James R. Barlow
16e4d342d2
Bug fix: --force-ocr should still run on pages with no images
...
Useful for people who want to reprocess text.
This also requires --oversample because DPI is undefined. To be fixed
in next commit.
2016-07-27 15:06:49 -07:00
James R. Barlow
bbd02926e1
Add helpful error message for PDFs that use algorithm 4
2016-06-23 13:13:17 -07:00
James R. Barlow
349ec5c81f
Provide more helpful error message if pypdf can't merge pages
2016-04-28 14:02:12 -07:00
James R. Barlow
fe14cb57c0
Fix ruffus exception output
...
I found this issue in ruffus 2.6.3
https://github.com/bunbun/ruffus/issues/65
also discussed here
https://github.com/bunbun/ruffus/pull/67
ruffus 2.6.3 RethrownJobError don't follow the normal conventions and
so its exception causes problems when they cross process boundaries.
This change carefully examines the various forms of ruffus exception
objects that can appear in 2.6.3 and parses them more carefully. It
also removes any direct posting of the exception to the logger because
this triggers another serializing of the exception object, mutating it
further.
2016-04-28 00:38:50 -07:00
James R. Barlow
e877d37ac8
--rotate-pages: Only apply rotation if we're reasonable confident
...
Take the threshold from tesseract's default value for -psm 1.
2016-04-14 13:49:44 -07:00
James R. Barlow
322085933b
unpaper: fix check for missing and old versions, add test case
2016-03-10 15:37:09 -08:00
James R. Barlow
6a380ee99c
Fix temporary file placed in wrong folder
2016-02-27 00:51:47 -08:00
James R. Barlow
dad2198394
Log information about detected page orientations in a summary line
2016-02-26 01:07:59 -08:00
James R. Barlow
e40fdc502d
Always dump stack trace for unexpected errors
2016-02-26 01:06:59 -08:00
James R. Barlow
71fbda8bf6
Adjust page orientation parsing to deal with change in Tess 3.04.01
2016-02-20 01:32:56 -08:00
James R. Barlow
f3b0434a87
Improve ability to capture error messages from tesseract on a crash
2016-02-19 03:48:49 -08:00
James R. Barlow
d4ef3411e0
Suppress --pdf-renderer tesseract warning in Docker image
...
Since the corrected font is provided in the Docker image, there's no
reason to show the warning.
2016-02-17 01:03:20 -08:00
James R. Barlow
60b2eb1455
Fix JPEG DPI: Pillow expects dpi=(x,y)
2016-02-16 07:29:20 -08:00
James R. Barlow
71e493a810
Fix case of JPEG missing DPI field
2016-02-16 05:29:32 -08:00
James R. Barlow
c50e3f1329
Complain about older tesseracts that don't have sharp2.ttf installed
2016-02-15 16:43:41 -08:00
James R. Barlow
7c691c21ab
Fix image layer rotation for pages with nonzero crop boxes
2016-02-10 17:48:33 -08:00
James R. Barlow
4ec51729d8
Partial fix for images not anchored to (0, 0)
2016-02-10 17:14:48 -08:00
James R. Barlow
6510bcad19
DPI information not transferred automatically from PNG to JPEG
2016-02-09 02:18:54 -08:00
James R. Barlow
1928a64cae
Better logging output for autorotation
2016-02-08 23:42:25 -08:00
James R. Barlow
1ba8b1aa4b
unpaper is lousy at deskewing, so let leptonica do it
2016-02-08 15:26:33 -08:00
James R. Barlow
2752bda80b
Merge branch 'feature/leptdeskew' into feature/logging
...
Need leptonica for testing now, I think
# Conflicts:
# ocrmypdf/tesseract.py
# requirements.txt
# setup.py
2016-02-08 12:34:48 -08:00
James R. Barlow
d30a879e2d
Fix test suite by running select_image_for_pdf unconditionally
...
The purpose of this change that caused the problem was a minor
optimization for the tesseract renderer path that had it pull an image
from select_image_for_pdf so that it could use a JPEG instead of PNG,
instead of taking it from preprocess_clean where it would only get a PNG
and make large files.
2016-02-08 02:33:03 -08:00
James R. Barlow
b0114c9174
More logging improvements
2016-02-08 01:31:15 -08:00
James R. Barlow
d2ba8c501f
Restore invisibletext for normal output
2016-02-08 01:14:39 -08:00
James R. Barlow
6a7ed7d359
Make logging output a lot more useful
2016-02-08 00:58:14 -08:00
James R. Barlow
6289afa1a6
Better: custom logging factory to avoid whatever ruffus is doing
2016-02-08 00:18:52 -08:00
James R. Barlow
9bb6fa04cb
Return logging to a semblance of normalcy
2016-02-08 00:09:31 -08:00
James R. Barlow
afb6f6f5c9
Render preview as .jpg instead of .png
...
Smaller file size of JPEG seems to help performance, although the
difference is only about 1%.
2016-02-07 15:49:10 -08:00
James R. Barlow
178aee4687
Make rotation optional (for now it's off, possibly should be on)
2016-02-07 15:43:45 -08:00
James R. Barlow
8484caddfb
Tweak pipeline, allowing --pdf-renderer to use JPEGs instead of PNGs
2016-02-07 15:36:51 -08:00
James R. Barlow
08313316de
Cleanup auto-rotation
2016-02-07 15:06:54 -08:00
James R. Barlow
1d0eca5c63
All four rotation directions working
2016-02-07 06:09:01 -08:00
James R. Barlow
fe89232a30
Fix autorotate for some lossless cases
2016-02-07 05:59:46 -08:00
James R. Barlow
4b51b521e2
Implement autorotate (provided lossless reconstruction is disabled)
...
Works for a single page file, probably
Although arguably rotation is not quite lossless, and the two could be
mutually exclusive anyway, so maybe this is it. Did not check in some
debugging changes (lossless=False, text debugging=True)
PyPDF seems to get merging wrong when one of the pages is rotated.
2016-02-07 03:27:33 -08:00
James R. Barlow
daa3916430
Fix img2pdf 0.2 usage
...
All tests pass when forced to rely on img2pdf, so seems okay
2016-02-05 15:13:26 -08:00
James R. Barlow
e9b87cefcc
Try img2pdf 0.2
2016-02-05 14:38:37 -08:00
James R. Barlow
b49f5a7d77
Support optionally using leptonica to deskew
...
unpaper doesn't seem to be good at deskewing. It fails on test case
with a lot of italics. I think it also struggles on pages with a lot
of whitespace. Leptonica continues to shine here.
However, this is only a first crack at Leptonica. The leptonica module
should be redone to use cffi (more extensible).
Also considering the possibility of making all Lept calls in a forked
process to insulate the calling process from C code crashes and the
messy redirect of stdout/stderr to read Leptonica's errors.
I don't think the redirect is a huge problem as long as multiprocesses
rather than multithreads are used. The ruffus child process that is
handling a page is single threaded and will not be affected by the
redirection. It just feels dirty. The main reason to consider a child
process is crash isolation.
2016-01-19 17:43:40 -08:00
James R. Barlow
37c508f3f8
Better versioning: no silly version files, but wrong ver in development
...
Small price to pay.
2016-01-19 16:07:52 -08:00
James R. Barlow
26e36422cc
More fiddling with version
2016-01-19 15:07:21 -08:00
James R. Barlow
f82cb002bc
Try automatic versioning with setuptools_scm
2016-01-19 13:27:18 -08:00
James R. Barlow
c1eb047a4b
Fix name of pdfa_def.ps
...
Used to include a copy of the parent dir's name.
2016-01-19 13:11:03 -08:00
James R. Barlow
626ca18f5c
Remove stale comment
2016-01-19 13:02:35 -08:00
James R. Barlow
a0952bfca3
Optimize: use img2pdf stream instead of repeated copies
2016-01-18 20:24:46 -08:00
James R. Barlow
fc0479f110
Fix all but test_oversample[hocr]
2016-01-15 15:46:47 -08:00
James R. Barlow
62728205b6
Implement image+text merging in other cases
...
5 failed, 28 passed
failures:
test_oversample[hocr], test_skip_ocr, test_skip_big, test_maximum_options[hocr],
test_blank_input_pdf,
2016-01-15 15:38:08 -08:00
James R. Barlow
dc0fb25e64
Render hocr page: no longer needs an image as input
2016-01-15 15:16:47 -08:00
James R. Barlow
7067110308
Add safety check to prevent merge from running when not sensible
2016-01-15 14:54:45 -08:00
James R. Barlow
599d889703
Implement "perfect reconstruction" - transfer page and watermark OCR layer
...
Works, does not account for changes to clean/deskew, etc.
Surprisingly, it works. PyPDF2 fixes since last attempt?
2016-01-15 14:39:12 -08:00
James R. Barlow
074c1d71b4
Activate --tesseract-pagesegmode
2016-01-11 17:19:32 -08:00
James R. Barlow
1fca9a004d
Adjust command line parameters
...
Was splitting each argument to --tesseract-config into a list of single
character strings
2016-01-11 16:57:19 -08:00
James R. Barlow