This repo is the first tentative steps towards a reproducible-build longevity data pipeline for The Longevity Initiative.
If you want to help out, please get in touch!
-
Architecture: current pipeline and proposed design.
-
Implementation plan: phased checklist, starting with the headline life-expectancy graph.
-
Local development: environment setup, pinned dependencies, and verification.
Local builds need no storage credentials or .env file. Follow the
local development guide to install dependencies and run
the pipeline. Local data, metadata, and chart previews are written
to the gitignored output/ directory.
For cloud uploads, copy env.example to .env in the repository root and fill
in the four B2_* settings. Keep credentials in .env, which is gitignored.
Build stages never upload. Use python src/pipeline/publish.py --manifest <path>
to publish explicitly; see the development guide for dry runs and recovery.
src/pipeline/: Python ingestion, standardisation, chart builders, and templates.src/charts/: browser JavaScript maintained by this project.tests/: automated regression tests.docs/: architecture, implementation plan, and development instructions.data/: reviewed snapshot provenance, required in Git before publication.output/: fully ignored local data and chart builds, including vendor libraries.
All generated files live in output/. Edit source in src/. Use
record_metadata.py --manifest <path> to prepare provenance for review and
commit; publication verifies it matches Git before accessing the bucket.