An interactive EDA session¶

Stub. This notebook is part of the documentation scaffold. The final version will mirror bench/session_vs_oneshot.py — reusing the module-level default connection, running several follow-up queries against the same catalogs, and using Catalog.save(...) to materialize an intermediate crossmatch.

Planned flow:

1. acid.init(catalogs, workers=8) to pin the default connection against the catalogs from Your first crossmatch. (Optional — the first acid.open() / acid.sql.query() lazy-inits one.)
2. Run three follow-up queries against the warm pool. The default connection's worker pool and registry are reused across queries; the pickling and startup costs are paid once.
3. Catalog.save("./out/nearby", name="nearby") on a heavy 3-way XMATCH; then iterate on filters and aggregates against nearby (resolvable by name in subsequent acid.sql / acid.open calls) -- subsecond per query.
4. Use a with acid.in_cone((ra, dec), radius=1*u.deg): block to scope materialization to a small patch while developing, then materialize outside the block for the full-sky run — the same query object, no rebuild.
5. Switch a result to Polars (r.to_polars()) for a quick group-by-magnitude analysis.

See Connections for the lifecycle and the cone rules; this notebook will be the worked example.