List Series

All datasets with their latest value and changes.

A GROUP is returned as ONE entry carrying a facets array — not as N flat rows and not omitted. The first cut of this dropped grouped series entirely, which silently removed rppi_national_houses (the panel's national comparison line) the moment property gained facets. Returning 334 flat rows instead was the other extreme: it made this the slowest endpoint in the app.

include_facets=true additionally returns every facet as its own top-level row.

stat_key: restrict to one stat (e.g. 'crime', 'sale_price'); omit for every series site-wide. See /api/v1/stats for valid keys. Added for the MCP tool package's rank_areas -- an unfiltered call measures ~6.5MB (every series, every geography); ranking one stat across areas needs only a slice of that. Plain default (not Query(...)) deliberately: this function is also called directly as a plain Python function in this codebase's own tests, and a Query(...) object survives unresolved as the literal default there, not as None -- the exact bug already caught and fixed once on series_for_geography's own stat_key filter earlier this session.