Odin Dahlström
b7d63f3dc1
Process ocr_caption lines
2025-01-30 17:49:06 +01:00
James R. Barlow
d87db6cad0
Merge branch 'pr/pajowu/1446'
2025-01-01 17:50:33 -08:00
James R. Barlow
f71a5ffd61
hocr: comment typo
2024-12-23 01:46:00 -08:00
Kara Engelhardt
74253e5fc8
hocr: only add space if boxwidth is positive
2024-12-11 16:23:45 +01:00
James R. Barlow
0f82d7223e
Fix some typing issues
2024-10-27 12:31:16 -07:00
James R. Barlow
653c4ffb45
hocr: accept multiple spaces in bounding boxes
...
Fixes #1322
2024-05-31 16:22:27 -07:00
James R. Barlow
28be50136c
hocr: If a line box's coords are invalid, log and error and don't render
...
Addresses [Bug]: Crash on multiple .pdf files #1312
Not actually a fix, but at least it will get us better diagnostics. Appears old Tesseract 4.x generates bad line boxes at times.
2024-05-18 23:32:18 -07:00
James R. Barlow
6a746a1cbb
ruff linting/Python 3.10 cleanup
2024-02-14 12:41:51 -08:00
James R. Barlow
602f5632cb
hocr: fix unbound local
2024-01-07 01:33:36 -08:00
James R. Barlow
9498f01f59
hocr: Pass text direction for standalone words
2024-01-06 16:21:21 -08:00
James R. Barlow
7e608fd1df
Use Courier as debug font
2024-01-06 14:24:25 -08:00
James R. Barlow
b71d16dd96
Draw RTL text with reversed matrix instead of reversing characters
2023-12-24 01:34:03 -08:00
James R. Barlow
fd593eb5e9
Reversing character order for RTL helps output
2023-12-24 01:24:44 -08:00
James R. Barlow
18353e1e94
Also fix space rendering
2023-12-24 00:09:47 -08:00
James R. Barlow
9adcad84da
Overhaul debug rendering to support possible use of Helvetica for testing
2023-12-23 23:47:35 -08:00
James R. Barlow
f2714586d8
Fix upside down glyphless font
2023-12-23 23:44:37 -08:00
James R. Barlow
e6095a9949
Fix text rendering issue with new hOCR text renderer
2023-12-20 12:26:06 -08:00
James R. Barlow
43618e6b3f
Move canvas API to pikepdf and import it
2023-12-02 19:42:35 -08:00
James R. Barlow
e97f89de3b
Refactor font so glyphless isn't as hard coded
2023-12-02 08:55:01 -08:00
James R. Barlow
11d3e32f1e
Fix hocrtransform CLI
2023-12-02 08:08:29 -08:00
James R. Barlow
d217856166
Make hocrdebug work, and try to handle CJK spacing better
2023-11-21 00:33:02 -08:00
James R. Barlow
e2be457e9b
Avoid divzero
2023-11-20 23:08:00 -08:00
James R. Barlow
4850f486d2
Make text API more like an accessor
2023-11-20 22:59:50 -08:00
James R. Barlow
729c7febd9
Fix placement of spaces in debug mode
2023-11-20 22:44:12 -08:00
James R. Barlow
6c6aca2f1e
Refactor save_state
2023-11-20 22:29:21 -08:00
James R. Barlow
c69823f496
Refactor; accumulate content stream as bytes rather than discrete pikepdf objects
2023-11-20 22:11:59 -08:00
James R. Barlow
73f8f6aac8
Add RTL output - seems to work, but debug does not
2023-11-20 20:28:07 -08:00
James R. Barlow
d944254e45
hocr: typing cont'd
2023-11-20 17:07:52 -08:00
James R. Barlow
f7ddffe554
hocr: typing
2023-11-20 16:52:55 -08:00
James R. Barlow
03669183d7
Rationalize canvas interface
2023-11-20 15:54:13 -08:00
James R. Barlow
74e101a2fa
Improve canvas interface with chaining
2023-11-20 14:42:48 -08:00
James R. Barlow
532cf18ad3
Restructure hocrtransform submodule to avoid having everything in __init__
2023-11-20 00:57:58 -08:00
James R. Barlow
0b90b697e2
More tidying
2023-11-20 00:43:43 -08:00
James R. Barlow
6be7c5f7c8
Fix colors and space box rendering
2023-11-20 00:30:54 -08:00
James R. Barlow
db2e5132e6
Remove some obsolete parameters
2023-11-20 00:10:55 -08:00
James R. Barlow
b14f6f778a
Tidying new hOCR renderer
2023-11-19 23:51:27 -08:00
James R. Barlow
a9466c4f58
Improve word box positioning
2023-11-19 23:51:27 -08:00
James R. Barlow
d9ae453a63
Significantly improvement overall
2023-11-19 23:51:27 -08:00
James R. Barlow
9841e09233
More adjustments
2023-11-19 23:51:27 -08:00
James R. Barlow
0ca314e066
Replace Rect with pikepdf.Rectangle, migrate line matrix to page
2023-11-19 23:51:27 -08:00
James R. Barlow
d7680cae27
Correcting Matrix logic helps
...
The good: don't have to do inverse and intermediate transforms.
The bad: skew looks bad, partly because the hOCR coordinate system is inconsistent around skew?
2023-11-19 23:51:27 -08:00
James R. Barlow
491b6bdb1f
Remove concept of HOCR_OK_LANGS
2023-11-19 23:51:27 -08:00
James R. Barlow
8d1e75017e
Remote reportlab backend and make reportlab a test-only dependency
2023-11-19 23:51:27 -08:00
James R. Barlow
e5df8e1315
Nearly pixel perfect
2023-11-19 23:51:27 -08:00
James R. Barlow
d739b91aef
Tidy up
2023-11-19 23:51:27 -08:00
James R. Barlow
686cfb2539
Add rendering of space between boxes
2023-11-19 23:51:27 -08:00
James R. Barlow
2633716bb7
Render interword spaces separately and avoid box overlap
2023-11-19 23:51:27 -08:00
James R. Barlow
0a07c0a44e
Fix more things
2023-11-19 23:51:27 -08:00
James R. Barlow
2ca6e110ca
Fix private accessors, rename pdf to canvas
2023-11-19 23:51:27 -08:00
James R. Barlow
334a07c839
Refactor debug printing
2023-11-19 23:51:27 -08:00
Odin Dahlström
James R. Barlow
Kara Engelhardt