Inspectable data processing workflows. This repository is an independently inspectable project; customer adoption, production scale and commercial readiness are not claimed without evidence.
CSV input → ingestion/transformation/validation/evaluation graph → stored run state → dashboard and metrics.
Primary implementation and review locations: app/main.py, app, tests. Dependency manifests and .github/workflows/ specify installation and automated checks. Read the source for exact contracts and data models.
From . in a configured virtual environment:
pip install -e ".[dev]"
pytest -qFrom the repository root, run python scripts/repository_check.py for documentation and tracked-file checks. CI evidence is available in GitHub Actions. Green hygiene checks alone do not mean application tests passed.
Execution is in-process rather than a durable distributed queue. In-memory state resets on restart. Optional S3/DynamoDB paths need configured resources.
Use the README's current run instructions and configuration examples. Keep provider credentials outside Git. Test changes against controlled fixtures before enabling external services. Health checks indicate process/service state, not end-to-end correctness.
Review checklist distinguishes repository evidence from outstanding human and deployment validation. Report measured workload, environment and method with any performance claim. Document incident fixes through reproducible issues and regression tests; do not invent user counts or peer reviews.
No repository-wide reuse license has been selected. Public visibility alone does not grant an open-source reuse license. Ownership and third-party asset rights must be confirmed before licensing this project.