Skip to content

A fast, helpful, and open-source document parser

Commits on main

48bf74e
docs(wasm): correct the OCR default and setup instructions
Choudhry Abdullah committed 1 month ago
b1c4b2e
fix: don't treat short CJK OCR words as chart labels
Diego-Romero-NTT committed 1 month ago
409afad
fix: recognize dash bullets as list markers
moah0911 committed 1 month ago
af158e2
fix: keep glyphs whose ink misses a clip by under 2% of the em
Pierre-Loic doulcet committed 1 month ago
f5e3d38
fix: judge clipped glyphs by their ink box, not the loose box
Pierre-Loic doulcet committed 1 month ago
8c7ca55
4cba0f2
37b0cbd
52a767e
924d204
fix: clip table block and cell boxes to the page
Christopher Atlan committed 1 month ago
aba75de
fix: extract word boxes when extract_blocks is on
Christopher Atlan committed 1 month ago
08eb974
Merge pull request #482 from run-llama/fix/clipped-native-text
Pierre-Loic Doulcet committed 1 month ago
b2a8735
fix: preserve partially clipped glyphs in full
Pierre-Loic doulcet committed 1 month ago
289e9c5
fix: exclude glyphs hidden by rectangular PDF clips
Pierre-Loic doulcet committed 1 month ago
aed33eb
v2.15.1
Logan Markewich committed 1 month ago
ab88796
fix: is_garbled checks for broken fonts
Logan Markewich committed 1 month ago
0f500f8
v2.15.0
Logan Markewich committed 1 month ago
d3a7917
fix(ocr): attach word boxes to OCR text items
Yi-111-a committed 1 month ago
26e7921
fix: preserve UTF-8 in sparse line compression
sanjaibalajee committed 1 month ago
0c8d9a8