- Replace per-ticker yf.Ticker.info with EDGAR XBRL frames API (~50 bulk calls)
- Add nightly cache_builder.py: fetches analyst/news/sector data for all US stocks
- Add .github/workflows/cache_refresh.yml: runs cache_builder Mon-Fri at 2am EST
- Fix sector/industry: cache_builder stores sector from yfinance; screener reads it
- Fix PyInstaller: add collect_all('tzdata'), set upx=False to prevent archive corruption
- Fix EDGAR row format: use dict keys (cik/val) instead of array indices
- Fix yfinance MultiIndex: squeeze single-ticker DataFrame to Series
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Nolan Kovacs