menekis/OCRmyPDF
Public Access
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
1,866 Commits 5 Branches 303 Tags
5ed05e08b1f71e663316a4b2c8b1c84967c22701
Commit Graph
1866 Commits
This Branch
This Branch
All Branches
Author SHA1 Message Date
James R. Barlow 5ed05e08b1 Fix "no languages" test and misuse of os.environ 2018-11-09 01:57:11 -08:00
James R. Barlow 58b26f6715 Leptonica: learn to despeckle 1bpp images 2018-11-07 01:49:13 -08:00
James R. Barlow 806daf4284 leptonica: reduce boilerplate for PIX (2/2) 2018-11-06 20:33:40 -08:00
James R. Barlow c64bc9329e leptonica: reduce boilerplate for wrapper classes (except PIX) 2018-11-06 20:12:09 -08:00
James R. Barlow dd01745519 Leptonica: add masked threshold fn 2018-11-06 19:31:06 -08:00
James R. Barlow 501ce726e7 Fix two failing tests 2018-11-06 11:16:08 -08:00
James R. Barlow 03076e89ce Leptonica: reduce verbosity, more error trapping, more garbage collection 2018-11-06 11:10:59 -08:00
James R. Barlow 02f37293ee Integrate barcode masking 2018-11-05 13:01:13 -08:00
James R. Barlow 590942ad14 Leptonica: Add barcode API 2018-11-05 01:48:38 -08:00
James R. Barlow 2ac028c759 test: Add a basic redo OCR test 2018-11-04 15:54:41 -08:00
James R. Barlow 2125b5bfab Remove text detection from our parser interpret_contents 
It's redundant now
 2018-11-04 15:47:55 -08:00
James R. Barlow b96532caa4 Only do detailed page analysis when needed by --redo-ocr 2018-11-04 15:40:49 -08:00
James R. Barlow 995fc58466 Move Ghostscript text analysis into its own module 2018-11-04 14:55:48 -08:00
James R. Barlow c023cae299 Make pdfminer Type3 patch conditional on PScript5.dll 
It appears that PDFs created by this software have a bug in their BBox
which will cause us to misjudge the space occupied by the font.

Other programs probably work around this by ignoring BBox and reading
each character procedure.
 2018-11-04 01:53:53 -07:00
James R. Barlow 8b9ab25125 coverage: test compile leptonica 2018-11-02 01:55:25 -07:00
James R. Barlow 77e87abe8f coverage: ensure get_orientation is checked 2018-11-02 01:32:20 -07:00
James R. Barlow 3be02e1e8d coverage: improve leptonic; don't create objects with null pointers 2018-11-02 01:10:10 -07:00
James R. Barlow 64c9ede979 leptonica: barcodes, BOXA 2018-11-02 00:42:01 -07:00
James R. Barlow 5b8d197812 coverage: make it more likely timeout is tested 2018-11-02 00:41:15 -07:00
James R. Barlow 2cba62dc4f coverage: ensure rotation is actually tested 2018-11-02 00:40:56 -07:00
James R. Barlow 288e28328f coverage: add qpdf 2018-11-02 00:37:33 -07:00
James R. Barlow b8214b3c49 coverage: exclude unicodefun.py 2018-11-02 00:33:08 -07:00
James R. Barlow 8681693994 Set up code coverage (it works with multiprocessing now!) 2018-11-02 00:31:50 -07:00
James R. Barlow 1364c63b7c Fix failure to pickle file with AcroForm 2018-11-01 20:07:53 -07:00
James R. Barlow 4ba9e8fe25 Add AcroForm detection 2018-10-30 22:28:44 -07:00
James R. Barlow a195713bb4 Throw exception on corrupt text 2018-10-30 16:35:09 -07:00
James R. Barlow 600d31a907 Require pikepdf 0.3.7 2018-10-30 16:22:05 -07:00
James R. Barlow be31cec332 Add corrupt text warning (when using --redo-ocr) 2018-10-30 16:19:58 -07:00
James R. Barlow 22a7cd3421 Add argument checks for --redo-ocr 2018-10-30 16:19:13 -07:00
James R. Barlow 8b61d2d521 pdfminer: If font descent claims to be positive, treat it as negative 2018-10-30 14:40:53 -07:00
James R. Barlow 559e5269d2 Ensure inline image is parsed correctly 
Requires pikepdf > 0.3.6
 2018-10-29 23:30:53 -07:00
James R. Barlow ebf6acb318 pdfminer patch: Type3 font height calculation is incorrect 
Not sure where it goes wrong or why it needs special treatment, but
this does address it.
 2018-10-29 22:27:25 -07:00
James R. Barlow 7acd75f013 pipeline: fix bbox coordinates 2018-10-29 22:26:37 -07:00
James R. Barlow 93623b2226 Refactor TextboxInfo 2018-10-29 14:46:40 -07:00
James R. Barlow d71fd089cb layout: allow names beginning with /i0123 for now 
Showed up in GGastro2.pdf. Need to check if this pattern has valid
Unicode mappings but allow for now.
 2018-10-29 14:45:59 -07:00
James R. Barlow 05aa43c856 Require pdfminer 2018-10-29 12:45:15 -07:00
James R. Barlow de80fb6bc8 Fix some failing tests after --redo-ocr changes 2018-10-29 11:49:38 -07:00
James R. Barlow 8e396f4be2 Document --redo-ocr more accurately 2018-10-29 02:03:58 -07:00
James R. Barlow efec6da377 Fix error on serializing bad character markers 
(Since they held a reference to their font, which in turn, had an
open file handle.)
 2018-10-29 02:02:00 -07:00
James R. Barlow 00ef53195e Fix corrupt Unicode mapping detection's false positives 2018-10-29 01:30:19 -07:00
James R. Barlow f564aaf485 Remove only_ocr_text 2018-10-28 22:41:18 -07:00
James R. Barlow 5ac2d31d0d Redo OCR can now handle visible and invisible text, so adjust accordingly 
Still can't filter out corrupt text
 2018-10-28 14:06:25 -07:00
James R. Barlow fda890ab47 pdfinfo: further layout improvements 
Rather than grouping visible/invisible in a custom analysis step,
use pdfminer's analysis and iterate.
Make iteration predicate and return more generic.
 2018-10-28 14:05:50 -07:00
James R. Barlow e6d64be890 pdfinfo: formatting 2018-10-27 23:22:44 -07:00
James R. Barlow 0e4d978d20 pdfinfo: all -> not any 2018-10-27 23:22:28 -07:00
James R. Barlow b12c2cfedf Fix handling of Type3 fonts with no ToUnicode mapping 2018-10-27 01:24:48 -07:00
James R. Barlow 58cc70725e Reorganize around getting bboxes for visible/invisible text 2018-10-26 01:07:02 -07:00
James R. Barlow 339afb02aa --redo-ocr now works in the presence of printable text 2018-10-25 16:53:47 -07:00
James R. Barlow 7ba0ff5c36 Fix strip invisible text bug: missing BT operator 2018-10-25 16:52:23 -07:00
James R. Barlow ff41fbf673 Add pdfminer based layout analysis 2018-10-25 12:42:35 -07:00