Implement the ReceiptParser parsing heuristics #23

New issue

Closed

opened 2026-07-05 19:20:26 +01:00 by TimCane · 0 comments

TimCane commented

2026-07-05 19:20:26 +01:00

(Migrated from github.com)

Goal

Pure Parse(OcrResult) -> ParsedReceipt {Items, Bill, Warnings}: price extraction, keyword rows (tax/tip/service/total with lowest-total-wins), item rows with quantity stripping, discard warnings, currency guess.

References

docs/06-ocr-service.md#parsing

Done when

Advances M3 A1/A2 (plausible items, corpus).

Part of M3: OCR pipeline.

## Goal Pure Parse(OcrResult) -> ParsedReceipt {Items, Bill, Warnings}: price extraction, keyword rows (tax/tip/service/total with lowest-total-wins), item rows with quantity stripping, discard warnings, currency guess. ## References - [docs/06-ocr-service.md#parsing](https://github.com/TimCane/bill-splitter/blob/main/docs/06-ocr-service.md#parsing) ## Done when - Advances M3 A1/A2 (plausible items, corpus). --- Part of **M3: OCR pipeline**.