OCR: poppler + gosseract; optional Paddle; retire docling #6

Open
opened 2026-08-13 14:05:47 +01:00 by eSlider · 0 comments
Owner

Why

Mail convert uses isolated docling (onnx segfault). Go default is Tesseract.

Do

  • Keep pdftotext -layout first. Scans: pdftoppm + gosseract (eng+deu). imaging preprocess, no gocv in v1.
  • Optional ocr-paddle profile (PP-OCRv5 ONNX). Docling leaves the default path.

Test

  • Synthetic printed-text PNG fixtures. gosseract tests tagged / OCR image job.
## Why Mail convert uses isolated docling (onnx segfault). Go default is Tesseract. ## Do - Keep `pdftotext -layout` first. Scans: `pdftoppm` + gosseract (`eng+deu`). imaging preprocess, no gocv in v1. - Optional `ocr-paddle` profile (PP-OCRv5 ONNX). Docling leaves the default path. ## Test - Synthetic printed-text PNG fixtures. gosseract tests tagged / OCR image job.
Sign in to join this conversation.
No labels
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: eSlider/2dph#6