Two related schema fixes from the code review:
- users.lang gets a single-column index. The ai_log_job and
email_digest_job both SELECT DISTINCT on this column every cycle;
even at low cardinality an index is the right shape.
- quotes_daily.symbol widened to VARCHAR(128) to match quotes.symbol
(widened back in 0005). Long Eurostat/ONS symbols would silently
truncate during rollup otherwise.
Models updated to match (User.lang gains index=True, QuoteDaily.symbol
goes to String(128)).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Giorgio Gilestro
eb31d09782