Stadium and capacity lookup

stadium_capacity_lookup_template.csv now contains the user-supplied stadium and capacity records for 52 historical club-season-tier combinations. Keep this file as the maintained lookup; despite its original template name, the gold script reads it on every run.

Rows match Team, Season and Tier. The gold script fills missing Stadium and Capacity independently, preserving populated values from the scraped dimension. It applies the lookup before stadium/capacity comparisons used in baseline rules. Repeated observations of the same club-season receive the same lookup values.

Keys must be unique and populated. Seasons use YYYY-YY (en dashes and full end years are normalised). Tier must be an integer from 1 to 4, and capacity a whole number from 500 to 150000. Decimal representations such as 23287.00 are accepted. Incomplete or invalid rows stop the run before gold outputs are overwritten.

Gold CSV, Parquet and the event comparison report expose Stadium_Lookup_Applied, Capacity_Lookup_Applied, Stadium_SourceURL, Capacity_SourceURL and Stadium_Capacity_Lookup_Notes. These flags identify user-supplied fills. Source URLs and notes are preserved as provided; most URLs currently refer to homepages and do not independently establish the capacity for the stated historical season. Use the specific supporting page when updating a record’s evidence.

Run gold_too_messy_to_melt_script.py after editing the lookup. This rebuilds gold CSV/Parquet, the event comparison report and missing_fields_by_team_season.csv. The scraper’s dimension and the populated lookup file are separate inputs.