Tracks how sirius-db/sirius evolves over time — stars, forks, traffic (including referrers and popular content), day-to-day issue/PR/commit activity, open issue/PR label trends (priority and other, with stale-item drill-down), and (eventually) release download counts.
GitHub's own Insights tab and traffic API only retain a rolling window (traffic data is dropped
after 14 days), so anything not collected is gone permanently. This repo collects a daily
snapshot and commits it to main, then publishes a static site from the accumulated history at
www.sirius-db.com/sirius-stats.
.github/workflows/collect.ymlruns once daily at0030 UTC(plus manualworkflow_dispatch) and never collects the current, still-in-progress day -- it only ever finalizes fully-elapsed days. Each firing scans the last 7 days for every currently-missing snapshot (not just the most recent) and collects all of them in one run, so a backlog from a missed/delayed firing actually closes instead of the pipeline staying permanently behind.scripts/refresh_traffic.pyalso runs every firing to keep already-collected snapshots' traffic fields in sync with GitHub's own rolling 14-day breakdown, which can lag or later revise a day's numbers. Both are committed together in a single commit -- seecollect.ymlfor the exact gating logic andBOOTSTRAP.md§5 for manually recovering a day that's aged out of the 7-day catch-up window..github/workflows/deploy.ymlruns on every push tomainand on manualworkflow_dispatch. Runsscripts/build_site.py, publishes the result to GitHub Pages..github/workflows/verify.ymlruns on every PR tomainand on manualworkflow_dispatch. Runs the same two scripts end-to-end without committing or deploying, so a broken change fails CI before merge -- including docs-only PRs, sinceverifyis a required status check and a path-filtered trigger would leave those permanently blocked.
See CONTRIBUTING.md for local dev setup, running the scripts, previewing site
changes, and customizing the UI. See DATA.md for the full schema of everything under
data/, including the recommended site/data.json entry point for historical analysis or
search. See BOOTSTRAP.md for one-time setup steps (backfills, secrets, branch
protection) outside the recurring workflows.