1
Workaround for unreliable auto-rotation with scanners that embed /Rotate metadata
Source: paperless-ngx/paperless-ngx#13461 · opened by @martindamboeck-cpu
Description Following up on the (now locked) discussions #6319 and #11635, which describe the same symptom — posting a workaround here since I can't comment there anymore. A reliable workaround for unreliable auto-rotation (pre-consume script) I ran into this exact issue after several days of testing: pages scanned upside-down (from a multifunction printer/scanner that embeds its own OCR text layer and /Rotate metadata) were consistently NOT being rotated correctly by Paperless-ngx, even with PAPERLESS_OCR_ROTATE_PAGES=true and various threshold settings. Root cause (as far as I could determine) The log line with existing rotation ⇨, page is facing ⇩, confidence X.XX - will rotate ↻ is misleading — it only means OCRmyPDF's OSD *detected* that rotation is needed, not that it was actually applied. In my testing, this happened consistently regardless of: • PAPERLESS_OCR_MODE (tried redo, force, auto) • Lowering PAPERLESS_OCR_ROTATE_PAGES_T…
No pledges yet. Be the first to back this.
Comments
Similar requests
[Feature] Optional removal of blank pages from the archive file (originals untouched)
1 vote · 0 comments
[Feature Request] Archive Export (Original File + Converted File + Unified Metadata)
3 votes · 0 comments
[Feature Request] Support for manipulating image files (.jpg, .png...) - crop and rotate
5 votes · 0 comments
[Feature Request] AI chat/RAG: cache the embedding model instead of reloading it on every query
1 vote · 0 comments
[Feature Request] Allow overriding the embeddings encoding_format for OpenAI-compatible LLM providers
4 votes · 0 comments
No comments yet.