ragflow
fix(mineru): use cached img_path in crop() to consume generated_images; manual.py tag parsing patched (other parsers unchanged)
#11855
Open
Open
fix(mineru): use cached img_path in crop() to consume generated_images; manual.py tag parsing patched (other parsers unchanged) #11855
shaoqing404
wants to merge
12 commits into
infiniflow:main
from
shaoqing404:fix/mineru-missing-images-submit
Fix: Generate missing images for MinerU text blocks using local crop
b443d34f
fix(mineru): use cached img_path in crop() to consume generated_images
eb004b62
fix(mineru): use consistent 0-1000 normalized coords for line_tag cac…
8049cb92
fix(mineru): robust coordinate conversion in crop() fallback for 0-10…
1c7bc475
Merge branch 'infiniflow:main' into fix/mineru-missing-images-submit
31b466fe
fix: Initialize imgs list in crop() fallback path
3bc3d82a
Merge branch 'infiniflow:main' into fix/mineru-missing-images-submit
b018ab6c
feat(mineru): implement smart crop with page-width fallback and nativ…
8a285d12
Merge branch 'infiniflow:main' into fix/mineru-missing-images-submit
3ce7b02d
fix: MinerU crop tag matching and manual.py bbox parsing
2d475053
shaoqing404
changed the title
yongtenglei
requested changes
on 2025-12-11
chore: increase image stitching thresholds to 20/4000px
02a4b79f
feat: enhance MinerU crop() with 3 major improvements
58792dfe
Login to write a write a comment.
Login via GitHub