menekis/OCRmyPDF
Public Access
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
1,467 Commits 5 Branches 303 Tags
acc6698ab3eda3d2a46ee92781d62fbb30afd1fd
Commit Graph
1467 Commits
This Branch
This Branch
All Branches
Author SHA1 Message Date
James R. Barlow acc6698ab3 Make XML metadata test actually work 2018-05-10 20:37:10 -07:00
James R. Barlow 606d3e6aa1 Remove tests that exercise obsolete features (tesseract, -g) 2018-05-10 20:33:32 -07:00
James R. Barlow 687a7954d6 test_main: uses leptonica 2018-05-10 19:05:31 -07:00
James R. Barlow 36a53a7b37 Weave: Unconditionally rotate and scale the text layerThis solves two issues. First, the text layer can end up being adifferent size, probably if the DPI is not an integer; scaling helps itfit slightly better. Second, other printable text on the page can end uphorizontally scaled or misaligned if we don't all of our drawing in aq/Q pair. 2018-05-10 19:03:31 -07:00
James R. Barlow 0a5982a902 PyMuPDF tweaks: don't clean 
In MuPDF 1.13 clean might be unreliable, so explicitly don't do it,
even though it doesn't cause trouble in 1.12.
 2018-05-10 18:50:52 -07:00
James R. Barlow 601863f9e9 Return to PyMuPDF 1.12.5 2018-05-10 18:47:10 -07:00
James R. Barlow c9ce731119 Fix DPI mismatch between OCR page and source page 2018-05-10 17:34:08 -07:00
James R. Barlow abed8e034e Add metadata preservation test from stash 2018-05-10 16:43:28 -07:00
James R. Barlow 63032d304d Revert "Since PyMuPDF 1.13.3 corrupts text, pin 1.12.5 and work around it" 
This reverts commit b0ce7c63dd.
 2018-05-10 16:27:17 -07:00
James R. Barlow a57ecede78 Refactor textareas to remove duplicate code 2018-05-10 16:26:52 -07:00
James R. Barlow b0ce7c63dd Since PyMuPDF 1.13.3 corrupts text, pin 1.12.5 and work around it 2018-05-10 16:10:24 -07:00
James R. Barlow d139a11c16 Weave: periodically save to prevent indefinite growth of open file list 2018-05-10 15:08:57 -07:00
James R. Barlow aef043db0b Revise parameter validation for output-type, pdf-renderer, lang 2018-05-10 14:53:22 -07:00
James R. Barlow b8f3ead541 Remove tesseract renderer entirely 
Grafting lets us work with older Tesseract versions as if they could use
sandwich, so there is no point in keeping it. It's been deprecated for a
long time now anyway.
 2018-05-10 14:06:13 -07:00
James R. Barlow e0bb898f29 Remove hocr debug renderer (-g) 
The fact that this produces additional pages makes it a maintenance
burden. hocr can be debugged using hocrtransform.
 2018-05-10 13:48:39 -07:00
James R. Barlow 45336c7c28 textareas: filter out images 2018-05-10 01:17:28 -07:00
James R. Barlow 20aabb2e83 When deciding if there is a text on a page, ignore the margins 
Margins may include watermarks or digital stamps on otherwise
text-free pages.
 2018-05-10 01:16:11 -07:00
James R. Barlow da80d3f354 Add unconditional (for now) whiteout of text areas 2018-05-07 17:37:46 -07:00
James R. Barlow 001c8d7678 Upgrade PyMuPDF version 2018-05-07 16:24:26 -07:00
James R. Barlow 423cef08bf Handle procset properly 2018-05-02 14:48:02 -07:00
James R. Barlow 04580accb4 Document aliasing of tesseract renderer 2018-05-02 14:47:38 -07:00
James R. Barlow 6376f77b8c Refactor, remove trigonometry 2018-05-02 12:30:34 -07:00
James R. Barlow e27e614ed9 Fixed rotation hard case 2018-05-02 01:32:11 -07:00
James R. Barlow b0c04704a1 Fixed all but one rotation case 2018-05-02 01:24:21 -07:00
James R. Barlow 6bb6bf8323 Fix correction angle used from wrong page 2018-05-02 01:00:30 -07:00
James R. Barlow e22fe8aefc Silence debug messages 2018-05-01 23:51:54 -07:00
James R. Barlow 76276f61e5 Split out rotation related tests 2018-05-01 23:51:35 -07:00
James R. Barlow bfd26e6ec6 Tests: confirm OCR layer copied 2018-05-01 23:16:41 -07:00
James R. Barlow d787e1ea0f ghostscript.py not saved in last commit 
Given importance of last one, confirmed that when the file is saved all tests pass too.
Passing is invariant with this change.
 2018-05-01 22:59:22 -07:00
James R. Barlow b5d7e9cbb0 Fix all issues with rotations 
All tests now pass
 2018-05-01 22:50:20 -07:00
James R. Barlow f3b6d9dcdf Fix a comment about Tesseract behavior in certain versions 2018-05-01 21:31:09 -07:00
James R. Barlow a9abe13185 Remove the old tesseract pdf_renderer 2018-05-01 17:31:34 -07:00
James R. Barlow 6b315e8315 Add ability to disable cache 2018-05-01 15:52:00 -07:00
James R. Barlow 37677de884 Fix regressions: pdfa.ps not used, PDF/A failures, handling of text layers with no font 2018-05-01 15:51:46 -07:00
James R. Barlow c7387de325 Fix auto rotate 2018-05-01 15:18:28 -07:00
James R. Barlow 2495b1e038 Refactor find font, get test cases working again 2018-05-01 14:48:41 -07:00
James R. Barlow 073ee52ce7 Use hocr and weave; eliminate old combine layers and merge pages 2018-05-01 14:21:53 -07:00
James R. Barlow 54150a14e9 Further elimination of tesseract renderer special casing 
We don't need to keep a "skip page" around anymore since
skipping means just not grafting on the text layer.
 2018-05-01 13:36:20 -07:00
James R. Barlow 88ff091cce Unify tesseract and sandwich renderer paths 
Since the new weaving method copies the font and content
stream from the Tesseract PDF, it doesn't matter if Tesseract
happens to have an image or not.
If Tesseract is text-only capable we use that feature for efficiency,
but ignore the image either way.
 2018-05-01 13:24:20 -07:00
James R. Barlow e87a5776f1 Remove now-unnecessary code to rotate pages 
Track only the decision to change rotation.
 2018-05-01 13:01:25 -07:00
James R. Barlow 0806ce6406 Fix rotation for unsplit (modulo --rotate-pages) 2018-04-30 20:58:42 -07:00
James R. Barlow 6409894a71 feature/unsplit-try-imagerotate 2018-04-30 20:48:59 -07:00
James R. Barlow e7286f6129 Unsplit now works with multipage, --force-ocr 2018-04-30 14:46:20 -07:00
James R. Barlow 2ab94b3151 unsplit: it's alive 
First successful file output.
 2018-04-28 01:57:41 -07:00
James R. Barlow 7ee90890ec Add copying of essential information from Tesseract textonly 2018-04-27 23:19:08 -07:00
James R. Barlow 8d2a917676 Page unsplit, development 2018-04-25 21:56:43 -07:00
James R. Barlow 44b4afa534 Begin conversion from page splititng to page markers 2018-04-23 22:57:50 -07:00
James R. Barlow 775be3933c Cherrypick merge_pages unification 2018-04-20 23:08:15 -07:00
Hugoandjbarlow83 d761d80750 Use more standard __version__ rather than PILLOW_VERSION (#257) 2018-04-19 23:35:32 -07:00
James R. Barlow 0b10db91be Fix regression: Disable Ghostscript JPEG passthrough entirely v6.1.5 2018-04-17 17:00:24 -07:00