5 Commits

Author	SHA1	Message	Date
shawn	c8ffab2c06	feat: connect-id suffix matching for OCR + bump SW cache to v10	2026-06-02 18:32:41 -04:00
shawn	8022c77b70	feat: add normalized matching for label photos → asset DB lookup ... - normalize_identifier() strips dots/dashes/prefixes, keeps alphanumeric - find_asset_by_normalized_id() searches serial_number, connect_id, equipment_id, barcode with normalized comparison - /api/ocr now returns matched_assets in addition to legacy machine_id - New /api/match-text endpoint for client-side text matching - scripts/match_label_photo.py CLI tool for OCR + DB matching - Vision model fixed (mimo-v2-omni at opencode.ai, was using truncated placeholder key)	2026-05-29 08:37:06 -04:00
shawn	a305766d65	fix: show '— No report yet' for assets without dex_report_date	2026-05-23 18:46:58 -04:00
shawn	c64baf2c45	classify_makes: add short 8-digit DN + 76/77 prefix rules ... Covers 3 additional edge cases found in unmatched: - 8-digit 114xxxx / 241xxxx → DN (short serial format) - 10-digit 76xxxx / 77xxxx → DN (no alpha suffix variant) Unknown make: 130 → 6 (95.4% classified) 6 remaining: garbage serials (00, 520, a×3) + RY01023590 (low-conf Royal)	2026-05-22 19:10:42 -04:00
shawn	4f33afb445	classify_makes.py: serial number → make/model/class classification ... Classifies Unknown-make assets by serial number prefix patterns, cross-referenced against Cantaloupe Excel export data. Handles: - Character substitution errors (S→5, I→1, O→0) - Double-strike/extra character cleanup (I5SS → 155) - 8 rule sets: Vendo, Crane (167/168/186/187/222/221/47x), DN/Dixie Narco, Royal (20xx/BA/CA/PA), USI, AMS, VE - Confidence levels: high/medium/low - Dry-run + --apply modes - --stats for post-classification summary Result: 121/130 unknown machines classified (93%) 76 Vendo, 14 DN, 12 Crane, 8 Royal, 6 USI, 1 AMS 9 remaining: garbage serials (a, 00, 520) + one-offs	2026-05-22 19:02:04 -04:00