Scan picker: ambiguous scans miss real printings (reverse holo, base ex) #540

New Issue

2026-08-30T22:11:04Z

gambit-admin commented

2026-08-30 22:11:04 +00:00

Follow-up to #539. When CardVision flags a scan print_ambiguous (ball patterns / stamped cards / look-alikes), the picker is built from CardVision's noisy prints list instead of our authoritative per-card printings, and my #539 sibling enrichment only feeds the confident-accept path.

Concrete failures Kris hit:

Vaporeon 022/131 (Prismatic, sv8pt5-22): reverse holo missing. It shares TCGplayer product id 610377 with the normal holo, so it's not a distinct scan candidate AND the sibling logic drops printings with no own product id. Scan was print_ambiguous=true. (Also CardVision offered a 151 Vaporeon look-alike — service issue #528.)
Latias ex 076/191 (Surging Sparks, sv8-76): only the two stamp printings shown, not the plain ex ($6.09). Same ambiguous-path issue + a label collision: playPokemonStampHolofoil normalizes to 'Holofoil', same as the base.

Fix: (1) feed the ambiguous/confirm scan path the same variant_siblings enrichment the accept path uses; (2) include finishes that share a product id (reverse holo), priced by their own variant_id; (3) distinct labels for stamp/cosmos printings.

Filed from a Claude Code session

Follow-up to #539. When CardVision flags a scan print_ambiguous (ball patterns / stamped cards / look-alikes), the picker is built from CardVision's noisy prints list instead of our authoritative per-card printings, and my #539 sibling enrichment only feeds the confident-accept path. Concrete failures Kris hit: - Vaporeon 022/131 (Prismatic, sv8pt5-22): reverse holo missing. It shares TCGplayer product id 610377 with the normal holo, so it's not a distinct scan candidate AND the sibling logic drops printings with no own product id. Scan was print_ambiguous=true. (Also CardVision offered a 151 Vaporeon look-alike — service issue #528.) - Latias ex 076/191 (Surging Sparks, sv8-76): only the two stamp printings shown, not the plain ex ($6.09). Same ambiguous-path issue + a label collision: playPokemonStampHolofoil normalizes to 'Holofoil', same as the base. Fix: (1) feed the ambiguous/confirm scan path the same variant_siblings enrichment the accept path uses; (2) include finishes that share a product id (reverse holo), priced by their own variant_id; (3) distinct labels for stamp/cosmos printings. --- Filed from a Claude Code session

gambit-admin added the bug ui claimed:kris labels 2026-08-30 22:11:18 +00:00

gambit-admin referenced this issue

2026-08-30 22:26:51 +00:00

Reprice keys on product id, not variant — link-less finishes reprice to base printing #541

gambit-admin commented

2026-08-30 22:44:08 +00:00

Fixed in PR #361 (merged, promoted to master e66f3ba, mobile OTA 1.3.6). Ambiguous scans now expand the lead match with the card's real printings, so Vaporeon 022/131 offers its reverse holo and Latias 076/191 shows the plain ex alongside the stamps (now distinctly labeled). Review surfaced a pre-existing reprice-by-product-id gap for link-less finishes → filed #541. Web/API deployed + healthy.

gambit-admin closed this issue

2026-08-30 22:44:08 +00:00

gambit-admin removed the claimed:kris label 2026-08-30 22:44:08 +00:00

gambit-admin referenced this issue

2026-09-23 00:28:02 +00:00

Review follow-ups: PR #540 — fix(import-parse): grades aren't card numbers; graders aren't assumed PSA; JP rows become matchable; sold dates recognised #784

Sign in to join this conversation.

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: gambit/gambit#540