Report the freshness of every MOSAIC data input and output

Read-only audit across all three layers of the data pipeline: the external scraper repos, the hand-maintained raw inputs, and the processed artifacts the pipeline produces. Answers "what is stale?" in one call, and touches nothing.

Usage

check_mosaic_data_freshness(
  root = NULL,
  stale_days = 30L,
  repo_stale_days = 14L,
  verbose = TRUE
)

Arguments

root: MOSAIC parent directory. Defaults to get_paths()$ROOT.
stale_days: Age in days beyond which a processed artifact is flagged. Default 30.
repo_stale_days: Age beyond which a source repo is flagged. Default 14, matching refresh_data_repos.
verbose: Print the report.

Value

Invisibly, a list with elements repos, manual, outputs and summary, each a data.frame.

Details

The third layer is the one nothing else covers. refresh_data_repos() reports source-repo staleness (and pulls as a side effect); check_mosaic_manual_inputs covers the manual raw inputs. Neither looks at what the pipeline produced, which is where vintage drift actually accumulates – a source can be current while the artifact derived from it is two years old.

Read-only

Unlike refresh_data_repos, this performs no git pull – it reads git log only. Safe to run on a schedule or before deciding whether a refresh is warranted.

What "vintage spread" means

The report counts how many distinct year-months the files in a directory were written in. A healthy directory was produced by one build and shows 1-2. A large spread means the directory is sediment from many partial runs and no single coherent build produced it – the files are individually valid but mutually inconsistent in what data they saw.

Examples

if (FALSE)  check_mosaic_data_freshness("~/MOSAIC")  # \dontrun{}