cell invasion by malaria parasites that is an important target of the human immune response.
- oviposition
to deposit or lay eggs.
- pre-erythrocytic
Stages of a malaria parasite that occur before the red blood cells are invaded. This includes the period starting from inoculation with sporozoites, through the liver-stage multiplication, up to infected hepatocyte rupture, and the first invasion of erythrocytes by merozoites.
- pre-erythrocytic vaccine (PEV)
A vaccine targeted at the earliest stage of malaria parasite development, namely before the blood-stage infection. Other vaccines target the asexual and sexual stages of the parasite.
- Plasmodium falciparum erythrocyte membrane protein 1 (PfEMP1)
A protein encoded by the malaria parasite and expressed on the surface of infected red blood cells. PfEMP1 adheres to various human receptors, allowing sequestration of parasitized red blood cells and avoidance of clearance in the spleen. PfEMP1 is an important antigenic target of the human adapted immune response. It exhibits a high degree of antigenic variation via a genetic switching mechanism evolved to escape immune pressure.
- schizont
Cells formed during the asexual stage of the life cycle of sporozoan protozoans, such as the malaria parasite.
- spatial repellent
An intervention that may be individually-distributed (protective indoors and outdoors), or node-distributed at the community-level, that repels mosquitoes from biting.
- sporogony
The asexual process of spore formation in parasitic sporozoans.
- sporozoite
a motile, spore-like stage of the malaria parasite that is the infectious agent introduced into the host.
- sugar-baited trap
Bait stations, often placed inside houses, that have an attractive and toxic sugar used to attract and kill sugar-feeding male and female mosquitoes.
- superinfection
The simultaneous infection with multiple strains of the same pathogen.
- transmission-blocking vaccine (TBV)
A vaccine that blocks gametocytes from infected humans by producing viable sporozoites in mosquitoes. Also referred to as a “sexual-stage” vaccine, since vaccines in this class block gametocytes from infected humans by producing viable sporozoites in mosquitoes.
- vaccine intervention types
In EMOD, vaccine intervention types are either “AcquisitionBlocking”, “TransmissionBlocking” or “MortalityBlocking”. For example, “mode-of-action”, “targeted parasite stages”, (such as pre-erythrocytic, asexual, or sexual), etc.
- vector
Insect or other living carrier that transmits an infectious agent. The vector life cycle and feeding cycle in the model are described in the article Eckhoff, Malaria Journal 2011, 10:303.
- vectorial capacity
The daily rate of all mosquitoes that would be infected after biting a single infectious host. This measure can be used to describe the transmission intensity of malaria, and serves as a rate at which future infections arise from a currently infected host.
- vector model
The vector model includes both a cohort model and an individual mosquito model. The model is a closed-loop feeding cycle where successful animal feeds (both indoors and outdoors) produce eggs, larva, vector mating, and then, adult mosquitoes. The vector model inherits the same human- infection model structure from the generic simulation type: uninfected, latent incubation, infectious, multiple immune variables, and superinfection. However, the transmission of infections is not between individual humans, but rather via the human-to-vector and vector- to-human pathways.
- Wolbachia
This is a genus of bacteria that infects many arthopod species (especially insects). It has been discovered that insects infected with Wolbachia have increased resistance to viral or other parasitic infections. Work is underway to determine if strains of modified Wolbachia can be used to control malaria, by effectively preventing mosquitoes from becoming infected with Plasmodium parasites.
EMOD source code installation¶
This section describes how to install the software needed to build the EMOD executable (Eradication.exe) or Eradication binary from the EMOD source code. This is necessary if you want to modify the source code to extend the capabilities of the model beyond what’s available with the latest EMOD release. For example, you may want to model a disease that isn’t currently supported in EMOD. You can build Eradication.exe from source code on computers running Windows 10, Windows Server 12, and Windows HPC Server 12 (64-bit) or build the Eradication binary on computers running CentOS 7.1 on Azure.
After building, you should run a single simulation to verify basic functionality. We recommend the 27_Vector_Sandbox scenario in the Regression directory, which is a simple five-year vector simulation with an indoor residual spraying (IRS) campaign in the third year, executed on a single-node geography that is based on Namawala, Tanzania. This simulation generally takes a few minutes to execute.
However, you can run a simulation using any of the subdirectories under Regression, which when used with the input data files provided by IDM, contain complete sets of files for EMOD simulations. See Run simulations and Use input data files for more information.
After that, we recommend running full regression tests to further verify that EMOD is behaving as expected and that none of the source code changes inadvertently changed the EMOD functionality. See Regression testing for more information.
The EMOD executable (Eradication.exe) is tested using Windows 10, Windows Server 12, and Windows HPC Server 12 (64-bit). Windows HPC Server is used for testing remote simulations on a high-performance computing (HPC) cluster and the other Windows operating systems are used to test local simulations.
The Eradication binary is tested and supported on a CentOS 7.1 on Azure virtual machine. It has also been successfully built and run on Ubuntu, SUSE, and Arch, but has not been tested and is not supported on those Linux distributions.
Prerequisites for working with EMOD source code¶
This section describes the software packages or utilities must be installed to build the EMOD executable (Eradication.exe) or Eradication binary from source code and run regression tests. It assumes you have already installed the required software to run EMOD simulations. If you have not, see EMOD installation.
If additional software is needed for the prerequisite software due to your specific environment, the installer for the prerequisite software should provide instructions. For example, if Microsoft MPI v8 requires additional Visual C++ redistributable packages, the installer will display that information.
Note
IDM does not provide support or guarantees for any third-party software, even software that we recommend you install. Send feedback if you encounter any issues, but any support must come from the makers of those software packages and their user communities.
Install prerequisites for building Eradication.exe¶
The following software packages are required to build the Eradication.exe from EMOD source code on Windows 10, Windows Server 12, and Windows HPC Server 12 (64-bit).
Visual Studio¶
Purchase a license from Microsoft or use an MSDN subscription to install Visual Studio 2015 (Professional, Premium, or Ultimate).
While you can use a free copy of Visual Studio Community, IDM does not test on or support this version.
Select the Visual C++ tools for Windows desktop during installation.
Python¶
Python is required for building the disease-specific Eradication.exe and running Python scripts.
Note
EMOD does not support the current version of Python.
Install Python 2.7.11 or 2.7.12. See https://www.python.org/downloads/ for instructions.
In the Customize Python dialog box, verify that Add python.exe to PATH is selected to add Python to the PATH environment variable on your computer.
From Control Panel, select Advanced system settings, and then click Environment Variables.
Add a new variable called IDM_PYTHON_PATH and set it to the directory where you installed Python, and then click OK.
Open a Command Prompt window and type the following to verify installation:
python --version
The Python package manager, pip, is installed as part of Python 2.7.11 or 2.7.12 and is used to install other software packages.
HPC SDK¶
Install the Microsoft HPC Pack 2012 SDK (64-bit). See https://www.microsoft.com/en-us/download/details.aspx?id=36043 for instructions.
Boost¶
Go to https://sourceforge.net/projects/boost/files/boost/1.61.0/ and select one of the compressed files.
Unpack the libraries to the location of your choice.
From Control Panel, select Advanced system settings, and then click Environment Variables.
Add a new variable called IMD_BOOST_PATH and set it to the directory where you installed Boost, and then click OK.
SCons¶
SCons is required for the building disease-specific Eradication.exe and is optional for the monolithic Eradication.exe that includes all simulation types.
Go to http://www.lfd.uci.edu/~gohlke/pythonlibs#scons and select the WHL file for SCons 2.5.0 that is compatible with Python 2.7.11 or 2.7.12.
Save the file to your Python installation directory.
Open a Command Prompt window and navigate to the Python installation directory, then enter the following, substituting the name of the specific matplotlib WHL file you downloaded:
pip install scons-2.x.x-py2-none-any.whl
Install prerequisites for building the Eradication binary¶
For CentOS 7.1 on Azure, all prerequisites for building the Eradication binary are installed by the setup script described in Install EMOD on CentOS on Azure. If you originally installed EMOD without including the source code and input data files that are optional for running simulations using a pre-built Eradication binary, rerun the script and install those.
Install prerequisites for running regression tests¶
The following plotting software is required for running regression tests, where graphs of model output are created both before and after source code changes are made to see if those changes created a discrepancy in the regression test output. For more information, see Regression testing. For CentOS 7.1 on Azure, all Python packages are installed by the setup script, but you may want to install R or MATLAB.
We recommended that you download some of the Python packages from http://www.lfd.uci.edu/~gohlke/pythonlibs, a page compiled by Christoph Gohlke, University of California, Irvine. The libraries there are compiled Windows binaries, including the 64-bit versions required by EMOD.
Python utilities¶
The Python utilities dateutil, six, and pyparsing provide text parsing and datetime functionality.
Open a Command Prompt window.
Enter the following commands:
pip install python-dateutil pip install pyparsing
NumPy¶
The NumPy package adds support for large, multi-dimensional arrays and matrices to Python.
Go to http://www.lfd.uci.edu/~gohlke/pythonlibs/#numpy and select the WHL file for NumPy 1.11.3 (64-bit) that is compatible with Python 2.7.11 or 2.7.12.
Save the file to your Python installation directory.
Open a Command Prompt window and navigate to the Python installation directory, then enter the following, substituting the name of the specific NumPy WHL file you downloaded:
pip install numpy-1.x.x+mkl-cp27m-win_amd64.whl
Matplotlib¶
Matplotlib is a Python library for making publication quality plots using syntax familiar to MATLAB users. Matplotlib also uses NumpPy for numeric manipulation. Output formats include PDF, Postscript, SVG, and PNG, as well as a screen display.
Go to http://www.lfd.uci.edu/~gohlke/pythonlibs/#matplotlib and select the WHL file for Matplotlib 1.5.3 (64-bit) that is compatible with Python 2.7.11 or 2.7.12.
Save the file to your Python installation directory.
Open a Command Prompt window and navigate to the Python installation directory, then enter the following, substituting the name of the specific matplotlib WHL file you downloaded:
pip install matplotlib-1.x.x+mkl-cp27m-win_amd64.whl
R¶
The IDM test team uses R 3.2.0 (64-bit) for regression testing, but it is considered optional.
R is a free software environment for statistical computing and graphics.
Go to https://www.r-project.org/ and install R 3.2.0 (64-bit).
MATLAB¶
The IDM test team uses MATLAB R2015a and the MATLAB Statistics and Machine Learning Toolbox™ R2015a for regression testing, but they are both considered optional.
MATLAB is a high-level language and interactive environment for numerical computation, visualization, and programming. The MATLAB Statistics and Machine Learning Toolbox™ provides functions and applications to describe, analyze and model data using statistics and machine learning algorithms.
Go to http://www.mathworks.com/products/matlab/ and install MATLAB R2015a.
If desired, go to https://www.mathworks.com/products/statistics.html and install the MATLAB Statistics and Machine Learning Toolbox™ R2015a.
Download the EMOD source code¶
The EMOD source code is available on GitHub. The EMOD source includes the source code, Visual Studio solution, sample configuration files, as well as regression test and other files needed to fully build and test the EMOD executable (Eradication.exe).
You can have multiple versions of the EMOD source code in separate directories on your local computer. For example, you might want to download a new release of EMOD but also keep a previous release of the source. In the following examples, the source code is downloaded to the directory EMOD at C:/IDM, but you can save to any location you want.
You can use a Git client of your choice to download from GitHub, however, these instructions show how to use Git GUI and Git Bash.
Install Git GUI and Git Bash¶
To install Git GUI and Git Bash, download a 64-bit version of Git from https://git-scm.com/download. On the Select Components installer window, you can select one or both of Git GUI Here for a GUI or Git Bash Here for a command window.
Use Git GUI to download the EMOD source¶
Launch the Git GUI application and click Clone Existing Repository.
From the Clone Existing Repository window:
In Source Location, enter https://github.com/InstituteforDiseaseModeling/EMOD
In Target Directory, enter the location and target directory name: C:/IDM/EMOD
Click Clone. Git GUI will create the directory and download the source code.
Close the Git GUI window when the download completes.
Use Git Bash to download the EMOD source¶
Note
For a list of the Git Bash commands, you can type git help git from Git Bash, or git help
<command> for information about a specific command.
Launch the Git Bash application. From the command line:
Go to the location where you want your copy of the EMOD source located:
cd C:\IDM
Clone the repository from GitHub:
git clone https://github.com/InstituteforDiseaseModeling/EMOD
Git Bash will copy the EMOD source to a directory named EMOD by default as a different directory name was not specified after the URL.
Verify that all directories on https://github.com/InstituteforDiseaseModeling/EMOD are now reflected on your local clone of the repository. The table below describes each of the directories.
Directory |
Description |
|---|---|
Dependencies |
The ComputeClusterPack. |
Eradication |
Eradication project. Main project for Visual Studio solution. |
Regression |
Scripts to perform regression tests for modified code. It also includes a copy of the Quick Start Scenarios directory that contains simulation configuration, batch, input, and script files that are associated with the training tutorials. |
Scripts |
Scripts used in file creation. |
UnitTest++ |
Software for early release of additional regression testing. |
baseReportLib |
Static library of classes used by EMOD built-in reports and custom reports. |
cajun |
Cajun C++ API for JavaScript Object Notation (JSON) project. |
campaign |
Campaign project. |
componentTests |
Tests that work in conjunction with the UnitTest++ directory. |
interventions |
Interventions project. |
rapidjson |
Rapid JSON project. |
reporters |
Collection of custom reporter projects. |
snappy |
Compression utilities used by the EMOD software. |
utils |
Miscellaneous utilities. |
Additionally, the repository includes additional files, such as the Visual Studio solution file, SCons scripts, the README, and more.
Build EMOD from source code¶
You can build the Eradication.exe for Windows 10, Windows Server 12, and Windows HPC Server 12 (64-bit) using Microsoft Visual Studio or SCons. You can build the Eradication binary for CentOS 7.1 on Azure using SCons.
If you want full debugging support, you must build using Visual Studio. However, Visual Studio is only capable of a monolithic build that includes all supported simulation types.
EMOD supports the following simulation types for modeling a variety of diseases:
Generic disease (GENERIC_SIM)
Vector-borne diseases (VECTOR_SIM)
Malaria (MALARIA_SIM)
Tuberculosis (TB_SIM)
Sexually transmitted infections (STI_SIM)
HIV (HIV_SIM)
If you want to create a disease-specific build, you must build using SCons. However, SCons builds build only the release version without extensive debugging information.
Build a monolithic Eradication.exe with Visual Studio¶
You can use the Microsoft Visual Studio solution file in the EMOD source code repository to build the monolithic version of the EMOD executable (Eradication.exe), which can be either a release or debug build.
Open the EradicationKernel solution in Visual Studio.
From the ** Solution Configurations** ribbon, select Debug or Release.
On the Build menu, select Build Solution.
Eradication.exe will be in a subdirectory of the Eradication directory.
Build Eradication.exe or Eradication binary with SCons¶
SCons is a software construction tool that is an alternative to the well-known “Make” build tool. It is implemented in Python and the SCons configuration files, SConstruct and SConscript, are executed as Python scripts. This allows for greater flexibility and extensibility, such as the creation of custom SCons abilities just for EMOD. For more information on Scons, see www.scons.org.
Open a command window.
Go to the directory where EMOD is installed:
cd C:\IDM\EMOD
Run the following command to build Eradication.exe:
For a monolithic build:
scons --Release
For a disease-specific build, specify one of the supported disease types using the
--Diseaseflag:scons --Release --Disease=Vector
The executable will be placed, by default, in the subdirectory build\x64\Release\Eradication\ of your local EMOD source.
Additionally, you can parallelize the build process with the --jobs flag. This flag indicates
the number of jobs that can be run at the same time (a single core can only run one job at a time).
For example:
scons --Release --Disease=Vector --jobs=4
EMOD architecture¶
Epidemiological MODeling software (EMOD) is implemented in C++ and has two subsystems: the environment and the simulation controller. the environment contains the interfaces to the simulation controller subsystem and the external inputs and outputs. the simulation controller contains the epidemiological model (simulation and campaign management), and reporters that capture output data and create reports from a simulation. The following figure shows a high-level view of the system components of EMOD and how they are related to each other.
EMOD system components¶
Warning
If you modify the source code to add or remove configuration or campaign parameters, you may need to update the code used to produce the schema. You must also verify that your simulations are still scientifically valid.
Environment subsystem¶
The environment subsystem provides access to global resources and most of the external input and output interfaces. Output reports are the only output not handled through the environment subsystem. It consists of a a singleton utility class, Environment. The environment subsystem consists of five logical components.
The following figure shows a high-level view of the system components of EMOD and how they are related to each other.
EMOD system components¶
Input file readers¶
Provide small utility functions for reading in the input data, configuration, and campaign files. The actual parsing of the configuration and campaign files is done by the configuration manager component. The parsing of the input data files is done by the model classes that consume that data, which are part of the simulation controller subsystem.
Configuration manager¶
Parses the configuration files, stores the parsed values for system-global configuration values, and provides a central container resource for the configuration as JSON to those classes that need to parse out the remaining configuration values locally for themselves. It retains the contents of the configuration file and the directories provided on the command line.
Error handler¶
Provides an application-level exception library and a centralized mechanism for handling all errors that occur during program execution. These errors are ultimately reported through the logging component.
Logger¶
Writes the output logs and errors. Output reports, such as time-series channel data reports, spatial channel data reports, progress and status summary, and custom reports, are handled by the simulation controller. For information and setting the logging level, see Set log levels.
Message Passing Interface (MPI)¶
EMOD is designed to support large-scale multi-node simulations by running multiple instances of EMOD in parallel across multiple compute nodes on a cluster. During initialization, geographical nodes are assigned to compute nodes. Individuals who migrate between geographical nodes that are not on the same compute node are migrated via a process of serialization, network I/O, and deserialization. The network communication is handled through a mixture of direct calls to the MPI library and use of Boost’s MPI facilities. This component provides the system-wide single interface to MPI, caching the number of tasks and rank of current process from the MPI environment.
Simulation controller subsystem¶
The simulation controller is the top-level structure for the epidemiological model. The controller’s capabilities are simple, running a single simulation in a single time direction at a constant rate. It exists to support future capabilities, such as running multiple simulations, pausing a simulation, or bootstrapping a simulation from an archived simulation. It contains two components: simulation and reporters.
The following figure shows a high-level view of the system components of EMOD and how they are related to each other.
EMOD system components¶
Simulation¶
The simulation component contains core functionality that models the behavior of a disease without any interventions and extended functionality to include migration, climate, or other input data to create a more realistic simulation. Disease transmission may be more or less complex depending on the disease being modeled.
Campaign management¶
The simulation component also includes a campaign manager sub-component for including disease interventions in a simulation.
Reporter¶
The reporter component creates output reports for both simulation-wide aggregate reporting and spatial reporting.
Simulation core components¶
The simulation component contains core functionality that models the behavior of a disease without any interventions and extended functionality to include migration, climate, or other input data to create a more realistic simulation. Disease transmission may be more or less complex depending on the disease being modeled.
Warning
If you modify the source code to add or remove configuration or campaign parameters, you may need to update the code used to produce the schema. You must also verify that your simulations are still scientifically valid.
Each generic EMOD simulation contains the following core base classes:
- Simulation
Created by the simulation controller via a SimulationFactory with each run of EMOD.
- Node
Corresponds to a geographic area. Each simulation maintains a collection of one or more nodes.
- IndividualHuman
Represents a human being. Creates Susceptibility and Infection objects for the collection of individuals it maintains. The file name that defines this class is “Individual” and you may see it likewise shortened in diagrams.
- Susceptibility
Manages an individual’s immunity.
- Infection
Represents an individual’s infection with a disease.
For vector simulations, the following corresponding classes are derived from generic classes:
- SimulationVector
Creates NodeVector objects instead of generic Node objects.
- NodeVector
Creates IndividualHumanVector objects instead of generic IndividualHuman objects and creates and manages VectorPopulation objects to model the mosquito vectors.
- IndividualHumanVector
Represents a human being and provides the additional layer of functionality for how vectors interact with individual humans.
- VectorPopulation
The mosquito species at each node, which can be represented by a collection of cohorts that counts the population of a specific state of mosquitoes (VectorCohort) or by a collection of individual agent mosquitoes (VectorIndividual).
- SusceptibilityVector
Represents a human being’s susceptibility to vector-borne disease.
- InfectionVector
Represents a human being’s infection with a vector-borne disease.
Substitute these classes wherever you see the generic base classes in the architecture documentation.
For malaria simulations, the following corresponding classes are derived from vector classes:
- SimulationMalaria
Creates NodeMalaria objects instead of generic Node objects.
- NodeMalaria
Creates IndividualHumanMalaria objects instead of generic IndividualHuman objects and provides various malaria-specific counters for the purposes of reporting.
- IndividualHumanMalaria
Represents a human being and provides the additional layer of functionality for how malaria vectors interact with individual humans.
- SusceptibilityMalaria
Represents a human being’s susceptibility to malaria. It contains much of the intra-host model, by modeling the specific details of an individual’s immune system in the context of the malaria infection life cycle. It is highly configurable, and interacts closely with InfectionMalaria objects and the IndividualHumanMalaria object, its parent. From a software point of view, SusceptibilityMalaria derives from SusceptibilityVector, but in reality SusceptibilityVector provides minimal epidemiological functionality.
- InfectionMalaria
Represents a human being’s infection with malaria. It is the other part of the detailed intra-host malaria model. It models the progression of the malaria parasite through sporozoite, schizont, hepatocyte, merozoite, and gametocyte stages. InfectionMalaria objects are contained with IndividualMalaria objects. There can be multiple such objects. They all interact closely with the SusceptibilityMalaria object. From a software point of view, InfectionMalaria derives from InfectionVector, but in reality InfectionVector provides minimal epidemiological functionality.
For generic simulations, human-to-human transmission uses a homogeneous contagion pool for each node. Every individual in the node sheds disease into the pool and acquires disease from the pool. For vector-borne diseases, disease transmission is more complex as it must take into account the vector life cycle. See Disease transmission.
The relationship between these classes is captured in the following figure.
Simulation components¶
After the simulation is initialized, all objects in the simulation are updated at each time step, typically a single day. Each object implements a method Update that advances the state of the objects it contains, as follows:
Controller updates Simulation
Simulation updates Nodes
Node updates IndividualHuman
IndividualHuman updates Susceptibility, Infections, and InterventionsContainer
InterventionsContainer updates Interventions
As a stochastic model, EMOD uses a random number seed for all simulations. The Simulation object has a data member (RNG) that is an object maintaining the state of the random number generator for the parent Simulation object. The only generator currently supported is pseudo-DES. The random seed is initialized from the configuration parameter Run_Number and from the process MPI rank. All child objects needing access to the RNG must be provided an appropriate (context) pointer by their owners.
The Simulation class contains the following methods:
Method |
Description |
|---|---|
Populate() |
Initializes the simulation. The Populate method initializes the simulation using both the configuration file and the demographic files. Populate calls through to populateFromDemographics to enable the Simulation object to create one or many Node objects populated with IndividualHumans as dictated by the demographics file, in conjunction with the sampling mode and value dictated by the configuration file. If the configuration file indicates that a migration and a climate model are to be used, those input file are also read. Populate also initializes all Reporters. |
Update() |
Advances the state of nodes. |
Simulation object hierarchy¶
For multi-core parallelization, the demographics file is read in order on each process and identity of each node and is compared with a policy assigning nodes to processes embodied in objects implementing InitialLoadBalancingScheme. If the initial load balancing scheme allows a node for the current rank, the node is created via addNewNodeFromDemographics. After all nodes have been created and propagated, the NodeRankMaps are merged across all processes. See Load-balancing file structure for more information.
Nodes are model abstractions that represent a population of individuals that interact in a way that does not depend on their geographic location. However, they represent a geographic location with latitude and longitude coordinates, climate information, migration links to other nodes, and miscellaneous demographic information. The Node is always the container for IndividualHumans and the contagion pool. The Node provides important capabilities for how IndividualHumans are created and managed. It can also contain a Climate object and Migration links if those features are enabled. The climate and migration settings are initialized based on the information in the input data files.
The Node class contains the following methods:
Method |
Description |
|---|---|
PopulateFromDemographics() |
The entry point that invokes populateNewIndividualsFromDemographics(InitPop), which adds individuals to a simulation and initializes them. PopulateNewIndividualsFromBirth() operates similarly, but can use different distributions for demographics and initial infections. |
Update() |
Advances the state of individuals. |
updateInfectivity() |
The workhorse of the simulation, it processes the list of all individuals attached to the Node object and updates the force of infection data members in the contagion pool object. It calls a base class function updatePopulationStatistics, which processes all individuals, sets the counters for prevalence reporting, and calls IndividualHuman::GetInfectiousness for all IndividualHuman objects. The code in GetInfectiousness governs the interaction of the IndividualHuman with the contagion pool object. The rest of the code in updateInfectivity processes the contagion beyond individual contributions. This can include decay of persisting contagion, vector population dynamics, seasonality, etc. This is also where the population-summed infectivity must be scaled by the population in the case of density-independent transmission. |
updateVitalDynamics() |
Manages community level vital dynamics, primarily births, since deaths occur at the individual level. |
By default, an IndividualHuman object is created, tracked, and updated for every person within a node. To reduce memory usage and processing time, you may want to sample such that each IndividualHuman object represents multiple people. There are several different sampling strategies implemented, with different strategies better suited for different simulations. See Sampling for more information.
If migration is enabled, at the end of the Node update, the Node moves all migrating individuals to a separate migration queue for processing. Once the full simulation time step is completed, all migrating individuals are moved from the migration queue and added to their destination nodes.
The IndividualHuman class corresponds to human beings within the simulation. Individuals are always contained by a Node object. Each IndividualHuman object may represent one or more human beings, depending on the sampling strategy and value chosen.
The IndividualHuman class contains components for susceptibility, infection, and interventions. Infection and Susceptibility cooperate to represent the detailed dynamics of infection and immune mechanisms. Every IndividualHuman contains a Susceptibility object that represents the state of the immune system over time. Only an infected IndividualHuman contains an Infection object, and may contain multiple Infection objects.. Susceptibility is passed to initialize the infection immunology in Infection::InitInfectionImmunology(). The state of an individual’s susceptibility and infection are updated with Update() methods. Disease-specific models have additional derived classes with properties and methods to represent specifics of the disease biology.
The InterventionsContainer is the mediating structure for how interventions interrupt disease transmission or progression. Campaign distribution results in an Intervention object being added to an individual’s InterventionsContainer, where it remains unless and until it is removed. When an IndividualHuman calls Update(), the InterventionsContainer is updated and its effects are applied to the IndividualHuman. These effects are used in the individual, infection, and susceptibility update operations. If migration is enabled, at the end of each individual’s update step EMOD checks if the individual is scheduled for migration (IndividualHuman::CheckForMigration()), setting a flag accordingly.
The IndividualHuman class contains the following methods:
Method |
Description |
|---|---|
Update() |
Advances the state of both the infection and the immune system and then registers any necessary changes in an individual’s state resulting from those dynamics (that is, death, paralysis, or clearance). It also updates intrinsic vital dynamics, intervention effects, migration, and exposure to infectivity of the appropriate social network. |
ExposeToInfectivity() |
Passes the IndividualHuman to the ExposeIndividual() function if it is exposed to infectivity at a time step. |
UpdateInfectiousness() |
Advances the quantity of contagion deposited to the contagion pool by an IndividualHuman at each time step of their infectious period. This is explained in more detail below. |
Transmission of disease is mediated through a pool mechanism which tracks abstract quantities of contagion. The pool mediates individual acquisition and transmission of infections as well as external processes that modify the infectivity dynamics external to individuals. The pool provides basic mechanisms for depositing, decaying, and querying quantities of contagion which are associated with a specific StrainIdentity. The pool internally manages a separate ContagionPopulation for each possible antigen identity. ContagionPopulations have further structure and manage an array of contagion quantities for each substrain identity.
Each IndividualHuman has a sampling weight \(W_i\) and a total infectiousness \(X_i\), the rate at which contacts with the infectious individual become infected. This rate can be modified by transmission-reducing immunity or heterogeneous contact rates, which are gathered in \(Y_{i,transmit}\), and the transmission-reducing effects of interventions, such as transmission- blocking vaccines in the factor \(Z_{i,transmit}\). The sampling weight \(W_i\) is not included in the probability of acquiring a new infection. Sample particles are simulated as single individuals, their weighting \(W_i\) is used to determine their effects upon the rest of the simulation.The total infectiousness \(T\) of the local human population is then calculated as:
For simulation of population density-independent transmission, individual infectiousness \(X_i\) includes the average contact rate of the infectious individual, so this total infectiousness is divided by the population \(P\) to get the force of infection \(FI=\frac{T}{P}\) for each individual in the population. The base rate of acquisition of new infections per person is then \(FI\), which can be modified for each individual \(I\) by their characteristics \(Y_{i,acquire}\) and interventions \(Z_{i,acquire}\). Over a time step \(\Delta t\), the probability of an individual acquiring a new infection is then:
A new infection receives an incubation period and infectious period from the input configuration (a constant at present, but possibly from specified distributions in the future) and a counter tracks the full latency, which is possible when simulating individual particles. After the incubation period, the counter for the infectious period begins, during which time the infection contributes to the individual’s infectiousness \(X_i\).
For vector-borne diseases, during an Update() for a node, the infectiousness of vectors is calculated on the local human population regarding the vector life cycle, along with the effects their interventions, such as bednets. Weather data from the Climate object is then used to update the available larval habitat for each local vector species. Multiple local vector species are supported, and after the weather updates, each vector species is advanced through a time step with the VectorPopulation::TimeStep() method. This calculates the life cycle and vector infection dynamics, along with the feeding cycle. It updates the indoor and outdoor bites, and indoor and outdoor infectious bites on the local human population. Each Individual in the local human population is then advanced in an Update(), propagating its infections forward in time, updating its interventions and infectivity, and acquiring any new infections.
NodeVector::updateInfectivity() calls the VectorPopulation::TimeStep() method for mortality adjustments due to weather and the human population and, when completed, processes each list. The effects of the human population are accounted for in VectorPopulation::Update_Host_Effects(), which determines the outcomes for indoor and outdoor attempted feeds on each individual human, and then on the human population as a whole, weighting by any heterogeneous biting. The processing order is:
All infectious female mosquitoes
All infected female mosquitoes
All adult uninfected female mosquitoes
All adult male mosquitoes
Immature mosquitoes
All mosquito larvae
All mosquito eggs
After each list is updated, all indoor and outdoor bites, both infectious and non-infectious, are tallied and updated within the Node infectivity objects.
Within a list update, each cohort of identical-state mosquitoes or each individual-agent mosquito experiences risk of mortality, may progress through any stage of development, such as sporogony (for infected) or larval maturation (for larval), and may attempt a blood feed (for adult females). For mosquitoes attempting a blood feed, outcomes are governed by the calculated host effects, with sequential conditional draws for choice of host type, choice of indoor or outdoor feeding location, and feed outcome. The potential feed outcomes are die before feeding, no host found, die during feeding, die post-feeding, and successful feed. Successful feeds result in ovipositions, and eggs laid start the next generation of mosquitoes.
Campaign management¶
Campaigns are structured into campaign events that determine when, where, and to whom the intervention will be distributed and interventions that determine what is distributed, for example vaccines or other treatments. This section describes the software architecture used for managing campaigns.
Campaigns can be very simple, such as a single intervention with fixed parameters at a point in time, or a complex, adaptive, repetitive, and responsive combination of interventions. A Campaign is a collection of events that modify a simulation. An Event is a distribution within a campaign, and an Intervention is an item that is given out, such as a vaccine, a drug or a bednet.
Warning
If you modify the source code to add or remove configuration or campaign parameters, you may need to update the code used to produce the schema. You must also verify that your simulations are still scientifically valid.
Contents
A campaign event triggers the campaign management system to evaluate when, where, and to whom the intervention will be distributed. This section describes the SimulationEventContext, CampaignEvent, NodeSet, and EventCoordinator classes used to manage this process.
For more information on setting up campaign files, see Create a campaign file.
The Simulation class has a nested helper class called SimulationEventContextHost, referred to as SimulationEventContext. It serves as the interface to, or manager of, CampaignEvents for the Simulation class.
Incoming: The Simulation to SimulationEventContext interface contains the following methods:
Method |
Description |
|---|---|
LoadCampaignFromFile() |
Loads the campaign file during simulation initialization (t=0). |
RegisterEventCoordinator() |
Registers new event coordinators. Also used by CampaignEvent during activation to notify the SimulationEventContext of a new active EventCoordinator. |
Update() |
Advances the state at each time step. |
VisitNodes() |
Called from CampaignEvent during its activation. CampaignEvent is not a Node container, so it needs to communicate with the Simulation class to do anything with nodes. |
Outgoing: The SimulationEventContext communicates with CampaignEvents and EventCoordinators, which are described below.
The CampaignEvent class exists primarily as an initialization-time, minimal intelligence container with a 1-to-1 mapping for CampaignEvents found in the JSON campaign file. They do not do much in the way of run-time campaign distribution or control.
Incoming: The public interface for CampaignEvent contains the following methods:
Method |
Description |
|---|---|
CreateInstance() |
For each campaign event that the SimulationEventContext finds in the campaign file, it creates a corresponding CampaignEvent instance via CampaignEventFactory. The nested JSON elements in a campaign event are stored as an opaque (that is, unparsed) JSON object known as Event_Coordinator_Config. Immediate parameters like Start_Day are parsed out of the JSON and stored in explicit variables in the CampaignEvent class. |
Dispatch() |
In its steady-state, the SimulationEventContext checks the start day for all existing CampaignEvents and invokes this method to when the time step matches the start day. |
Outgoing: The CampaignEvent communicates with two other classes: NodeSet and EventCoordinator.
The NodeSet class parses the Nodeset_Config and helps the EventCoordinator determine if a given node is included in a particular intervention distribution. It can be as simple as NodeSetAll, which includes all nodes, or something much more involved. Usual cases include NodeSetPolygon and NodeSetCommunityList.
Incoming: The public interface for NodeSet contains the following methods:
Method |
Description |
|---|---|
Contains() |
The caller passes a Node via the INodeEventContext interface and gets back a Boolean. The NodeSet class completely encapsulates whatever logic is used to evaluate node membership. |
EventCoordinator is a base type that contains most of the functionality in the event distribution mechanism. There are two actual EventCoordinator classes implemented, but you can new implementations easily by creating a new EventCoordinator class.
Incoming: There are four classes or class groups that make calls into EventCoordinator:
CampaignEvents, which create and activate them.
The SimulationEventContext that invokes their steady-state methods.
Nodes (via the NodeEventContext helper class)
Interventions, which get EventCoordinator-level configuration data needed for intervention execution.
The public interface for EventCoordinator contains the following methods:
Method |
Description |
|---|---|
AddNode() |
Invoked by CampaignEvent::Dispatch() to allow the EventCoordinator to act as a node container and have access to all its constituent nodes via the INodeEventContext interface. Note The AddNode calls actually come via a delegate function that is in the EventCoordinator, but is invoked from the VisitNodes method up in the SimulationEventContext. The CampaignEvent passes the delegate function it wants used when it calls SimulationEventContext::VisitNodes() . This circuitous journey is needed because EMOD operates on a node-by-node basis and the SimulationEventContext, not the CampaignEvent, is the primary node container. |
CreateInstance() |
During their instantiation, CampaignEvents create EventCoordinators as the CampaignEvents parse the Event_Coordinator_Config, which specifies the class name and is passed by SimulationEventContext. Nothing then happens until CampaignEvent::Dispatch() is called. |
IsFinished() |
When an EventCoordinator determines that it is finished distributing its interventions, it sets an internal flag. The SimulationEventContext queries the IsFinished() function on each registered EventCoordinator and disposes of it if true. An EventCoordinator can activate and complete in a single time step, or it could last the entire length of the simulation. |
Update() |
The SimulationEventContext calls this on all registered EventCoordinators to advance their state at each time step. This method exists for global EventCoordinator communication and logic. |
UpdateNodes() |
Distributes the actual intervention. For the simplest possible use case, an EventCoordinator will distribute an intervention on its start day and then finish in that same time step. However, an EventCoordinator may implement repeating interventions, phased distributions, or sustained distributions that are contingent on node or individual attributes. |
Outgoing: The outgoing function calls consist of the distribution of the intervention either to individuals or nodes. For the interface between the EventCoordinator and the actual Intervention, the EventCoordinator calls InterventionFactory and passes it the remaining unparsed JSON from the campaign file. Namely, this is the Intervention_Config, the actual intervention-level campaign parameters including the intervention class. The InterventionFactory returns the newly created intervention instance for each individual via a one of two types of IDistributableIntervention interfaces.
Interventions are the actual objects aimed at reducing the spread of disease, such as a vaccine or drug. THey can be either distributed to individuals or to nodes. For example, a vaccine is an individual-distributed intervention and a larvicide is a node-distributed intervention. Intervention is an abstract type class and specific interventions are classes that implement either IDistributableIntervention, for individual-distributed interventions, or INodeDistributableIntervention, for node-distributed interventions.
The determination of whether an intervention is individual-distributed or node-distributed is contained by logic in InterventionFactory. At the beginning of the EventCoordinator::UpdateNodes()**method, the **InterventionFactory is queried with the unparsed JSON campaign file to see if it returns an INodeDistributableInterface pointer. If it does, the intervention is executed as a node-distributed intervention; if it does not, the individual- distributed intervention execution is used.
Incoming: The interface to the abstract type Intervention class is the Distribute() method. It functions slightly differently for each intervention type.
Outgoing: Intervention is an abstract type that interacts with Nodes or IndividualHumans with an interface specific to that intervention type. The intervention calls QueryInterface on the container to ask for a consumer interface, for example IBednetConsumer. If supported, a Give method specific to that intervention will be called, for example, GiveBednet(this) to give itself to this individual. It is then up to the internal IndividualHuman or Node logic to decide what to do with this intervention.
An individual-distributed intervention implements IDistributableIntervention, a container for the actual parameter name-value pairs from the campaign file, that is contained by an IndividualContainer, an IndividualHuman helper class. The EventCoordinator calls IDistributableIntervention::Distribute(), passing the pointer IIndividualHumanDistributionContext, which provides an ISupports interface to the individual by which the intervention can request a consumer interface. This interface can then be used with Give().
In the EventCoordinator::UpdateNodes() method, the EventCoordinator organizes the distribution of Interventions to IndividualHumans mediated by Nodes, which contain the IndividualHuman objects. The EventCoordinator also applies campaign parameters at the event coordinator level (for example, Target_Demographic, Demographic_Coverage, Num_Repetitions, and Days_Between_Reps) which mainly consists of filtering the distribution to individuals based on individual attributes.
A node-distributed intervention implements INodeDistributableIntervention, a container for the actual parameter name-value pairs from the campaign file, that is contained by a Node. The EventCoordinator calls INodeDistributableIntervention::Distribute(), passing the pointer INodeEventContext, which provides an ISupports interface to the individual by which the intervention can request a consumer interface. This interface can then be used with Give(). Similar to individual-distributed interventions, the EventCoordinator also applies campaign parameters for filtering the distribution.
The architectural diagram below illustrates how campaign file settings are processed by an EMOD simulation.
EMOD campaign architecture¶
Reporter component¶
After running simulations, EMOD creates output reports that contain the model results. Two methods of coordinated reporting are implemented: simulation-wide aggregated reporting and spatial reporting. See Output report structure for more information about the different output reports available.
Simulation-wide aggregate reporting is the most commonly used. This reporting method generates the output written to InsetChart.json. Conceptually, for each time step, the value of a named channel is derived from values provided for that channel by each node. The values are accumulated (summed) over all nodes and then transformed (often normalized by an internal parameter or another channel) just prior to writing. This is implemented through the simulation’s insetChartDataAccumulator.
Basic usage of the NodeTimestepDataAccumulator is explained in Report.h. The simulation is responsible for calling Begin/EndTimestep() and collecting and writing out the data at the end of the simulation. The accumulation calls occur in Node::accumulateInsetChartData().
The second method is spatial reporting which is facilitated by SpatialReporter. In spatial reporting, each node again produces values for named channels, but no simulation-wide accumulation takes place. Instead, the values of each channel for each node are written to a binary table (ReportNNNN.dat) where NNNN is the time step. The format of this file is simple and can be summarized as the maximally packed layout of this structure:
struct ReportDataFormat
{
int32_t num_nodes;
int32_t num_channels;
float data[num_nodes*num_channels];
}
Num_Channels is the number of user channels implicitly generated by the number of different channel names requested in calls to Node::reportSpatialDataReportChannels() plus two additional channels for longitude and latitude. JSON metadata for the ReportNNNN.dat files is written to Report.header.dat after the first time step. It does not include entries for latitude and longitude.
In general, the value of channel \(c\) for node \(n\), starting at zero, is found at \(data[n\times num\_channels+c]\). The latitude is \(data[n\times num\_channels+0]\) and the longitude is \(data[n\times num\_channels+1]\).
The report data files are written after every time step at the request of the Controller by calling WriteTimestep(). Under MPI, the default implementation reduces all the data to rank 0 and writes the combined data out to file on rank 0.
See Custom reporters for information about how to use a custom reporter.
Debugging and testing¶
When you build Eradication.exe or Eradication binary from the EMOD source code, it’s important to debug the code and run regression tests to be sure your changes didn’t inadvertently change EMOD functionality. You can run simulations directly from Visual Studio to step through the code, troubleshoot any build errors you encounter, and run the standard IDM regression tests or create a new regression test.
Warning
If you modify the source code to add or remove configuration or campaign parameters, you may need to update the code used to produce the schema. You must also verify that your simulations are still scientifically valid.
Set log levels¶
EMOD simulations output several files, including log files. See Error and logging files for more information on the logging files that are created by a simulation. By default, logging is at the “INFO” level, but you can set the level at which you want logging information generated. If you see redundant log messages, you can also throttle logging to eliminate them.
There are five log levels. The level chosen will log messages for that level and all higher levels (levels of smaller numeric value). You can set default log levels across the entire Eradication.exe or for individual files in the Eradication.exe; individual file settings will override the default. For example, if you want to increase logging for a particular file you are debugging without increasing logging across the entire Eradication.exe, you can set the default log level to “WARNING” and set the log level for the file to “DEBUG”. The standard output will reflect the default log level for all settings except “ERROR” or “WARNING”. For example:
Log-levels:
Default -> INFO
Eradication -> INFO
The following log levels are available:
- ERROR (1)
This is the highest level. Only errors, including critical errors, are logged to standard output.
- WARNING (2)
Warnings and errors are logged to standard output.
- INFO (3)
This is the default logging level for the executable and is set implicitly. Informational messages, warnings, and errors are logged to standard output.
- DEBUG (4)
Debug information, informational messages, warnings, and errors are logged to standard output. Debug level logging messages do not require a debug build of Eradication.exe. This setting will generate a large number of messages and may impact performance. If you need the debugging messages, we recommend that you limit the number of files with this setting and not set it as the default across the entire Eradication.exe.
- VALID (5)
This is the lowest level. Validation information, debug information, informational messages, warnings, and errors are logged to standard output. This log level is for very low-level debugging of specific values, including values in tight loops. Because it could severely impact performance, this level will only output to standard output during a debug build of Eradication.exe.
Set default Eradication.exe level¶
To set the default log level across the entire Eradication.exe, add the parameter logLevel_default to your configuration file and set it to the log level desired. For example:
{
"logLevel_default": "WARNING"
}
Set level for individual files¶
You can set the log level for an individual file, if it supports the ability. This is typically only used for debugging after making source code changes. This setting will override the default set using logLevel_default, whether set to a higher or lower level.
In the C++ files that make up the EMOD source code, open the file for which you want to set the log level.
Search for
static const char * _module.“Module” refers to the C++ file and is not the same as an EMODule. If this string isn’t present, you cannot set a log level for the file.
Look for logging macros in the file, such as LOG_WARN(x).
If macros are present, this indicates that the file supports that log level setting. The file may not support all log levels. Logging macros can be found in utils/Log.h. You can also add your own logging messages to the file using the logging macros.
In the configuration file, add the logLevel_<FileName> parameter with the log level desired. For example:
{ "logLevel_SimulationVectorExport":"DEBUG" }
Throttle redundant logging¶
It’s possible that multiple identical logging messages are generated by a file of Eradication.exe. This is normal behavior, but the verbosity can be unwanted or unnecessary. You can eliminate duplicate messages by turning on “throttling”, which retains only one logging message instead of a series of duplicate messages from each file.
Only one copy of the message will be generated when it is one of a series of duplicates from the same file. The identical messages need not be output one right after one another to be throttled. If an identical message is output from a different file, it will still be generated. Throttling can only be enabled across the entire Eradication.exe, not per file. It is off by default.
To enable throttling, add the following parameter and value to your configuration file:
{
"Enable_Log_Throttling": 1
}
For example, the following log messages are seen with throttling turned off:
00:00:00 [0] [D] [FileA] identical message: I'm in FileA
00:00:00 [0] [D] [FileB] another message from B
00:00:00 [0] [D] [FileA] identical message: I'm in FileA
00:00:00 [0] [D] [FileB] different message from B
00:00:00 [0] [D] [FileA] identical message: I'm in FileA
00:00:00 [0] [D] [FileB] yet another message from B.
00:00:00 [0] [D] [FileA] identical message: I'm in FileA
With throttling on, the repeated messages from file A are removed, even though they are intermixed with other log messages from file B:
00:00:00 [0] [D] [FileA] identical message: I'm in FileA
00:00:00 [0] [D] [FileB] another message from B
00:00:00 [0] [D] [FileB] different message from B
00:00:00 [0] [D] [FileB] yet another message from B.
Run debug simulations in Visual Studio¶
If you have modified the EMOD source code, you can run simulations directly from Visual Studio as part of debugging your changes by using the built-in debugger. For example, you may want to run one or more of the simulations in the Regression directory to verify that the results match those obtained prior to your changes.
Open the EradicationKernel solution in Visual Studio.
On the Solution Explorer pane, right-click the Eradication project and select Properties.
On the Property Pages window, in the Configuration Properties pane, click Debugging.
Set the Command Arguments and Working Directory to the appropriate values and click OK. See Run a simulation using the command line for more information.
On the Debug menu, do one of the following:
Click Start Without Debugging to run the simulation in Release mode, where the simulation runs with essentially the same behavior and performance as running it at the command line.
Click Start Debugging to run the simulation in Debug mode, where you have the ability to add break points and step through the code, inspecting the values of different variables throughout the simulation.
Troubleshooting EMOD builds¶
If you encounter any of the following warnings or errors when attempting to build the EMOD executable (Eradication.exe) or Eradication binary, see the information below to resolve the issue.
If you need assistance, you can contact support for help with solving issues. You can contact Institute for Disease Modeling (IDM) support at support@idmod.org. When submitting the issue, please include any error information.
Unknown compiler version¶
If you encounter the warning “Unknown compiler version - please run the configure tests and report the results” when attempting to build the Eradication.exe or Eradication binary, this indicates you are using a version of Boost that is no longer supported. Install Boost 1.61.0.
Inconsistent DLL linkage¶
If you see the following warning on some files, “c:python27includepymath.h(22): warning C4273: ‘round’: inconsistent dll linkage”, this indicates that you are using a version of Python that is no longer supported. Install Python 2.7.11 or 2.7.12.
Error 255¶
Check to see if you have any white spaces in the path to your local EMOD source code. If you do, remove the white spaces in all of the directory names in the path.
Error LNK4272¶
If you see the error “LNK4272 library machine type ‘X86’ conflicts with target machine type ‘x64’”, this indicates that you need to uninstall 32-bit Python and reinstall 64-bit Python. Install Python 2.7.11 or 2.7.12.
Regression testing¶
After building the EMOD executable (Eradication.exe), it’s important to verify that Eradication.exe is performing properly. Regression testing is a method by which the built code is tested to see if it has “regressed” or moved backwards in any way, such as previously reported (and fixed) issues reappearing.
Within the EMOD Regression directory there are many subdirectories that correspond to different disease scenarios in a variety of locations. Each of these contains the configuration and campaign files needed to run the simulation and the reference output, which represents the expected results. These reference outputs have been calculated by the scientific researchers modeling each scenario.
EMOD regression testing consists of running one or more of these simulations and comparing the output to the reference output. If the two outputs are comparable, the test passes; if they differ, the test fails. Because EMOD is stochastic, a passing test will fall within some acceptable range of the reference output, rather than be an identical match. If a regression test fails, EMOD will produce a matplotlib chart of the first 16 channels in the InsetChart.json output report. You can then review the charts to identify the problem.
If you want to quickly compare a simulation output to the reference output, you can also run any of the regression scenarios as a typical simulation, as described in Run simulations. However, this will not include the comparison and pass/fail evaluation that regression_test.py conducts. In addition, if you choose to do this, be sure to specify a different output directory, such as “testing”, so as not to overwrite the reference output.
Run regression tests¶
The Python script regression_test.py runs a suite of regression simulations and compares the output to the reference output for each simulation. It is set up to run the simulations on an HPC cluster; however, you can run modify the script to run tests locally. However, the script was written with remote execution in mind and running it locally can be time-consuming. Running the entire regression suite locally will take several hours on average.
The regression scenarios, script, configuration file, and other relevant files are all in the Regression directory. Be aware that many of these tests, due to abnormally high or low values, will produce output that should not be considered scientifically accurate.
Modify the configuration file, regression_test.cfg, for your environment, setting the values for the location of the working directory, input data files, binary file, and cluster settings.
For local Windows simulations, set the values under [WINDOWS].
For local CentOS on Azure simulations, set the values under [POSIX].
For simulations on IDM HPC clusters, no changes are necessary if your username and password are cached locally.
For simulations on your own HPC cluster, create [HPC-<cluster>] and [ENVIRONMENT-<cluster>] sections for your cluster that contain the same variables as shown for IDM HPC clusters.
Select the suite of regression tests you want to run. This is indicated by a JSON file in the following format:
{ "tests": [{ "path": "Relative path to test directory." }, { "path": "Relative path to test directory." }] }
You can use one of the JSON files in the Regression directory or create your own. The sanity.json file is recommended for quickly testing a wide range of EMOD functionality.
From the Regression directory, open a Command Prompt window and run the regression test script, regression_test.py. It requires the name of the regression suite (without the .json extension) and the relative path to Eradication.exe. For example:
regression_test.py sanity ..\Eradication\x64\Release\Eradication.exe
In addition, you may need to include the following optional arguments depending on your testing environment or how Eradication.exe was built.
Argument
Default
Description
--perfFalse
Measure Eradication.exe performance.
--hidegraphsFalse
Suppress pop-up graphs in case of validation failures.
--debugFalse
Use the debug path for EMODules.
--labelAdd a custom suffix for HPC job name.
--configregression_test.cfg
The regression test configuration file.
--disable-schema-testTrue
Include to suppress schema testing, which is on by default.
--use-dllsFalse
Use EMODules when running tests.
--all-outputsFalse
Use all output JSON files for validation, not just InsetChart.json.
--dll-pathThe path to the root directory of the EMODules to use.
--skip-emodule-checkFalse
Skip checking if EMODules on the cluster are up-to-date, which can be slow.
--sconsFalse
Indicate that this is a SCons build so custom DLLs are found in the build/64/Release directory.
--localFalse
Run all simulations locally.
Review the output and examine any failures.
EMOD will output the standard error and logging files, StdErr.txt and StdOut.txt, produced from any simulation (see Error and logging files). In addition, regression_test.py will output time.txt under the regression test working directory and report_xxxx.xml under Regression/reports. The time report contains the EMOD version and total run time of the simulation. The regression report is in JUnit standard form and captures run information, including pass/fail/complete and time to complete.
If a simulation completes saying the run passed but the channel order was different than the reference output, this is considered a pass. However, if any output completes but does not match the reference output, this is considered a failure and a matplotlib chart of the output will appear in a pop-up window. The chart will appear immediately after the simulation, before the entire suite of regression tests completes. You can manipulate the output of the charts, such as adjusting the scale of the plots, zooming or panning, and so forth, through the icons at the bottom of the chart window.
If any of the regression tests fail and you have not made any changes to the EMOD source code, email support@idmod.org.
Create a new regression test¶
You can create a new regression based off one of the ones included with the EMOD source code using the steps below.
Under the Regression directory, create a new subdirectory.
Copy the contents of the regression test that you want to base the new test on into the subdirectory.
Modify the configuration, campaign, and demographic files as desired.
Create the reference output by doing one of the following:
Modify the InsetChart.json file to match the output you expect.
Run simulations until you have an acceptable InsetChart.json that you wish to use as the reference.