PDF-to-PDF comparison
Compares BOM PDFs against supplier order confirmation PDFs and highlights mismatched quantities.
Windows desktop QA tool · BOM validation · PDF parsing
LAPSRS BOM Validator compares BOM PDFs with supplier order confirmation PDFs, detects quantity discrepancies, and exports human-reviewable Excel reports for procurement and industrial QA workflows.
$ lapsrs validate bom.pdf supplier-confirmation.pdf
✓ OCR repair heuristics applied
✓ Matched supplier lines against BOM entries
! 2 discrepancies require human review
✓ Excel report generated
Built for messy procurement documents
LAPSRS is designed around real PDF failure modes: OCR damage, formatting noise, split lines, order confirmation quirks, and edge cases that still need manual sign-off.
Compares BOM PDFs against supplier order confirmation PDFs and highlights mismatched quantities.
Uses targeted cleanup logic to recover common extraction errors before matching.
Surfaces uncertain matches for human review instead of hiding ambiguous results.
Generates discrepancy workbooks suitable for QA handover, review, and record keeping.
Validated against A–M datasets, including real discrepancies, OCR issues, and invalid/multi-document cases.
PyInstaller EXE and Inno Setup installer builds are already working for Windows distribution.
How it works
Select the BOM PDF and supplier confirmation PDF inside the desktop app.
Tesseract, Poppler, parser rules, and OCR repair heuristics normalise document text.
Line items are compared while known risky cases are escalated rather than over-trusted.
An Excel discrepancy report is generated for procurement QA review.
Screenshots
The placeholders below are intentionally styled like product screenshots. Replace the SVGs in assets/screenshots with real app captures later.
Ready for release wiring
Once the installer is uploaded to a release, replace YOUR-USERNAME/YOUR-REPO in this site with your actual GitHub repository path.
Version history
Draft branch for split-length and equivalent-length matching improvements.
Improved OCR repair, parser confidence, mismatch detection, and regression harness behaviour.
PyInstaller EXE builds and Inno Setup installer workflow confirmed working.
Support
For best results, include the BOM PDF, supplier confirmation PDF, generated discrepancy report, app version, and a short note describing the expected result.