Skip to content

fix: decouple barcode detection from OCR engines

Riccardo Di Mambro requested to merge feature/barcode-engine-decoupling into master

Summary

Decouples barcode detection from the selected OCR engine so that barcode processing behaves consistently with Tesseract and Paddle OCR profiles.

Changes

  • introduces a dedicated barcode service;
  • removes barcode detection dependency from OCR engine selection;
  • updates document processing integration;
  • adds regression tests;
  • adds technical documentation.

Validation

  • targeted barcode suite: 20 passed, 2 known warnings;
  • correlated OCR/hOCR regression suite: 67 passed, 2 known warnings;
  • tracked full suite: 90 passed, 11 skipped, 14 pre-existing quality-endpoint fixture errors;
  • no new failures or collection errors in tracked tests;
  • branch and remote tracking reference are aligned;
  • whitespace check passed;
  • tracked working tree and index are clean.

Scope

This merge request does not contain the subsequent hOCR bounding-box integrity correction.

Merge request reports