diff --git a/.gitignore b/.gitignore index 610fe588..e545c422 100644 --- a/.gitignore +++ b/.gitignore @@ -32,6 +32,7 @@ htmlcov/ *.profile /*.pdf /*.qdf +/*.png /scratch.py IDEAS log/ diff --git a/docs/advanced.rst b/docs/advanced.rst index bc84c5a0..a0f7b7d8 100644 --- a/docs/advanced.rst +++ b/docs/advanced.rst @@ -13,7 +13,9 @@ If a page in a PDF seems to have text, by default OCRmyPDF will exit without mod If ``--skip-text`` is issued, then no OCR will be performed on pages that already have text. The page will be copied to the output. This may be useful for documents that contain both "born digital" and scanned content, or to use OCRmyPDF to normalize and convert to PDF/A regardless of their contents. -If ``--force-ocr`` is issued, then all pages will be rasterized to images, discarding any hidden OCR text, and rasterizing any printable text. This is useful for redoing OCR, for fixing OCR text with a damaged character map (text is selectable but not searchable), and destroying redacted information. +If ``--redo-ocr`` is issued, then a detailed text analysis is performed. Text is categorized as either visible or invisible. Invisible text (OCR) is stripped out. Then an image of each page is created with visible text masked out. The page image is sent for OCR, and any additional text is inserted as OCR. If a file contains a mix of text and bitmap images that contain text, OCRmyPDF will locate the additional text in images without disrupting the existing text. + +If ``--force-ocr`` is issued, then all pages will be rasterized to images, discarding any hidden OCR text, and rasterizing any printable text. This is useful for redoing OCR, for fixing OCR text with a damaged character map (text is selectable but not searchable), and destroying redacted information. Any forms and vector graphics will be rasterized as well. Time and image size limits diff --git a/docs/cookbook.rst b/docs/cookbook.rst index db2c2bf5..996062d9 100644 --- a/docs/cookbook.rst +++ b/docs/cookbook.rst @@ -57,7 +57,6 @@ You can increase (decrease) the parameter ``--rotate-pages-threshold`` to make p If the page is "just a little off horizontal", like a crooked picture, then you want ``--deskew``. ``--rotate-pages`` is for when the cardinal angle is wrong. - OCR languages other than English """""""""""""""""""""""""""""""" @@ -70,7 +69,6 @@ By default OCRmyPDF assumes the document is English. Language packs must be installed for all languages specified. See :ref:`Installing additional language packs `. - Produce PDF and text file containing OCR text """"""""""""""""""""""""""""""""""""""""""""" @@ -116,7 +114,6 @@ If you have multiple images, you must use ``img2pdf`` to convert the images to P ImageMagick ``convert`` can also convert a group of images to PDF, but in the author's experience it takes a long time, transcodes unnecessarily and gives poor results. - Image processing ---------------- @@ -132,6 +129,8 @@ OCRmyPDF perform some image processing on each page of a PDF, if desired. The s * ``--clean-final`` uses unpaper to clean up pages before OCR and inserts the page into the final output. You will want to review each page to ensure that unpaper did not remove something important. +* ``-mask-barcodes`` will "cover up" any barcodes detected in the image of a page. Barcodes are known to confuse Tesseract OCR and interfere with the recognition of text on the same baseline as a barcode. The output file will contain the unaltered image of the barcode. + .. note:: In many cases image processing will rasterize PDF pages as images, potentially losing quality. @@ -140,7 +139,6 @@ OCRmyPDF perform some image processing on each page of a PDF, if desired. The s ``--clean-final`` and ``-remove-background`` may leave undesirable visual artifacts in some images where their algorithms have shortcomings. Files should be visually reviewed after using these options. - OCR and correct document skew (crooked scan) """""""""""""""""""""""""""""""""""""""""""" @@ -167,28 +165,22 @@ If you set ``--tesseract-timeout 0`` OCRmyPDF will apply its image processing wi ocrmypdf --tesseract-timeout=0 --remove-background input.pdf output.pdf -Redo OCR -"""""""" +Redo existing OCR +""""""""""""""""" -To redo OCR on a file OCRed with other OCR software or a previous version of OCRmyPDF and/or Tesseract, you may use the ``--force-ocr`` argument. Normally, OCRmyPDF does not modify files that already appear to contain OCR text. +To redo OCR on a file OCRed with other OCR software or a previous version of OCRmyPDF and/or Tesseract, you may use the ``--redo-ocr`` argument. (Normally, OCRmyPDF will exit with an error if asked to modify a file with OCR.) + +This may be helpful for users who want to take advantage of accuracy improvements in Tesseract 4.0 for files they previously OCRed with an earlier version of Tesseract and OCRmyPDF. .. code-block:: bash - ocrmypdf --force-ocr input.pdf output.pdf + ocrmypdf --redo-ocr input.pdf output.pdf -Note that the method above will force rasterization of all pages, potentially reducing quality or losing vector content. +This method will replace OCR without rasterizing, reducing quality or removing vector content. If a file contains a mix of pure digital text and OCR, digital text will be ignored and OCR will be replaced. As such this mode is incompatible with image processing options, since they alter the appearance of the file. -To ensure quality is preserved, one could extract all of the images and rebuild the PDF for a lossless transformation. This recipe does not work when PDFs contain multiple images per page, as many do in practice. It will also lose any page rotation information. - -.. code-block:: bash - - pdfimages -all old-ocr.pdf prefix # extract all images - img2pdf -o temp.pdf prefix* # construct new PDF from the images - # review the new PDF to ensure it visually matches the old one - ocrmypdf --output-type pdf temp.pdf new-ocr.pdf - -``--output-type pdf`` is used here to avoid using Ghostscript which will also rasterize images. +In some cases, existing OCR cannot be detected or replaced. Files produced by OCRmyPDF v2.2 or earlier, for example, are internally represented as having visible text with an opaque image drawn on top. This situation cannot be detected. +If ``--redo-ocr`` does not work, you can use ``--force-ocr``, which will force rasterization of all pages, potentially reducing quality or losing vector content. Improving OCR quality --------------------- @@ -199,7 +191,6 @@ Rotating pages and deskewing helps to ensure that the page orientation is correc OCR quality will suffer if the resolution of input images is not correct (since the range of pixel sizes that will be checked for possible fonts will also be incorrect). - PDF optimization ---------------- diff --git a/docs/introduction.rst b/docs/introduction.rst index 7038d3fc..c4b4f52d 100644 --- a/docs/introduction.rst +++ b/docs/introduction.rst @@ -64,7 +64,7 @@ In the case of a PDF that is nothing other than a container of images (no rotati OCRmyPDF uses several strategies depending on input options and the input PDF itself, but generally speaking it rasterizes a page for OCR and then grafts the OCR back onto the original. As such it can handle complex PDFs and still preserve their contents as much as possible. -OCRmyPDF also supports a many, many edge cases that have cropped over several years of development. We support PDF features like images inside of Form XObjects, and pages with UserUnit scaling. We support rare image formats like non-monochrome 1-bit images. Thanks to pikepdf and QPDF, we auto-repair PDFs that are damaged. (Not that you need to know what any of these are! You should be able to throw any PDF at it.) +OCRmyPDF also supports a many, many edge cases that have cropped over several years of development. We support PDF features like images inside of Form XObjects, and pages with UserUnit scaling. We support rare image formats like non-monochrome 1-bit images. We warn about files you may not to OCR. Thanks to pikepdf and QPDF, we auto-repair PDFs that are damaged. (Not that you need to know what any of these are! You should be able to throw any PDF at it.) Limitations @@ -76,20 +76,20 @@ OCRmyPDF is limited by the Tesseract OCR engine. As such it experiences these l * It is not capable of recognizing handwriting. * It may find gibberish and report this as OCR output. * If a document contains languages outside of those given in the ``-l LANG`` arguments, results may be poor. -* It is not always good at analyzing the natural reading order of documents. For example, it may fail to recognize that a document contains two columns and join text across the columns. +* It is not always good at analyzing the natural reading order of documents. For example, it may fail to recognize that a document contains two columns, and may try to join text across columns. * Poor quality scans may produce poor quality OCR. Garbage in, garbage out. * It does not expose information about what font family text belongs to. OCRmyPDF is also limited by the PDF specification: * PDF encodes the position of text glyphs but does not encode document structure. There is no markup that divides a document in sections, paragraphs, sentences, or even words (since blank spaces are not represented). As such all elements of document structure including the spaces between words must be derived heuristically. Some PDF viewers do a better job of this than others. -* Because some popular open source PDF viewers have a particularly hard time with spaces betweem words, OCRmyPDF appends a space to each text element as a workaround. While this mixes document structure with graphical information that ideally should be left to the PDF viewer to interpret, it improves compatibility with some viewers and does not cause problems for better ones. +* Because some popular open source PDF viewers have a particularly hard time with spaces betweem words, OCRmyPDF appends a space to each text element as a workaround (when using ``--pdf-renderer hocr``). While this mixes document structure with graphical information that ideally should be left to the PDF viewer to interpret, it improves compatibility with some viewers and does not cause problems for better ones. Ghostscript also imposes some limitations: * PDFs containing JBIG2-encoded content will be converted to CCITT Group4 encoding, which has lower compression ratios, if Ghostscript PDF/A is enabled. * PDFs containing JPEG 2000-encoded content will be converted to JPEG encoding, which may introduce compression artifacts, if Ghostscript PDF/A is enabled. -* Ghostscript may transcode grayscale and color images, either lossy to lossless or lossless to lossy, based on an internal algorithm. This behavior can be suppressed by setting ``--pdfa-image-compression`` to ``jpeg`` or ``lossless`` to set all images to one type or the other. Ghostscript has no option to maintain the input image's format. +* Ghostscript may transcode grayscale and color images, either lossy to lossless or lossless to lossy, based on an internal algorithm. This behavior can be suppressed by setting ``--pdfa-image-compression`` to ``jpeg`` or ``lossless`` to set all images to one type or the other. Ghostscript has no option to maintain the input image's format. (Ghostscript 9.25+ can copy JPEG images without transcoding them; earlier versions will transcode.) Regarding OCRmyPDF itself: diff --git a/docs/release_notes.rst b/docs/release_notes.rst index ba5c6bf6..d36cf4d4 100644 --- a/docs/release_notes.rst +++ b/docs/release_notes.rst @@ -14,6 +14,33 @@ Note that it is licensed under GPLv3, so scripts that ``import ocrmypdf`` and ar replace: `#$1 `_ +v7.3.0 +------ + +- Added a new feature ``--redo-ocr`` to detect existing OCR in a file, remove it, and redo the OCR. This may be particularly helpful for anyone who wants to take advantage of OCR quality improvements in Tesseract 4.0. Note that OCR added by OCRmyPDF before version 3.0 cannot be detected since it was not properly marked as invisible text in the earliest versions. OCR that constructs a font from visible text, such as Adobe Acrobat's ClearScan. + +- OCRmyPDF's content detection is generally more sophisticated. It learns more about the contents of each PDF and makes better recommendations: + + - OCRmyPDF can now detect when a PDF contains text that cannot be mapped to Unicode (meaning it is readable to human eyes but copy-pastes as gibberish). In these cases it recommends ``--force-ocr`` to make the text searchable. + + - PDFs containing vector objects are now rendered at more appropriate resolution for OCR. + + - We now exit with an error for PDFs that contain Adobe LiveCycle Designer's dynamic XFA forms. Currently the open source community does not have tools to work with these files. + + - OCRmyPDF now warns when a PDF that contains Adobe AcroForms, since such files probably do not need OCR. It can work with these files. + +- Added three new **experimental** features. The name, syntax and behavior of these arguments is subject to change. They may also be incompatible with some other features. + + - ``--remove-vectors`` which strips out vector graphics. This can improve OCR quality since OCR will not search artwork for readable text; however, it currently removes "text as curves" as well. + + - ``--mask-barcodes`` to detect and suppress barcodes in files. We have observed that barcodes can interfere with OCR. + + - ``--threshold`` which uses a more sophisticated thresholding algorithm than is currently in use in Tesseract OCR. This works around a `known issue in Tesseract `_ with text on bright backgrounds. + +- Fixed an issue where an error message was not reported when the installed Ghostscript was very old. + +- New dependency: pdfminer.six 20181108. + v7.2.1 ------ @@ -152,6 +179,15 @@ v7.0.0 + It may be necessary to separately ``pip install pycparser`` to avoid `another Python 3.7 issue `_. +v6.2.5 +------ + +- Disable a failing test due to Tesseract 4.0rc1 behavior change. Previously, Tesseract would exit with an error message if its configuration was invalid, and OCRmyPDF would intercept this message. Now Tesseract issues a warning, which OCRmyPDF v6.2.5 may relay or ignore. (In v7.x, OCRmyPDF will respond to the warning.) + +- This release branch no longer supports using the optional PyMuPDF installation, since it was removed in v7.x. + +- This release branch no longer supports macOS. macOS users should upgrade to v7.x. + v6.2.4 ------ diff --git a/requirements/dev.txt b/requirements/dev.txt index e74bb3a7..4faf987d 100644 --- a/requirements/dev.txt +++ b/requirements/dev.txt @@ -1,4 +1,4 @@ check-manifest >= 0.35 twine >= 1.8.1 -coverage >= 4.4 +coverage >= 4.5 GitPython == 2.1.3 diff --git a/requirements/main.txt b/requirements/main.txt index c4f7021a..3b9884b1 100644 --- a/requirements/main.txt +++ b/requirements/main.txt @@ -1,11 +1,13 @@ # requirements.txt can be used to replicate the developer's build environment # setup.py lists a separate set of requirements that are looser to simplify # installation +chardet == 3.0.4 cffi == 1.11.5 -img2pdf == 0.3.0 -pikepdf == 0.3.4 +img2pdf == 0.3.1 +pdfminer.six == 20181108 +pikepdf == 0.3.7 Pillow >= 5.0.0, != 5.1.0 ; sys_platform == "darwin" -pycparser == 2.18 +pycparser == 2.19 python-xmp-toolkit == 2.0.1 -reportlab == 3.4.0 -ruffus == 2.7.0 +reportlab == 3.5.9 +ruffus == 2.8.0 diff --git a/requirements/test.txt b/requirements/test.txt index c1d250b5..f28432a3 100644 --- a/requirements/test.txt +++ b/requirements/test.txt @@ -1,4 +1,4 @@ -pytest >= 3.2 +pytest == 3.9.3 pytest-helpers-namespace pytest-xdist pytest-cov diff --git a/setup.py b/setup.py index 5f1cb5d0..611d4c16 100644 --- a/setup.py +++ b/setup.py @@ -249,8 +249,9 @@ setup( ], install_requires=[ 'cffi >= 1.9.1', # must be a setup and install requirement - 'img2pdf >= 0.2.4, < 0.4', # pure Python, so track HEAD closely - 'pikepdf >= 0.3.5, < 0.4', + 'img2pdf >= 0.3.0, < 0.4', # pure Python, so track HEAD closely + 'pdfminer.six == 20181108', + 'pikepdf >= 0.3.7, < 0.4', 'Pillow >= 4.0.0, != 5.1.0 ; sys_platform == "darwin"', # Pillow < 4 has BytesIO/TIFF bug w/img2pdf 0.2.3 # block 5.1.0, broken wheels diff --git a/src/ocrmypdf/__main__.py b/src/ocrmypdf/__main__.py index 3650787c..7424cb1d 100755 --- a/src/ocrmypdf/__main__.py +++ b/src/ocrmypdf/__main__.py @@ -244,19 +244,38 @@ preprocessing.add_argument( '--oversample', metavar='DPI', type=numeric(int, 0, 5000), default=0, help="Oversample images to at least the specified DPI, to improve OCR " "results slightly") +preprocessing.add_argument( + '--remove-vectors', action='store_true', + help="EXPERIMENTAL. Mask out any vector objects in the PDF so that they " + "will not be included in OCR. This can eliminate false characters.") +preprocessing.add_argument( + '--mask-barcodes', action='store_true', + help="EXPERIMENTAL. Mask out any barcodes that appear in the PDF so they are not " + "considered during OCR. Barcodes can introduce false characters into " + "OCR.") +preprocessing.add_argument( + '--threshold', action='store_true', + help="EXPERIMENTAL. Threshold image to 1bpp before sending it to Tesseract for OCR. Can " + "improve OCR quality compared to Tesseract's thresholder.") ocrsettings = parser.add_argument_group( "OCR options", "Control how OCR is applied") ocrsettings.add_argument( '-f', '--force-ocr', action='store_true', - help="Rasterize any fonts or vector objects on each page, apply OCR, and " + help="Rasterize any text or vector objects on each page, apply OCR, and " "save the rastered output (this rewrites the PDF)") ocrsettings.add_argument( '-s', '--skip-text', action='store_true', help="Skip OCR on any pages that already contain text, but include the " "page in final output; useful for PDFs that contain a mix of " "images, text pages, and/or previously OCRed pages") +ocrsettings.add_argument( + '--redo-ocr', action='store_true', + help="Attempt to detect and remove the hidden OCR layer from files that " + "were previously OCRed with OCRmyPDF or another program. Apply OCR " + "to text found in raster images. Existing visible text objects will " + "not be changed. If there is no existing OCR, OCR will be added.") ocrsettings.add_argument( '--skip-big', type=numeric(float, 0, 5000), metavar='MPixels', help="Skip OCR on pages larger than the specified amount of megapixels, " @@ -461,12 +480,18 @@ def check_options_output(options, log): ) lossless_reconstruction = False - if options.pdf_renderer in ('hocr', 'sandwich'): - if not any((options.deskew, options.clean_final, options.force_ocr, - options.remove_background)): - lossless_reconstruction = True + if not any((options.deskew, options.clean_final, options.force_ocr, + options.remove_background)): + lossless_reconstruction = True options.lossless_reconstruction = lossless_reconstruction + if not options.lossless_reconstruction and options.redo_ocr: + raise argparse.ArgumentError( + None, + "--redo-ocr is not currently compatible with --deskew, " + "--clean-final, and --remove-background" + ) + def check_options_sidecar(options, log): if options.sidecar == '\0': @@ -511,10 +536,15 @@ def check_options_preprocessing(options, log): def check_options_ocr_behavior(options, log): - if options.force_ocr and options.skip_text: + exclusive_options = sum( + [(1 if opt else 0) + for opt in (options.force_ocr, options.skip_text, options.redo_ocr) + ] + ) + if exclusive_options >= 2: raise argparse.ArgumentError( None, - "Error: --force-ocr and --skip-text are mutually exclusive.") + "Error: choose only one of --force-ocr, --skip-text, --redo-ocr.") def check_options_optimizing(options, log): diff --git a/src/ocrmypdf/_pipeline.py b/src/ocrmypdf/_pipeline.py index 80cd9093..3d63cc47 100644 --- a/src/ocrmypdf/_pipeline.py +++ b/src/ocrmypdf/_pipeline.py @@ -162,6 +162,10 @@ def repair_and_parse_pdf( options = context.get_options() copyfile(input_file, output_file) + detailed_page_analysis = False + if options.redo_ocr: + detailed_page_analysis = True + try: pdfinfo = PdfInfo(output_file, log=log) except pikepdf.PasswordError as e: @@ -193,6 +197,25 @@ def repair_and_parse_pdf( "high page count files. Python 3.6 or newer is recommended." ) + if pdfinfo.has_acroform: + if options.redo_ocr: + log.error( + "This PDF has a user fillable form. --redo-ocr is not " + "currently possible on such files." + ) + raise PriorOcrFoundError() + else: + log.warning( + "This PDF has a fillable form. Chances are it is a pure digital " + "document that does not need OCR." + ) + if not options.force_ocr: + log.info( + "Use the option --force-ocr to produce an image of the " + "form and all filled form fields. The output PDF will be " + "'flattened' and will no longer be fillable." + ) + context.set_pdfinfo(pdfinfo) log.debug(pdfinfo) @@ -243,7 +266,7 @@ def is_ocr_required(pageinfo, log, options): if pageinfo.has_text: msg = "{0:4d}: page already has text! – {1}" - if not options.force_ocr and not options.skip_text: + if not options.force_ocr and not (options.skip_text or options.redo_ocr): log.error(msg.format(page, "aborting (use --force-ocr to force OCR)")) raise PriorOcrFoundError() @@ -251,6 +274,18 @@ def is_ocr_required(pageinfo, log, options): log.info(msg.format(page, "rasterizing text and running OCR anyway")) ocr_required = True + elif options.redo_ocr: + if pageinfo.has_corrupt_text: + log.warning(msg.format( + page, + "some text on this page cannot be mapped to characters: " + "consider using --force-ocr instead") + ) + raise PriorOcrFoundError() # Wrong error but will do for now + else: + log.info(msg.format(page, + "redoing OCR")) + ocr_required = True elif options.skip_text: log.info(msg.format(page, "skipping all processing on this page")) @@ -493,7 +528,8 @@ def rasterize_with_ghostscript( ghostscript.rasterize_pdf( input_file, output_file, xres=canvas_dpi, yres=canvas_dpi, raster_device=device, log=log, page_dpi=(page_dpi, page_dpi), - pageno=page_number(input_file), rotation=correction) + pageno=page_number(input_file), rotation=correction, + filter_vector=options.remove_vectors) def preprocess_remove_background( @@ -559,12 +595,13 @@ def select_ocr_image( user.""" image = infiles[0] - if context.get_options().force_ocr: + options = context.get_options() + pageinfo = get_pageinfo(image, context) + + if options.force_ocr: re_symlink(image, output_file, log) return - pageinfo = get_pageinfo(image, context) - with Image.open(image) as im: from PIL import ImageColor from PIL import ImageDraw @@ -575,22 +612,39 @@ def select_ocr_image( xres, yres = im.info['dpi'] log.debug('resolution %r %r', xres, yres) - for textarea in pageinfo.get_textareas(): + + mask = None # Exclude both visible and invisible text from OCR + if options.redo_ocr: + mask = True # Mask visible text, but not invisible text + + for textarea in pageinfo.get_textareas(visible=mask, corrupt=None): # Calculate resolution based on the image size and page dimensions # without regard whatever resolution is in pageinfo (may differ or # be None) bbox = textarea - pixcoords = [Decimal(bbox[0]) / Decimal(72) * xres, - Decimal(bbox[1]) / Decimal(72) * yres, - Decimal(bbox[2]) / Decimal(72) * xres, - Decimal(bbox[3]) / Decimal(72) * yres] - pixcoords = [int(c) for c in pixcoords] + xscale, yscale = float(xres) / 72.0, float(yres) / 72.0 + pixcoords = [bbox[0] * xscale, + im.height - bbox[3] * yscale, + bbox[2] * xscale, + im.height - bbox[1] * yscale] + pixcoords = [int(round(c)) for c in pixcoords] log.debug('blanking %r', pixcoords) draw.rectangle(pixcoords, fill=white) #draw.rectangle(pixcoords, outline=pink) - del draw + if options.mask_barcodes or options.threshold: + pix = leptonica.Pix.frompil(im) + if options.threshold: + pix = pix.masked_threshold_on_background_norm() + if options.mask_barcodes: + barcodes = pix.locate_barcodes() + for barcode in barcodes: + decoded, rect = barcode + log.info('masking barcode %s %r', decoded, rect) + draw.rectangle(rect, fill=white) + im = pix.topil() + # Pillow requires integer DPI dpi = round(xres), round(yres) im.save(output_file, dpi=dpi) diff --git a/src/ocrmypdf/_unicodefun.py b/src/ocrmypdf/_unicodefun.py index e80d3d21..6834884d 100644 --- a/src/ocrmypdf/_unicodefun.py +++ b/src/ocrmypdf/_unicodefun.py @@ -38,7 +38,7 @@ import sys import codecs -def verify_python3_env(): +def verify_python3_env(): # pragma: no cover """Ensures that the environment is good for unicode on Python 3.""" # PEP 538 changes in Python 3.7 should make this wrangling unnecessary diff --git a/src/ocrmypdf/_weave.py b/src/ocrmypdf/_weave.py index 7d08ffd6..995878b6 100644 --- a/src/ocrmypdf/_weave.py +++ b/src/ocrmypdf/_weave.py @@ -43,8 +43,54 @@ def _update_page_resources(*, page, font, font_key, procset): resources['/ProcSet'] = procset +def strip_invisible_text(pdf, page, log): + stream = [] + in_text_obj = False + render_mode = 0 + text_objects = [] + + page.page_contents_coalesce() + for operands, operator in pikepdf.parse_content_stream(page, ''): + if not in_text_obj: + if operator == pikepdf.Operator('BT'): + in_text_obj = True + render_mode = 0 + text_objects.append((operands, operator)) + else: + stream.append((operands, operator)) + else: + if operator == pikepdf.Operator('Tr'): + render_mode = operands[0] + text_objects.append((operands, operator)) + if operator == pikepdf.Operator('ET'): + in_text_obj = False + if render_mode != 3: + stream.extend(text_objects) + text_objects.clear() + + def convert(op): + try: + return op.unparse() + except AttributeError: + return str(op).encode('ascii') + + lines = [] + + for operands, operator in stream: + if operator == pikepdf.Operator('INLINE IMAGE'): + iim = operands[0] + line = iim.unparse() + else: + line = b' '.join(convert(op) for op in operands) + b' ' + operator.unparse() + lines.append(line) + + content_stream = b'\n'.join(lines) + page.Contents = pikepdf.Stream(pdf, content_stream) + + def _weave_layers_graft( - *, pdf_base, page_num, text, font, font_key, procset, rotation, log): + *, pdf_base, page_num, text, font, font_key, procset, rotation, + strip_old_text, log): """Insert the text layer from text page 0 on to pdf_base at page_num""" log.debug("Grafting") @@ -109,6 +155,9 @@ def _weave_layers_graft( new_text_layer = pikepdf.Stream(pdf_base, pdf_text_contents) + if strip_old_text: + strip_invisible_text(pdf_base, base_page, log) + base_page.page_contents_add(new_text_layer, prepend=True) _update_page_resources( @@ -336,10 +385,11 @@ def weave_layers( if text and font: # Graft the text layer onto this page, whether new or old + strip_old = context.get_options().redo_ocr _weave_layers_graft( pdf_base=pdf_base, page_num=page_num, text=text, font=font, font_key=font_key, rotation=text_misaligned, procset=procset, - log=log + strip_old_text=strip_old, log=log ) # Correct the rotation if applicable diff --git a/src/ocrmypdf/exec/ghostscript.py b/src/ocrmypdf/exec/ghostscript.py index 60a0f3c6..d1abfe96 100644 --- a/src/ocrmypdf/exec/ghostscript.py +++ b/src/ocrmypdf/exec/ghostscript.py @@ -101,7 +101,7 @@ def extract_text(input_file, pageno=1): def rasterize_pdf(input_file, output_file, xres, yres, raster_device, log, - pageno=1, page_dpi=None, rotation=None): + pageno=1, page_dpi=None, rotation=None, filter_vector=False): """Rasterize one page of a PDF at resolution (xres, yres) in canvas units. The image is sized to match the integer pixels dimensions implied by @@ -116,6 +116,8 @@ def rasterize_pdf(input_file, output_file, xres, yres, raster_device, log, :param log: :param pageno: page number to rasterize (beginning at page 1) :param page_dpi: resolution tuple (x, y) overriding output image DPI + :param rotation: 0, 90, 180, 270: clockwise angle to rotate page + :param filter_vector: if True, remove vector graphics objects :return: """ res = xres, yres @@ -134,6 +136,7 @@ def rasterize_pdf(input_file, output_file, xres, yres, raster_device, log, '-dFirstPage=%i' % pageno, '-dLastPage=%i' % pageno, '-r{0}x{1}'.format(str(int_res[0]), str(int_res[1])), + ] + (['-dFILTERVECTOR'] if filter_vector else []) + [ '-o', tmp.name, '-dAutoRotatePages=/None', # Probably has no effect on raster '-f', diff --git a/src/ocrmypdf/exec/qpdf.py b/src/ocrmypdf/exec/qpdf.py index d0c8b9c5..92522491 100644 --- a/src/ocrmypdf/exec/qpdf.py +++ b/src/ocrmypdf/exec/qpdf.py @@ -19,6 +19,7 @@ from subprocess import CalledProcessError, STDOUT, PIPE, run from functools import lru_cache from . import get_version +from ..helpers import fspath @lru_cache(maxsize=1) @@ -30,7 +31,7 @@ def check(input_file, log=None): args_qpdf = [ 'qpdf', '--check', - input_file + fspath(input_file) ] if log is None: diff --git a/src/ocrmypdf/exec/tesseract.py b/src/ocrmypdf/exec/tesseract.py index 1dce012c..13ff838b 100644 --- a/src/ocrmypdf/exec/tesseract.py +++ b/src/ocrmypdf/exec/tesseract.py @@ -21,11 +21,11 @@ import shutil from functools import lru_cache from collections import namedtuple from textwrap import dedent -from subprocess import CalledProcessError, TimeoutExpired, check_output, STDOUT +from subprocess import CalledProcessError, TimeoutExpired, check_output, STDOUT, run, PIPE from contextlib import suppress from ..exceptions import MissingDependencyError, TesseractConfigError -from ..helpers import page_number +from ..helpers import page_number, fspath from . import get_version OrientationConfidence = namedtuple( @@ -92,22 +92,33 @@ def psm(): @lru_cache(maxsize=1) def languages(): + def lang_error(output): + msg = dedent("""Tesseract failed to report available languages. + Output from Tesseract: + ----------- + """) + msg += output + print(msg, file=sys.stderr) + args_tess = [ 'tesseract', '--list-langs' ] try: - langs = check_output( - args_tess, universal_newlines=True, stderr=STDOUT) + proc = run( + args_tess, universal_newlines=True, stdout=PIPE, stderr=STDOUT, + check=True + ) + output = proc.stdout except CalledProcessError as e: - msg = dedent("""Tesseract failed to report available languages. - Output from Tesseract: - ----------- - """) - msg += e.output - print(msg, file=sys.stderr) + lang_error(e.output) raise MissingDependencyError from e - return set(lang.strip() for lang in langs.splitlines()[1:]) + + header, *rest = output.splitlines() + if not header.startswith('List of available languages'): + lang_error(output) + raise MissingDependencyError + return set(lang.strip() for lang in rest) def tess_base_args(langs, engine_mode): @@ -124,7 +135,7 @@ def tess_base_args(langs, engine_mode): def get_orientation(input_file, engine_mode, timeout: float, log): args_tesseract = tess_base_args(['osd'], engine_mode) + [ psm(), '0', - input_file, + fspath(input_file), 'stdout' ] diff --git a/src/ocrmypdf/leptonica.py b/src/ocrmypdf/leptonica.py index 436281d7..644acd70 100644 --- a/src/ocrmypdf/leptonica.py +++ b/src/ocrmypdf/leptonica.py @@ -20,14 +20,16 @@ # # Python FFI wrapper for Leptonica library -import argparse -import sys -import os -import logging -import warnings -from tempfile import TemporaryFile +from collections.abc import Sequence from ctypes.util import find_library from functools import lru_cache +from io import BytesIO +from tempfile import TemporaryFile +import argparse +import logging +import os +import sys +import warnings from .lib._leptonica import ffi from .helpers import fspath @@ -38,6 +40,7 @@ lept = ffi.dlopen(find_library('lept')) logger = logging.getLogger(__name__) +lept.setMsgSeverity(lept.L_SEVERITY_WARNING) def stderr(*objs): """Shorthand print to stderr.""" @@ -103,6 +106,8 @@ class _LeptonicaErrorTrap: raise FileNotFoundError() if 'pixWrite: stream not opened' in leptonica_output: raise LeptonicaIOError() + if 'index not valid' in leptonica_output: + raise IndexError() raise LeptonicaError(leptonica_output) return False @@ -116,7 +121,43 @@ class LeptonicaIOError(LeptonicaError): pass -class Pix: +class LeptonicaObject: + """General wrapper for Leptonica objects + + When Leptonica returns an object, we bundled it in a wrapper class, which + manages its memory. The wrapper class assumes that it will be calling some + sort of lept.thingDestroy() function when the instance is deleted. Most + Leptonica objects are reference counted, and destroy decrements the + refcount. + + Most of the time, when Leptonica returns something, we wrap and it the job + is done. When wrapping objects that came from a Leptonica container, like + a PIXA returning PIX, the subclass must clone the object before passing it + here, to maintain the reference count. + + CFFI ensures that the destroy function is called at garbage collection time + so we do not need to mess with __del__. + """ + + cdata_destroy = lambda cdata: None + LEPTONICA_TYPENAME = '' + + def __init__(self, cdata): + if not cdata: + raise ValueError('Tried to wrap a NULL ' + self.LEPTONICA_TYPENAME) + self._cdata = ffi.gc(cdata, self._destroy) + + @classmethod + def _destroy(cls, cdata): + """Destroy some cdata""" + # Leptonica API uses double-pointers for its destroy APIs to prevent + # dangling pointers. This means we need to put our single pointer, + # cdata, in a temporary CDATA**. + pp = ffi.new('{} **'.format(cls.LEPTONICA_TYPENAME), cdata) + cls.cdata_destroy(pp) + + +class Pix(LeptonicaObject): """ Wrapper around leptonica's PIX object. @@ -135,18 +176,17 @@ class Pix: modified objects. This allows convenient chaining: >>> Pix.open('filename.jpg').scale((0.5, 0.5)).deskew().show() - """ - def __init__(self, pix): - self._pix = ffi.gc(pix, Pix._pix_destroy) + LEPTONICA_TYPENAME = "PIX" + cdata_destroy = lept.pixDestroy def __repr__(self): - if self._pix: + if self._cdata: s = "" - return s.format(self._pix.w, self._pix.h, self._pix.d, - int(ffi.cast('intptr_t', self._pix)), - '(colormapped)' if self._pix.colormap else '') + return s.format(self._cdata.w, self._cdata.h, self._cdata.d, + int(ffi.cast('intptr_t', self._cdata)), + '(colormapped)' if self._cdata.colormap else '') else: return "" @@ -159,7 +199,7 @@ class Pix: data = ffi.new('l_uint8 **') size = ffi.new('size_t *') - err = lept.pixWriteMemPng(data, size, self._pix, 0) + err = lept.pixWriteMemPng(data, size, self._cdata, 0) if err != 0: raise LeptonicaIOError("pixWriteMemPng") @@ -170,7 +210,7 @@ class Pix: data = ffi.new('l_uint32 **') size = ffi.new('size_t *') - err = lept.pixSerializeToMemory(self._pix, data, size) + err = lept.pixSerializeToMemory(self._cdata, data, size) if err != 0: raise LeptonicaIOError("pixSerializeToMemory") @@ -187,32 +227,38 @@ class Pix: cdata_bytes = ffi.new('char[]', state['data']) cdata_uint32 = ffi.cast('l_uint32 *', cdata_bytes) - pix = lept.pixDeserializeFromMemory( - cdata_uint32, len(state['data'])) + pix = lept.pixDeserializeFromMemory(cdata_uint32, len(state['data'])) Pix.__init__(self, pix) def __eq__(self, other): - return self.__getstate__() == other.__getstate__() + if not isinstance(other, Pix): + return NotImplemented + same = ffi.new('l_int32 *', 0) + with _LeptonicaErrorTrap(): + err = lept.pixEqual(self._cdata, other._cdata, same) + if err: + raise TypeError() + return bool(same[0]) @property def width(self): - return self._pix.w + return self._cdata.w @property def height(self): - return self._pix.h + return self._cdata.h @property def depth(self): - return self._pix.d + return self._cdata.d @property def size(self): - return (self._pix.w, self._pix.h) + return (self._cdata.w, self._cdata.h) @property def info(self): - return {'dpi': (self._pix.xres, self._pix.yres)} + return {'dpi': (self._cdata.xres, self._cdata.yres)} @property def mode(self): @@ -221,7 +267,7 @@ class Pix: return '1' elif self.depth >= 16: return 'RGB' - elif not self._pix.colormap: + elif not self._cdata.colormap: return 'L' else: return 'P' @@ -253,10 +299,21 @@ class Pix: with _LeptonicaErrorTrap(): lept.pixWriteImpliedFormat( os.fsencode(filename), - self._pix, jpeg_quality, jpeg_progressive) + self._cdata, jpeg_quality, jpeg_progressive) + + @classmethod + def frompil(self, pillow_image): + """Create a copy of a PIL.Image from this Pix""" + bio = BytesIO() + pillow_image.save(bio, format='png', compress_level=1) + py_buffer = bio.getbuffer() + c_buffer = ffi.from_buffer(py_buffer) + with _LeptonicaErrorTrap(): + pix = Pix(lept.pixReadMem(c_buffer, len(c_buffer))) + return pix def topil(self): - "Returns a PIL.Image version of this Pix" + """Returns a PIL.Image version of this Pix""" from PIL import Image # Leptonica manages data in words, so it implicitly does an endian @@ -269,17 +326,17 @@ class Pix: raw_mode = 'ABGR' elif self.mode == '1': raw_mode = '1;I' - pix = Pix(lept.pixEndianByteSwapNew(pix._pix)) + pix = Pix(lept.pixEndianByteSwapNew(pix._cdata)) else: raw_mode = self.mode - pix = Pix(lept.pixEndianByteSwapNew(pix._pix)) + pix = Pix(lept.pixEndianByteSwapNew(pix._cdata)) else: raw_mode = self.mode # no endian swap needed - size = (pix._pix.w, pix._pix.h) - bytecount = pix._pix.wpl * 4 * pix._pix.h - buf = ffi.buffer(pix._pix.data, bytecount) - stride = pix._pix.wpl * 4 + size = (pix._cdata.w, pix._cdata.h) + bytecount = pix._cdata.wpl * 4 * pix._cdata.h + buf = ffi.buffer(pix._cdata.data, bytecount) + stride = pix._cdata.wpl * 4 im = Image.frombytes(self.mode, size, buf, 'raw', raw_mode, stride) @@ -298,21 +355,21 @@ class Pix: for skew angle """ with _LeptonicaErrorTrap(): - return Pix(lept.pixDeskew(self._pix, reduction_factor)) + return Pix(lept.pixDeskew(self._cdata, reduction_factor)) def scale(self, scale_xy): "Returns the pix object rescaled according to the proportions given." with _LeptonicaErrorTrap(): - return Pix(lept.pixScale(self._pix, scale_xy[0], scale_xy[1])) + return Pix(lept.pixScale(self._cdata, scale_xy[0], scale_xy[1])) def rotate180(self): with _LeptonicaErrorTrap(): - return Pix(lept.pixRotate180(ffi.NULL, self._pix)) + return Pix(lept.pixRotate180(ffi.NULL, self._cdata)) def rotate_orth(self, quads): "Orthographic rotation, quads: 0-3, number of clockwise rotations" with _LeptonicaErrorTrap(): - return Pix(lept.pixRotateOrth(self._pix, quads)) + return Pix(lept.pixRotateOrth(self._cdata, quads)) def find_skew(self): """Returns a tuple (deskew angle in degrees, confidence value). @@ -322,7 +379,7 @@ class Pix: with _LeptonicaErrorTrap(): angle = ffi.new('float *', 0.0) confidence = ffi.new('float *', 0.0) - result = lept.pixFindSkew(self._pix, angle, confidence) + result = lept.pixFindSkew(self._cdata, angle, confidence) if result == 0: return (angle[0], confidence[0]) else: @@ -330,7 +387,7 @@ class Pix: def convert_rgb_to_luminance(self): with _LeptonicaErrorTrap(): - gray_pix = lept.pixConvertRGBToLuminance(self._pix) + gray_pix = lept.pixConvertRGBToLuminance(self._cdata) if gray_pix: return Pix(gray_pix) return None @@ -344,7 +401,7 @@ class Pix: """ with _LeptonicaErrorTrap(): return Pix(lept.pixRemoveColormapGeneral( - self._pix, removal_type, lept.L_COPY)) + self._cdata, removal_type, lept.L_COPY)) def otsu_adaptive_threshold( self, tile_size=(300, 300), kernel_size=(4, 4), scorefract=0.1): @@ -354,7 +411,7 @@ class Pix: p_pix = ffi.new('PIX **') result = lept.pixOtsuAdaptiveThreshold( - self._pix, + self._cdata, sx, sy, smoothx, smoothy, scorefract, @@ -371,22 +428,41 @@ class Pix: with _LeptonicaErrorTrap(): sx, sy = tile_size smoothx, smoothy = kernel_size - if mask is None: - mask = ffi.NULL + mask = ffi.NULL if isinstance(mask, Pix): - mask = mask._pix + mask = mask._cdata thresh_pix = lept.pixOtsuThreshOnBackgroundNorm( - self._pix, + self._cdata, mask, sx, sy, thresh, mincount, bgval, smoothx, smoothy, scorefract, ffi.NULL - ) - if thresh_pix == ffi.NULL: - return None + ) + return Pix(thresh_pix) + + def masked_threshold_on_background_norm( + self, mask=None, tile_size=(10, 15), thresh=100, mincount=50, + kernel_size=(2, 2), scorefract=0.1): + with _LeptonicaErrorTrap(): + sx, sy = tile_size + smoothx, smoothy = kernel_size + mask = ffi.NULL + if isinstance(mask, Pix): + mask = mask._cdata + + pix = Pix(lept.pixConvertTo8(self._cdata, 0)) + thresh_pix = lept.pixMaskedThreshOnBackgroundNorm( + pix._cdata, + mask, + sx, sy, + thresh, mincount, + smoothx, smoothy, + scorefract, + ffi.NULL + ) return Pix(thresh_pix) def crop_to_foreground( @@ -394,7 +470,7 @@ class Pix: showmorph=0, display=0, pdfdir=ffi.NULL): with _LeptonicaErrorTrap(): cropbox = Box(lept.pixFindPageForeground( - self._pix, + self._cdata, threshold, mindist, erasedist, @@ -403,11 +479,9 @@ class Pix: display, pdfdir)) - print(repr(cropbox)) - cropped_pix = lept.pixClipRectangle( - self._pix, - cropbox._box, + self._cdata, + cropbox._cdata, ffi.NULL) return Pix(cropped_pix) @@ -416,7 +490,7 @@ class Pix: self, mask=None, grayscale=None, gamma=1.0, black=0, white=255): with _LeptonicaErrorTrap(): return Pix(lept.pixCleanBackgroundToWhite( - self._pix, + self._cdata, mask or ffi.NULL, grayscale or ffi.NULL, gamma, @@ -427,7 +501,7 @@ class Pix: with _LeptonicaErrorTrap(): return Pix(lept.pixGammaTRC( ffi.NULL, - self._pix, + self._cdata, gamma, minval, maxval @@ -440,7 +514,7 @@ class Pix: target_pix = self.remove_colormap(lept.REMOVE_CMAP_BASED_ON_SRC) with _LeptonicaErrorTrap(): return Pix(lept.pixBackgroundNorm( - target_pix._pix, + target_pix._cdata, mask or ffi.NULL, grayscale or ffi.NULL, tile_size[0], @@ -467,7 +541,7 @@ class Pix: raise LeptonicaError("Leptonica version is too old") correlation = ffi.new('float *', 0.0) - result = lept.pixCorrelationBinary(pix1._pix, pix2._pix, + result = lept.pixCorrelationBinary(pix1._cdata, pix2._cdata, correlation) if result != 0: raise LeptonicaError("Correlation failed") @@ -476,25 +550,68 @@ class Pix: def generate_pdf_ci_data(self, type_, quality): "Convert to PDF data, with transcoding" p_compdata = ffi.new('L_COMP_DATA **') - result = lept.pixGenerateCIData(self._pix, type_, quality, 0, + result = lept.pixGenerateCIData(self._cdata, type_, quality, 0, p_compdata) if result != 0: raise LeptonicaError("Generate PDF data failed") return CompressedData(p_compdata[0]) def invert(self): - return Pix(lept.pixInvert(ffi.NULL, self._pix)) + return Pix(lept.pixInvert(ffi.NULL, self._cdata)) - @staticmethod - def _pix_destroy(pix): - p_pix = ffi.new('PIX **', pix) - lept.pixDestroy(p_pix) - # print('pix destroy ' + repr(pix)) + def locate_barcodes(self): + with _LeptonicaErrorTrap(): + pix = Pix(lept.pixConvertTo8(self._cdata, 0)) + pixa_candidates = PixArray(lept.pixExtractBarcodes(pix._cdata, 0)) + sarray = StringArray(lept.pixReadBarcodes(pixa_candidates._cdata, + lept.L_BF_ANY, + lept.L_USE_WIDTHS, + ffi.NULL, + 0)) + for n, s in enumerate(sarray): + decoded = s.decode() + if s.strip() == '': + continue + box = pixa_candidates.get_box(n) + left, top = box.x, box.y + right, bottom = box.x + box.w, box.y + box.h + yield (decoded, (left, top, right, bottom)) + + def despeckle(self, size): + if size == 2: + speckle2 = """ + oooo + oC o + o o + oooo + """ + sel1 = Sel.from_selstr(speckle2, 'speckle2') + sel2 = Sel.create_brick(2, 2, 0, 0, lept.SEL_HIT) + elif size == 3: + speckle3 = """ + ooooo + oC o + o o + o o + ooooo + """ + sel1 = Sel.from_selstr(speckle3, 'speckle3') + sel2 = Sel.create_brick(3, 3, 0, 0, lept.SEL_HIT) + else: + raise ValueError(size) + + pixhmt = Pix(lept.pixHMT(ffi.NULL, self._cdata, sel1._cdata)) + pixdilated = Pix(lept.pixDilate(ffi.NULL, pixhmt._cdata, sel2._cdata)) + + pixsub = Pix(lept.pixSubtract(ffi.NULL, self._cdata, pixdilated._cdata)) + return pixsub -class CompressedData: - def __init__(self, compdata): - self._compdata = ffi.gc(compdata, CompressedData._destroy) +class CompressedData(LeptonicaObject): + """Wrapper for L_COMP_DATA - abstract compressed image data""" + + LEPTONICA_TYPENAME = 'L_COMP_DATA' + cdata_destroy = lept.l_CIDataDestroy @classmethod def open(cls, path, jpeg_quality=75): @@ -509,64 +626,141 @@ class CompressedData: return CompressedData(p_compdata[0]) def __len__(self): - return self._compdata.nbytescomp + return self._cdata.nbytescomp def read(self): - buf = ffi.buffer(self._compdata.datacomp, self._compdata.nbytescomp) + buf = ffi.buffer(self._cdata.datacomp, self._cdata.nbytescomp) return bytes(buf) def __getattr__(self, name): - if hasattr(self._compdata, name): - return getattr(self._compdata, name) + if hasattr(self._cdata, name): + return getattr(self._cdata, name) raise AttributeError(name) def get_palette_pdf_string(self): "Returns palette pre-formatted for use in PDF" - buflen = len('< ') + len(' rrggbb') * self._compdata.ncolors + len('>') - buf = ffi.buffer(self._compdata.cmapdatahex, buflen) + buflen = len('< ') + len(' rrggbb') * self._cdata.ncolors + len('>') + buf = ffi.buffer(self._cdata.cmapdatahex, buflen) return bytes(buf) - @staticmethod - def _destroy(compdata): - pp = ffi.new('L_COMP_DATA **', compdata) - lept.l_CIDataDestroy(pp) + +class PixArray(LeptonicaObject, Sequence): + """Wrapper around PIXA (array of PIX)""" + + LEPTONICA_TYPENAME = 'PIXA' + cdata_destroy = lept.pixaDestroy + + def __len__(self): + return self._cdata[0].n + + def __getitem__(self, n): + with _LeptonicaErrorTrap(): + return Pix(lept.pixaGetPix(self._cdata, n, lept.L_CLONE)) + + def get_box(self, n): + with _LeptonicaErrorTrap(): + return Box(lept.pixaGetBox(self._cdata, n, lept.L_CLONE)) -class Box: - """Wrapper around Leptonica's BOX objects. +class Box(LeptonicaObject): + """Wrapper around Leptonica's BOX objects (a pixel rectangle) - See class Pix for notes about reference counting. + Uses x, y, w, h coordinates. """ - def __init__(self, box): - self._box = ffi.gc(box, Box._box_destroy) + LEPTONICA_TYPENAME = 'BOX' + cdata_destroy = lept.boxDestroy def __repr__(self): - if self._box: + if self._cdata: return ''.format( self.x, self.y, self.w, self.h) return '' @property def x(self): - return self._box.x + return self._cdata.x @property def y(self): - return self._box.y + return self._cdata.y @property def w(self): - return self._box.w + return self._cdata.w @property def h(self): - return self._box.h + return self._cdata.h - @staticmethod - def _box_destroy(box): - p_box = ffi.new('BOX **', box) - lept.boxDestroy(p_box) + +class BoxArray(LeptonicaObject, Sequence): + """Wrapper around Leptonica's BOXA (Array of BOX) objects.""" + + LEPTONICA_TYPENAME = 'BOXA' + cdata_destroy = lept.boxaDestroy + + def __repr__(self): + if not self._cdata: + return '' + boxes = (repr(box) for box in self) + return '' + + def __len__(self): + return self._cdata.n + + def __getitem__(self, n): + if not isinstance(n, int): + raise TypeError('list indices must be integers') + if 0 <= n < len(self): + return Box(lept.boxaGetBox(self._cdata, n, lept.L_CLONE)) + raise IndexError(n) + + +class StringArray(LeptonicaObject, Sequence): + """Leptonica SARRAY/string array""" + + LEPTONICA_TYPENAME = 'SARRAY' + cdata_destroy = lept.sarrayDestroy + + def __len__(self): + return self._cdata.n + + def __getitem__(self, n): + if 0 <= n < len(self): + return ffi.string(self._cdata.array[n]) + raise IndexError(n) + + +class Sel(LeptonicaObject): + """Leptonica 'sel'/selection element for hit-miss transform""" + + LEPTONICA_TYPENAME = 'SEL' + cdata_destroy = lept.selDestroy + + @classmethod + def from_selstr(cls, selstr, name): + lines = [line.strip() for line in selstr.split('\n') if line.strip()] + h = len(lines) + w = len(lines[0]) + lengths = set(len(line) for line in lines) + if len(lengths) != 1: + raise ValueError("All lines in selstr must be same length") + + repacked = ''.join(line.strip() for line in lines) + buf_selstr = ffi.from_buffer(repacked.encode('ascii')) + buf_name = ffi.from_buffer(name.encode('ascii')) + sel = lept.selCreateFromString(buf_selstr, h, w, buf_name) + return cls(sel) + + @classmethod + def create_brick(cls, h, w, cy, cx, type_): + sel = lept.selCreateBrick(h, w, cy, cx, type_) + return cls(sel) + + def __repr__(self): + selstr = ffi.gc(lept.selPrintToString(self._cdata), lept.lept_free) + return '' @lru_cache(maxsize=1) diff --git a/src/ocrmypdf/lib/_leptonica.py b/src/ocrmypdf/lib/_leptonica.py index b2611d5b..17a2f757 100644 --- a/src/ocrmypdf/lib/_leptonica.py +++ b/src/ocrmypdf/lib/_leptonica.py @@ -3,9 +3,9 @@ import _cffi_backend ffi = _cffi_backend.FFI('ocrmypdf.lib._leptonica', _version = 0x2601, - _types = b'\x00\x00\x0F\x0D\x00\x00\xCD\x03\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\xD0\x03\x00\x00\x00\x0F\x00\x00\x01\x0D\x00\x00\x01\x11\x00\x00\x00\x0F\x00\x00\x01\x0D\x00\x00\x01\x11\x00\x00\xCA\x03\x00\x00\x0F\x03\x00\x00\x00\x0F\x00\x00\x01\x0D\x00\x00\x01\x11\x00\x00\x01\x11\x00\x00\x00\x0F\x00\x00\x01\x0D\x00\x00\x01\x11\x00\x00\x01\x11\x00\x00\x01\x11\x00\x00\x00\x0F\x00\x00\x01\x0D\x00\x00\x01\x11\x00\x00\x01\x11\x00\x00\x01\x11\x00\x00\x0D\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x00\x0F\x00\x00\x01\x0D\x00\x00\x01\x11\x00\x00\x01\x11\x00\x00\x01\x11\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x00\x0F\x00\x00\x01\x0D\x00\x00\x01\x11\x00\x00\x01\x11\x00\x00\x0D\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x00\x0F\x00\x00\x01\x0D\x00\x00\x01\x11\x00\x00\x01\x11\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x00\x0F\x00\x00\x01\x0D\x00\x00\x01\x11\x00\x00\x01\x11\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x0D\x01\x00\x00\x02\x03\x00\x00\x00\x0F\x00\x00\x01\x0D\x00\x00\x01\x11\x00\x00\x0D\x01\x00\x00\x0D\x01\x00\x00\x00\x0F\x00\x00\x01\x0D\x00\x00\x01\x11\x00\x00\x07\x01\x00\x00\x00\x0F\x00\x00\x01\x0D\x00\x00\x01\x11\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x00\x0F\x00\x00\x01\x0D\x00\x00\x01\x11\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x00\x0F\x00\x00\x01\x0D\x00\x00\x08\x11\x00\x00\x00\x0F\x00\x00\x01\x0D\x00\x00\xDC\x03\x00\x00\x1C\x01\x00\x00\x00\x0F\x00\x00\xCF\x0D\x00\x00\x00\x0F\x00\x00\x49\x0D\x00\x00\x00\x0F\x00\x00\x02\x0D\x00\x00\x01\x11\x00\x00\x01\x11\x00\x00\x1F\x03\x00\x00\x00\x0F\x00\x00\x02\x0D\x00\x00\x01\x11\x00\x00\x01\x11\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x6E\x11\x00\x00\x6E\x11\x00\x00\x6E\x11\x00\x00\x00\x0F\x00\x00\x02\x0D\x00\x00\x01\x11\x00\x00\x6E\x11\x00\x00\x6E\x11\x00\x00\x00\x0F\x00\x00\x02\x0D\x00\x00\x01\x11\x00\x00\x49\x11\x00\x00\x49\x11\x00\x00\x00\x0F\x00\x00\x02\x0D\x00\x00\x01\x11\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x0D\x01\x00\x00\x07\x01\x00\x00\x49\x11\x00\x00\x00\x0F\x00\x00\x02\x0D\x00\x00\x01\x11\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\xCB\x03\x00\x00\x00\x0F\x00\x00\x02\x0D\x00\x00\x01\x11\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x6E\x11\x00\x00\x6E\x11\x00\x00\x00\x0F\x00\x00\x02\x0D\x00\x00\x01\x11\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x0D\x01\x00\x00\x01\x03\x00\x00\xA4\x11\x00\x00\x00\x0F\x00\x00\x02\x0D\x00\x00\x01\x11\x00\x00\xDB\x03\x00\x00\x65\x03\x00\x00\x00\x0F\x00\x00\x02\x0D\x00\x00\x08\x11\x00\x00\x01\x11\x00\x00\x07\x01\x00\x00\x92\x11\x00\x00\x00\x0F\x00\x00\x02\x0D\x00\x00\x08\x11\x00\x00\x01\x11\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x00\x0F\x00\x00\x02\x0D\x00\x00\xD9\x03\x00\x00\xAA\x11\x00\x00\x01\x11\x00\x00\x0D\x01\x00\x00\x00\x0F\x00\x00\xDF\x0D\x00\x00\x10\x11\x00\x00\x00\x0F\x00\x00\xDF\x0D\x00\x00\x92\x11\x00\x00\x00\x0F\x00\x00\xDF\x0D\x00\x00\xA4\x11\x00\x00\x00\x0F\x00\x00\xDF\x0D\x00\x00\xDF\x03\x00\x00\x00\x0F\x00\x00\x00\x09\x00\x00\xCC\x03\x00\x00\x01\x09\x00\x00\x02\x09\x00\x00\x03\x09\x00\x00\xD0\x03\x00\x00\x02\x01\x00\x00\x0E\x01\x00\x00\x00\x0B\x00\x00\x01\x0B\x00\x00\x02\x0B\x00\x00\x0B\x01\x00\x00\x05\x01\x00\x00\x03\x01\x00\x00\xCE\x03\x00\x00\xDA\x03\x00\x00\x04\x01\x00\x00\xDC\x03\x00\x00\x08\x01\x00\x00\x0C\x01\x00\x00\x06\x01\x00\x00\x00\x01', - _globals = (b'\xFF\xFF\xFF\x0BL_CLONE',2,b'\xFF\xFF\xFF\x0BL_COPY',1,b'\xFF\xFF\xFF\x0BL_COPY_CLONE',3,b'\xFF\xFF\xFF\x0BL_DEFAULT_ENCODE',0,b'\xFF\xFF\xFF\x0BL_FLATE_ENCODE',3,b'\xFF\xFF\xFF\x0BL_G4_ENCODE',2,b'\xFF\xFF\xFF\x0BL_INSERT',0,b'\xFF\xFF\xFF\x0BL_JP2K_ENCODE',4,b'\xFF\xFF\xFF\x0BL_JPEG_ENCODE',1,b'\xFF\xFF\xFF\x0BL_NOCOPY',0,b'\xFF\xFF\xFF\x0BREMOVE_CMAP_BASED_ON_SRC',4,b'\xFF\xFF\xFF\x0BREMOVE_CMAP_TO_BINARY',0,b'\xFF\xFF\xFF\x0BREMOVE_CMAP_TO_FULL_COLOR',2,b'\xFF\xFF\xFF\x0BREMOVE_CMAP_TO_GRAYSCALE',1,b'\xFF\xFF\xFF\x0BREMOVE_CMAP_WITH_ALPHA',3,b'\x00\x00\xBE\x23boxDestroy',0,b'\x00\x00\x67\x23getLeptonicaVersion',0,b'\x00\x00\xC1\x23l_CIDataDestroy',0,b'\x00\x00\xAC\x23l_generateCIDataForPdf',0,b'\x00\x00\xC7\x23lept_free',0,b'\x00\x00\x69\x23makePixelSumTab8',0,b'\x00\x00\x16\x23pixAnd',0,b'\x00\x00\x23\x23pixBackgroundNorm',0,b'\x00\x00\x1B\x23pixCleanBackgroundToWhite',0,b'\x00\x00\x0D\x23pixClipRectangle',0,b'\x00\x00\x94\x23pixColorFraction',0,b'\x00\x00\x59\x23pixColorMagnitude',0,b'\x00\x00\x0A\x23pixConvertRGBToLuminance',0,b'\x00\x00\x6B\x23pixCorrelationBinary',0,b'\x00\x00\x80\x23pixCountPixels',0,b'\x00\x00\x63\x23pixDeserializeFromMemory',0,b'\x00\x00\x50\x23pixDeskew',0,b'\x00\x00\xC4\x23pixDestroy',0,b'\x00\x00\x0A\x23pixEndianByteSwapNew',0,b'\x00\x00\x00\x23pixFindPageForeground',0,b'\x00\x00\x7B\x23pixFindSkew',0,b'\x00\x00\x2F\x23pixGammaTRC',0,b'\x00\x00\x8D\x23pixGenerateCIData',0,b'\x00\x00\x70\x23pixGetAverageMaskedRGB',0,b'\x00\x00\x36\x23pixGlobalNormRGB',0,b'\x00\x00\x12\x23pixInvert',0,b'\x00\x00\x54\x23pixMaskOverColorPixels',0,b'\x00\x00\x85\x23pixNumSignificantGrayColors',0,b'\x00\x00\x9D\x23pixOtsuAdaptiveThreshold',0,b'\x00\x00\x3E\x23pixOtsuThreshOnBackgroundNorm',0,b'\x00\x00\x60\x23pixRead',0,b'\x00\x00\x50\x23pixRemoveColormap',0,b'\x00\x00\x54\x23pixRemoveColormapGeneral',0,b'\x00\x00\x12\x23pixRotate180',0,b'\x00\x00\x50\x23pixRotateOrth',0,b'\x00\x00\x4B\x23pixScale',0,b'\x00\x00\xA7\x23pixSerializeToMemory',0,b'\x00\x00\xB2\x23pixWriteImpliedFormat',0,b'\x00\x00\xB8\x23pixWriteMemPng',0), - _struct_unions = ((b'\x00\x00\x00\xCA\x00\x00\x00\x02Box',b'\x00\x00\x02\x11x',b'\x00\x00\x02\x11y',b'\x00\x00\x02\x11w',b'\x00\x00\x02\x11h',b'\x00\x00\xDC\x11refcount'),(b'\x00\x00\x00\xCC\x00\x00\x00\x02L_Compressed_Data',b'\x00\x00\x02\x11type',b'\x00\x00\xD9\x11datacomp',b'\x00\x00\x65\x11nbytescomp',b'\x00\x00\xCF\x11data85',b'\x00\x00\x65\x11nbytes85',b'\x00\x00\xCF\x11cmapdata85',b'\x00\x00\xCF\x11cmapdatahex',b'\x00\x00\x02\x11ncolors',b'\x00\x00\x02\x11w',b'\x00\x00\x02\x11h',b'\x00\x00\x02\x11bps',b'\x00\x00\x02\x11spp',b'\x00\x00\x02\x11minisblack',b'\x00\x00\x02\x11predictor',b'\x00\x00\x65\x11nbytes',b'\x00\x00\x02\x11res'),(b'\x00\x00\x00\xCD\x00\x00\x00\x02Pix',b'\x00\x00\xDC\x11w',b'\x00\x00\xDC\x11h',b'\x00\x00\xDC\x11d',b'\x00\x00\xDC\x11spp',b'\x00\x00\xDC\x11wpl',b'\x00\x00\xDC\x11refcount',b'\x00\x00\x02\x11xres',b'\x00\x00\x02\x11yres',b'\x00\x00\x02\x11informat',b'\x00\x00\x02\x11special',b'\x00\x00\xCF\x11text',b'\x00\x00\xD8\x11colormap',b'\x00\x00\xDB\x11data'),(b'\x00\x00\x00\xCE\x00\x00\x00\x02PixColormap',b'\x00\x00\xC8\x11array',b'\x00\x00\x02\x11depth',b'\x00\x00\x02\x11nalloc',b'\x00\x00\x02\x11n')), - _enums = (b'\x00\x00\x00\xD2\x00\x00\x00\x16$1\x00L_DEFAULT_ENCODE,L_JPEG_ENCODE,L_G4_ENCODE,L_FLATE_ENCODE,L_JP2K_ENCODE',b'\x00\x00\x00\xD3\x00\x00\x00\x16$2\x00REMOVE_CMAP_TO_BINARY,REMOVE_CMAP_TO_GRAYSCALE,REMOVE_CMAP_TO_FULL_COLOR,REMOVE_CMAP_WITH_ALPHA,REMOVE_CMAP_BASED_ON_SRC',b'\x00\x00\x00\xD4\x00\x00\x00\x16$3\x00L_NOCOPY,L_INSERT,L_COPY,L_CLONE,L_COPY_CLONE'), - _typenames = (b'\x00\x00\x00\xCABOX',b'\x00\x00\x00\xCCL_COMP_DATA',b'\x00\x00\x00\xCDPIX',b'\x00\x00\x00\xCEPIXCMAP',b'\x00\x00\x00\x1Fl_float32',b'\x00\x00\x00\xD1l_float64',b'\x00\x00\x00\xD6l_int16',b'\x00\x00\x00\x02l_int32',b'\x00\x00\x00\xD5l_int64',b'\x00\x00\x00\xD7l_int8',b'\x00\x00\x00\xDEl_uint16',b'\x00\x00\x00\xDCl_uint32',b'\x00\x00\x00\xDDl_uint64',b'\x00\x00\x00\xDAl_uint8'), + _types = b'\x00\x00\x01\x0D\x00\x01\x33\x03\x00\x00\x00\x0F\x00\x00\x01\x0D\x00\x01\x34\x03\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x00\x0F\x00\x00\x01\x0D\x00\x01\x37\x03\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x01\x3F\x03\x00\x00\x00\x0F\x00\x00\x01\x0D\x00\x01\x38\x03\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x00\x0F\x00\x00\x04\x0D\x00\x00\x09\x11\x00\x00\x07\x01\x00\x00\x09\x03\x00\x00\x1A\x11\x00\x00\x00\x0F\x00\x00\x09\x0D\x00\x00\x09\x11\x00\x00\x00\x0F\x00\x00\x09\x0D\x00\x00\x09\x11\x00\x00\x01\x11\x00\x00\x01\x03\x00\x00\x00\x0F\x00\x00\x09\x0D\x00\x00\x09\x11\x00\x00\x09\x11\x00\x00\x00\x0F\x00\x00\x09\x0D\x00\x00\x09\x11\x00\x00\x09\x11\x00\x00\x09\x11\x00\x00\x00\x0F\x00\x00\x09\x0D\x00\x00\x09\x11\x00\x00\x09\x11\x00\x00\x09\x11\x00\x00\x0D\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x00\x0F\x00\x00\x09\x0D\x00\x00\x09\x11\x00\x00\x09\x11\x00\x00\x09\x11\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x00\x0F\x00\x00\x09\x0D\x00\x00\x09\x11\x00\x00\x09\x11\x00\x01\x3C\x03\x00\x00\x00\x0F\x00\x00\x09\x0D\x00\x00\x09\x11\x00\x00\x09\x11\x00\x00\x0D\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x00\x0F\x00\x00\x09\x0D\x00\x00\x09\x11\x00\x00\x09\x11\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x00\x0F\x00\x00\x09\x0D\x00\x00\x09\x11\x00\x00\x09\x11\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x0D\x01\x00\x00\x05\x03\x00\x00\x00\x0F\x00\x00\x09\x0D\x00\x00\x09\x11\x00\x00\x09\x11\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x0D\x01\x00\x00\x60\x11\x00\x00\x00\x0F\x00\x00\x09\x0D\x00\x00\x09\x11\x00\x00\x0D\x01\x00\x00\x0D\x01\x00\x00\x00\x0F\x00\x00\x09\x0D\x00\x00\x09\x11\x00\x00\x07\x01\x00\x00\x00\x0F\x00\x00\x09\x0D\x00\x00\x09\x11\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x00\x0F\x00\x00\x09\x0D\x00\x00\x09\x11\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x00\x0F\x00\x00\x09\x0D\x00\x00\x13\x11\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x00\x0F\x00\x00\x09\x0D\x00\x00\x10\x11\x00\x00\x00\x0F\x00\x00\x09\x0D\x00\x01\x4E\x03\x00\x00\x1C\x01\x00\x00\x00\x0F\x00\x00\x09\x0D\x00\x01\x50\x03\x00\x00\x1C\x01\x00\x00\x00\x0F\x00\x00\x13\x0D\x00\x00\x09\x11\x00\x00\x07\x01\x00\x00\x00\x0F\x00\x01\x3A\x0D\x00\x00\x09\x11\x00\x00\x07\x01\x00\x00\x07\x01\x00\x01\x3A\x03\x00\x00\x07\x01\x00\x00\x00\x0F\x00\x01\x3A\x0D\x00\x00\x13\x11\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x9C\x11\x00\x00\x07\x01\x00\x00\x00\x0F\x00\x00\x45\x0D\x00\x00\x10\x11\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x10\x11\x00\x00\x00\x0F\x00\x00\x45\x0D\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x00\x0F\x00\x01\x3E\x0D\x00\x00\x45\x11\x00\x00\x00\x0F\x00\x01\x3E\x0D\x00\x00\x00\x0F\x00\x00\x60\x0D\x00\x00\x00\x0F\x00\x00\x05\x0D\x00\x00\x09\x11\x00\x00\x04\x11\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x00\x0F\x00\x00\x05\x0D\x00\x00\x09\x11\x00\x00\x09\x11\x00\x00\x32\x03\x00\x00\x00\x0F\x00\x00\x05\x0D\x00\x00\x09\x11\x00\x00\x09\x11\x00\x00\x60\x11\x00\x00\x00\x0F\x00\x00\x05\x0D\x00\x00\x09\x11\x00\x00\x09\x11\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\xC3\x11\x00\x00\xC3\x11\x00\x00\xC3\x11\x00\x00\x00\x0F\x00\x00\x05\x0D\x00\x00\x09\x11\x00\x00\xC3\x11\x00\x00\xC3\x11\x00\x00\x00\x0F\x00\x00\x05\x0D\x00\x00\x09\x11\x00\x00\x60\x11\x00\x00\x60\x11\x00\x00\x00\x0F\x00\x00\x05\x0D\x00\x00\x09\x11\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x0D\x01\x00\x00\x07\x01\x00\x00\x60\x11\x00\x00\x00\x0F\x00\x00\x05\x0D\x00\x00\x09\x11\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x01\x35\x03\x00\x00\x00\x0F\x00\x00\x05\x0D\x00\x00\x09\x11\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\xC3\x11\x00\x00\xC3\x11\x00\x00\x00\x0F\x00\x00\x05\x0D\x00\x00\x09\x11\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x0D\x01\x00\x00\x1A\x11\x00\x00\x1A\x11\x00\x00\x00\x0F\x00\x00\x05\x0D\x00\x00\x09\x11\x00\x01\x4F\x03\x00\x00\x8E\x03\x00\x00\x00\x0F\x00\x00\x05\x0D\x00\x00\x10\x11\x00\x00\x09\x11\x00\x00\x07\x01\x00\x00\xEC\x11\x00\x00\x00\x0F\x00\x00\x05\x0D\x00\x00\x10\x11\x00\x00\x09\x11\x00\x00\x07\x01\x00\x00\x07\x01\x00\x00\x00\x0F\x00\x00\x05\x0D\x00\x00\x07\x01\x00\x00\x00\x0F\x00\x00\x05\x0D\x00\x01\x4D\x03\x00\x01\x04\x11\x00\x00\x09\x11\x00\x00\x0D\x01\x00\x00\x00\x0F\x00\x01\x53\x0D\x00\x00\x23\x11\x00\x00\x00\x0F\x00\x01\x53\x0D\x00\x00\x04\x03\x00\x00\x00\x0F\x00\x01\x53\x0D\x00\x00\xEC\x11\x00\x00\x00\x0F\x00\x01\x53\x0D\x00\x00\x1A\x11\x00\x00\x00\x0F\x00\x01\x53\x0D\x00\x00\x13\x03\x00\x00\x00\x0F\x00\x01\x53\x0D\x00\x00\x9C\x11\x00\x00\x00\x0F\x00\x01\x53\x0D\x00\x00\x45\x03\x00\x00\x00\x0F\x00\x01\x53\x0D\x00\x01\x53\x03\x00\x00\x00\x0F\x00\x00\x00\x09\x00\x00\x01\x09\x00\x01\x36\x03\x00\x00\x02\x09\x00\x00\x03\x09\x00\x00\x05\x09\x00\x00\x04\x09\x00\x01\x3B\x03\x00\x00\x06\x09\x00\x00\x07\x09\x00\x01\x3E\x03\x00\x01\x3F\x03\x00\x00\x02\x01\x00\x00\x0E\x01\x00\x00\x00\x0B\x00\x00\x01\x0B\x00\x00\x02\x0B\x00\x00\x03\x0B\x00\x00\x04\x0B\x00\x00\x05\x0B\x00\x00\x06\x0B\x00\x00\x60\x03\x00\x00\x0B\x01\x00\x00\x05\x01\x00\x00\x03\x01\x00\x01\x39\x03\x00\x01\x4E\x03\x00\x00\x04\x01\x00\x01\x50\x03\x00\x00\x08\x01\x00\x00\x0C\x01\x00\x00\x06\x01\x00\x00\x00\x01', + _globals = (b'\xFF\xFF\xFF\x0BL_BF_ANY',1,b'\xFF\xFF\xFF\x0BL_BF_CODABAR',9,b'\xFF\xFF\xFF\x0BL_BF_CODE128',2,b'\xFF\xFF\xFF\x0BL_BF_CODE2OF5',5,b'\xFF\xFF\xFF\x0BL_BF_CODE39',7,b'\xFF\xFF\xFF\x0BL_BF_CODE93',8,b'\xFF\xFF\xFF\x0BL_BF_CODEI2OF5',6,b'\xFF\xFF\xFF\x0BL_BF_EAN13',4,b'\xFF\xFF\xFF\x0BL_BF_EAN8',3,b'\xFF\xFF\xFF\x0BL_BF_UNKNOWN',0,b'\xFF\xFF\xFF\x0BL_BF_UPCA',10,b'\xFF\xFF\xFF\x0BL_CLONE',2,b'\xFF\xFF\xFF\x0BL_COPY',1,b'\xFF\xFF\xFF\x0BL_COPY_CLONE',3,b'\xFF\xFF\xFF\x0BL_DEFAULT_ENCODE',0,b'\xFF\xFF\xFF\x0BL_FLATE_ENCODE',3,b'\xFF\xFF\xFF\x0BL_G4_ENCODE',2,b'\xFF\xFF\xFF\x0BL_INSERT',0,b'\xFF\xFF\xFF\x0BL_JP2K_ENCODE',4,b'\xFF\xFF\xFF\x0BL_JPEG_ENCODE',1,b'\xFF\xFF\xFF\x0BL_NOCOPY',0,b'\xFF\xFF\xFF\x0BL_SEVERITY_ALL',1,b'\xFF\xFF\xFF\x0BL_SEVERITY_DEBUG',2,b'\xFF\xFF\xFF\x0BL_SEVERITY_ERROR',5,b'\xFF\xFF\xFF\x0BL_SEVERITY_EXTERNAL',0,b'\xFF\xFF\xFF\x0BL_SEVERITY_INFO',3,b'\xFF\xFF\xFF\x0BL_SEVERITY_NONE',6,b'\xFF\xFF\xFF\x0BL_SEVERITY_WARNING',4,b'\xFF\xFF\xFF\x0BL_USE_WIDTHS',1,b'\xFF\xFF\xFF\x0BL_USE_WINDOWS',2,b'\xFF\xFF\xFF\x0BREMOVE_CMAP_BASED_ON_SRC',4,b'\xFF\xFF\xFF\x0BREMOVE_CMAP_TO_BINARY',0,b'\xFF\xFF\xFF\x0BREMOVE_CMAP_TO_FULL_COLOR',2,b'\xFF\xFF\xFF\x0BREMOVE_CMAP_TO_GRAYSCALE',1,b'\xFF\xFF\xFF\x0BREMOVE_CMAP_WITH_ALPHA',3,b'\xFF\xFF\xFF\x0BSEL_DONT_CARE',0,b'\xFF\xFF\xFF\x0BSEL_HIT',1,b'\xFF\xFF\xFF\x0BSEL_MISS',2,b'\x00\x00\x00\x23boxClone',0,b'\x00\x01\x1B\x23boxDestroy',0,b'\x00\x01\x1E\x23boxaDestroy',0,b'\x00\x00\x03\x23boxaGetBox',0,b'\x00\x00\xB6\x23getLeptonicaVersion',0,b'\x00\x01\x21\x23l_CIDataDestroy',0,b'\x00\x01\x06\x23l_generateCIDataForPdf',0,b'\x00\x01\x30\x23lept_free',0,b'\x00\x00\xB8\x23makePixelSumTab8',0,b'\x00\x00\x29\x23pixAnd',0,b'\x00\x00\x36\x23pixBackgroundNorm',0,b'\x00\x00\x2E\x23pixCleanBackgroundToWhite',0,b'\x00\x00\x20\x23pixClipRectangle',0,b'\x00\x00\xEE\x23pixColorFraction',0,b'\x00\x00\x7D\x23pixColorMagnitude',0,b'\x00\x00\x1D\x23pixConvertRGBToLuminance',0,b'\x00\x00\x74\x23pixConvertTo8',0,b'\x00\x00\xC0\x23pixCorrelationBinary',0,b'\x00\x00\xDA\x23pixCountPixels',0,b'\x00\x00\x90\x23pixDeserializeFromMemory',0,b'\x00\x00\x74\x23pixDeskew',0,b'\x00\x01\x24\x23pixDestroy',0,b'\x00\x00\x42\x23pixDilate',0,b'\x00\x00\x1D\x23pixEndianByteSwapNew',0,b'\x00\x00\xC5\x23pixEqual',0,b'\x00\x00\x42\x23pixErode',0,b'\x00\x00\x94\x23pixExtractBarcodes',0,b'\x00\x00\x08\x23pixFindPageForeground',0,b'\x00\x00\xD5\x23pixFindSkew',0,b'\x00\x00\x47\x23pixGammaTRC',0,b'\x00\x00\xE7\x23pixGenerateCIData',0,b'\x00\x00\xCA\x23pixGetAverageMaskedRGB',0,b'\x00\x00\x4E\x23pixGlobalNormRGB',0,b'\x00\x00\x42\x23pixHMT',0,b'\x00\x00\x25\x23pixInvert',0,b'\x00\x00\x17\x23pixLocateBarcodes',0,b'\x00\x00\x78\x23pixMaskOverColorPixels',0,b'\x00\x00\x56\x23pixMaskedThreshOnBackgroundNorm',0,b'\x00\x00\xDF\x23pixNumSignificantGrayColors',0,b'\x00\x00\xF7\x23pixOtsuAdaptiveThreshold',0,b'\x00\x00\x62\x23pixOtsuThreshOnBackgroundNorm',0,b'\x00\x00\x98\x23pixProcessBarcodes',0,b'\x00\x00\x89\x23pixRead',0,b'\x00\x00\x9F\x23pixReadBarcodes',0,b'\x00\x00\x8C\x23pixReadMem',0,b'\x00\x00\x74\x23pixRemoveColormap',0,b'\x00\x00\x78\x23pixRemoveColormapGeneral',0,b'\x00\x00\xBA\x23pixRenderBoxa',0,b'\x00\x00\x25\x23pixRotate180',0,b'\x00\x00\x74\x23pixRotateOrth',0,b'\x00\x00\x6F\x23pixScale',0,b'\x00\x01\x01\x23pixSerializeToMemory',0,b'\x00\x00\x29\x23pixSubtract',0,b'\x00\x01\x0C\x23pixWriteImpliedFormat',0,b'\x00\x01\x15\x23pixWriteMemPng',0,b'\x00\x01\x27\x23pixaDestroy',0,b'\x00\x00\x12\x23pixaGetBox',0,b'\x00\x00\x84\x23pixaGetPix',0,b'\x00\x01\x2A\x23sarrayDestroy',0,b'\x00\x00\xAC\x23selCreateBrick',0,b'\x00\x00\xA6\x23selCreateFromString',0,b'\x00\x01\x2D\x23selDestroy',0,b'\x00\x00\xB3\x23selPrintToString',0,b'\x00\x01\x12\x23setMsgSeverity',0), + _struct_unions = ((b'\x00\x00\x01\x33\x00\x00\x00\x02Box',b'\x00\x00\x05\x11x',b'\x00\x00\x05\x11y',b'\x00\x00\x05\x11w',b'\x00\x00\x05\x11h',b'\x00\x01\x50\x11refcount'),(b'\x00\x00\x01\x34\x00\x00\x00\x02Boxa',b'\x00\x00\x05\x11n',b'\x00\x00\x05\x11nalloc',b'\x00\x01\x50\x11refcount',b'\x00\x00\x23\x11box'),(b'\x00\x00\x01\x36\x00\x00\x00\x02L_Compressed_Data',b'\x00\x00\x05\x11type',b'\x00\x01\x4D\x11datacomp',b'\x00\x00\x8E\x11nbytescomp',b'\x00\x01\x3E\x11data85',b'\x00\x00\x8E\x11nbytes85',b'\x00\x01\x3E\x11cmapdata85',b'\x00\x01\x3E\x11cmapdatahex',b'\x00\x00\x05\x11ncolors',b'\x00\x00\x05\x11w',b'\x00\x00\x05\x11h',b'\x00\x00\x05\x11bps',b'\x00\x00\x05\x11spp',b'\x00\x00\x05\x11minisblack',b'\x00\x00\x05\x11predictor',b'\x00\x00\x8E\x11nbytes',b'\x00\x00\x05\x11res'),(b'\x00\x00\x01\x37\x00\x00\x00\x02Pix',b'\x00\x01\x50\x11w',b'\x00\x01\x50\x11h',b'\x00\x01\x50\x11d',b'\x00\x01\x50\x11spp',b'\x00\x01\x50\x11wpl',b'\x00\x01\x50\x11refcount',b'\x00\x00\x05\x11xres',b'\x00\x00\x05\x11yres',b'\x00\x00\x05\x11informat',b'\x00\x00\x05\x11special',b'\x00\x01\x3E\x11text',b'\x00\x01\x4C\x11colormap',b'\x00\x01\x4F\x11data'),(b'\x00\x00\x01\x39\x00\x00\x00\x02PixColormap',b'\x00\x01\x31\x11array',b'\x00\x00\x05\x11depth',b'\x00\x00\x05\x11nalloc',b'\x00\x00\x05\x11n'),(b'\x00\x00\x01\x38\x00\x00\x00\x02Pixa',b'\x00\x00\x05\x11n',b'\x00\x00\x05\x11nalloc',b'\x00\x01\x50\x11refcount',b'\x00\x00\x1A\x11pix',b'\x00\x00\x04\x11boxa'),(b'\x00\x00\x01\x3B\x00\x00\x00\x02Sarray',b'\x00\x00\x05\x11nalloc',b'\x00\x00\x05\x11n',b'\x00\x00\x05\x11refcount',b'\x00\x01\x3D\x11array'),(b'\x00\x00\x01\x3C\x00\x00\x00\x02Sel',b'\x00\x00\x05\x11sy',b'\x00\x00\x05\x11sx',b'\x00\x00\x05\x11cy',b'\x00\x00\x05\x11cx',b'\x00\x01\x48\x11data',b'\x00\x01\x3E\x11name')), + _enums = (b'\x00\x00\x01\x41\x00\x00\x00\x16$1\x00L_DEFAULT_ENCODE,L_JPEG_ENCODE,L_G4_ENCODE,L_FLATE_ENCODE,L_JP2K_ENCODE',b'\x00\x00\x01\x42\x00\x00\x00\x16$2\x00REMOVE_CMAP_TO_BINARY,REMOVE_CMAP_TO_GRAYSCALE,REMOVE_CMAP_TO_FULL_COLOR,REMOVE_CMAP_WITH_ALPHA,REMOVE_CMAP_BASED_ON_SRC',b'\x00\x00\x01\x43\x00\x00\x00\x16$3\x00L_NOCOPY,L_INSERT,L_COPY,L_CLONE,L_COPY_CLONE',b'\x00\x00\x01\x44\x00\x00\x00\x16$4\x00L_USE_WIDTHS,L_USE_WINDOWS',b'\x00\x00\x01\x45\x00\x00\x00\x16$5\x00L_BF_UNKNOWN,L_BF_ANY,L_BF_CODE128,L_BF_EAN8,L_BF_EAN13,L_BF_CODE2OF5,L_BF_CODEI2OF5,L_BF_CODE39,L_BF_CODE93,L_BF_CODABAR,L_BF_UPCA',b'\x00\x00\x01\x46\x00\x00\x00\x16$6\x00L_SEVERITY_EXTERNAL,L_SEVERITY_ALL,L_SEVERITY_DEBUG,L_SEVERITY_INFO,L_SEVERITY_WARNING,L_SEVERITY_ERROR,L_SEVERITY_NONE',b'\x00\x00\x01\x47\x00\x00\x00\x16$7\x00SEL_DONT_CARE,SEL_HIT,SEL_MISS'), + _typenames = (b'\x00\x00\x01\x33BOX',b'\x00\x00\x01\x34BOXA',b'\x00\x00\x01\x36L_COMP_DATA',b'\x00\x00\x01\x37PIX',b'\x00\x00\x01\x38PIXA',b'\x00\x00\x01\x39PIXCMAP',b'\x00\x00\x01\x3BSARRAY',b'\x00\x00\x01\x3CSEL',b'\x00\x00\x00\x32l_float32',b'\x00\x00\x01\x40l_float64',b'\x00\x00\x01\x4Al_int16',b'\x00\x00\x00\x05l_int32',b'\x00\x00\x01\x49l_int64',b'\x00\x00\x01\x4Bl_int8',b'\x00\x00\x00\x05l_ok',b'\x00\x00\x01\x52l_uint16',b'\x00\x00\x01\x50l_uint32',b'\x00\x00\x01\x51l_uint64',b'\x00\x00\x01\x4El_uint8'), ) diff --git a/src/ocrmypdf/lib/compile_leptonica.py b/src/ocrmypdf/lib/compile_leptonica.py index 5f739f28..4418e7cf 100644 --- a/src/ocrmypdf/lib/compile_leptonica.py +++ b/src/ocrmypdf/lib/compile_leptonica.py @@ -19,7 +19,6 @@ from cffi import FFI ffibuilder = FFI() -ffibuilder.set_source("ocrmypdf.lib._leptonica", None) ffibuilder.cdef(""" typedef signed char l_int8; typedef unsigned char l_uint8; @@ -32,6 +31,8 @@ typedef double l_float64; typedef long long l_int64; typedef unsigned long long l_uint64; +typedef int l_ok; /*!< return type 0 if OK, 1 on error */ + struct Pix { l_uint32 w; /* width in pixels */ @@ -61,6 +62,17 @@ struct PixColormap }; typedef struct PixColormap PIXCMAP; +/*! Array of pix */ +struct Pixa +{ + l_int32 n; /*!< number of Pix in ptr array */ + l_int32 nalloc; /*!< number of Pix ptrs allocated */ + l_uint32 refcount; /*!< reference count (1 if no clones) */ + struct Pix **pix; /*!< the array of ptrs to pix */ + struct Boxa *boxa; /*!< array of boxes */ +}; +typedef struct Pixa PIXA; + struct Box { l_int32 x; @@ -72,6 +84,25 @@ struct Box }; typedef struct Box BOX; +/*! Array of Box */ +struct Boxa +{ + l_int32 n; /*!< number of box in ptr array */ + l_int32 nalloc; /*!< number of box ptrs allocated */ + l_uint32 refcount; /*!< reference count (1 if no clones) */ + struct Box **box; /*!< box ptr array */ +}; +typedef struct Boxa BOXA; + +/*! String array: an array of C strings */ +struct Sarray +{ + l_int32 nalloc; /*!< size of allocated ptr array */ + l_int32 n; /*!< number of strings allocated */ + l_int32 refcount; /*!< reference count (1 if no clones) */ + char **array; /*!< string array */ +}; +typedef struct Sarray SARRAY; /*! Pdf formatted encoding types */ enum { @@ -104,6 +135,18 @@ struct L_Compressed_Data }; typedef struct L_Compressed_Data L_COMP_DATA; +/*! Selection */ +struct Sel +{ + l_int32 sy; /*!< sel height */ + l_int32 sx; /*!< sel width */ + l_int32 cy; /*!< y location of sel origin */ + l_int32 cx; /*!< x location of sel origin */ + l_int32 **data; /*!< {0,1,2}; data[i][j] in [row][col] order */ + char *name; /*!< used to find sel by name */ +}; +typedef struct Sel SEL; + enum { REMOVE_CMAP_TO_BINARY = 0, /*!< remove colormap for conv to 1 bpp */ REMOVE_CMAP_TO_GRAYSCALE = 1, /*!< remove colormap for conv to 8 bpp */ @@ -122,10 +165,48 @@ enum { /*!< the array with clones (e.g., pix) */ }; +/*! Flags for method of extracting barcode widths */ +enum { + L_USE_WIDTHS = 1, /*!< use histogram of barcode widths */ + L_USE_WINDOWS = 2 /*!< find best window for decoding transitions */ +}; + +/*! Flags for barcode formats */ +enum { + L_BF_UNKNOWN = 0, /*!< unknown format */ + L_BF_ANY = 1, /*!< try decoding with all known formats */ + L_BF_CODE128 = 2, /*!< decode with Code128 format */ + L_BF_EAN8 = 3, /*!< decode with EAN8 format */ + L_BF_EAN13 = 4, /*!< decode with EAN13 format */ + L_BF_CODE2OF5 = 5, /*!< decode with Code 2 of 5 format */ + L_BF_CODEI2OF5 = 6, /*!< decode with Interleaved 2 of 5 format */ + L_BF_CODE39 = 7, /*!< decode with Code39 format */ + L_BF_CODE93 = 8, /*!< decode with Code93 format */ + L_BF_CODABAR = 9, /*!< decode with Code93 format */ + L_BF_UPCA = 10 /*!< decode with UPC A format */ +}; + +enum { + L_SEVERITY_EXTERNAL = 0, /* Get the severity from the environment */ + L_SEVERITY_ALL = 1, /* Lowest severity: print all messages */ + L_SEVERITY_DEBUG = 2, /* Print debugging and higher messages */ + L_SEVERITY_INFO = 3, /* Print informational and higher messages */ + L_SEVERITY_WARNING = 4, /* Print warning and higher messages */ + L_SEVERITY_ERROR = 5, /* Print error and higher messages */ + L_SEVERITY_NONE = 6 /* Highest severity: print no messages */ +}; + +enum { + SEL_DONT_CARE = 0, + SEL_HIT = 1, + SEL_MISS = 2 +}; + """) ffibuilder.cdef(""" PIX * pixRead ( const char *filename ); +PIX * pixReadMem ( const l_uint8 *data, size_t size ); PIX * pixScale ( PIX *pixs, l_float32 scalex, l_float32 scaley ); l_int32 pixFindSkew ( PIX *pixs, l_float32 *pangle, l_float32 *pconf ); l_int32 pixWriteImpliedFormat ( const char *filename, PIX *pix, l_int32 quality, l_int32 progressive ); @@ -137,6 +218,11 @@ pixWriteMemPng(l_uint8 **pdata, void pixDestroy ( PIX **ppix ); +l_ok +pixEqual(PIX *pix1, + PIX *pix2, + l_int32 *psame); + PIX * pixEndianByteSwapNew(PIX *pixs); @@ -157,6 +243,8 @@ l_int32 pixSerializeToMemory ( PIX *pixs, l_uint32 **pdata, size_t *pnbytes ); PIX * pixConvertRGBToLuminance(PIX *pixs); +PIX * pixConvertTo8(PIX *pixs, l_int32 cmapflag); + PIX * pixRemoveColormap(PIX *pixs, l_int32 type); l_int32 @@ -182,6 +270,18 @@ pixOtsuThreshOnBackgroundNorm(PIX *pixs, l_float32 scorefract, l_int32 *pthresh); +PIX * +pixMaskedThreshOnBackgroundNorm(PIX *pixs, + PIX *pixim, + l_int32 sx, + l_int32 sy, + l_int32 thresh, + l_int32 mincount, + l_int32 smoothx, + l_int32 smoothy, + l_float32 scorefract, + l_int32 *pthresh); + PIX * pixCleanBackgroundToWhite(PIX *pixs, PIX *pixim, @@ -265,13 +365,13 @@ pixGetAverageMaskedRGB(PIX *pixs, l_float32 *pgval, l_float32 *pbval); -PIX * +PIX * pixGlobalNormRGB(PIX * pixd, PIX * pixs, l_int32 rval, l_int32 gval, l_int32 bval, - l_int32 mapval); + l_int32 mapval); PIX * pixInvert(PIX * pixd, @@ -289,22 +389,101 @@ pixGenerateCIData(PIX *pixs, l_int32 ascii85, L_COMP_DATA **pcid); -l_int32 -l_generateCIDataForPdf(const char *fname, - PIX *pix, - l_int32 quality, +SARRAY * +pixProcessBarcodes(PIX *pixs, + l_int32 format, + l_int32 method, + SARRAY **psaw, + l_int32 debugflag); + +PIX * +pixaGetPix(PIXA *pixa, + l_int32 index, + l_int32 accesstype); + +BOX* +pixaGetBox (PIXA * pixa, + l_int32 index, + l_int32 accesstype ); + +PIXA * +pixExtractBarcodes(PIX *pixs, + l_int32 debugflag); + +BOXA * +pixLocateBarcodes ( PIX *pixs, l_int32 thresh, PIX **ppixb, PIX **ppixm ); + +SARRAY * +pixReadBarcodes(PIXA *pixa, + l_int32 format, + l_int32 method, + SARRAY **psaw, + l_int32 debugflag); + +l_int32 +l_generateCIDataForPdf(const char *fname, + PIX *pix, + l_int32 quality, L_COMP_DATA **pcid); -void + +BOX * +boxClone ( BOX *box ); + +BOX * +boxaGetBox ( BOXA *boxa, l_int32 index, l_int32 accessflag ); + +SEL * +selCreateFromString ( const char *text, l_int32 h, l_int32 w, const char *name ); + +SEL * +selCreateBrick ( l_int32 h, l_int32 w, l_int32 cy, l_int32 cx, l_int32 type ); + +char * +selPrintToString(SEL *sel); + +PIX * +pixDilate ( PIX *pixd, PIX *pixs, SEL *sel ); + +PIX * +pixErode ( PIX *pixd, PIX *pixs, SEL *sel ); + +PIX * +pixHMT ( PIX *pixd, PIX *pixs, SEL *sel ); + +PIX * +pixSubtract ( PIX *pixd, PIX *pixs1, PIX *pixs2 ); + +void boxDestroy(BOX **pbox); +void +boxaDestroy ( BOXA **pboxa ); + +void +pixaDestroy(PIXA **ppixa); + +l_ok +pixRenderBoxa ( PIX *pix, BOXA *boxa, l_int32 width, l_int32 op ); + void l_CIDataDestroy(L_COMP_DATA **pcid); void -lept_free(void *ptr); +sarrayDestroy(SARRAY **psa); + +void +lept_free(void *ptr); + +void selDestroy ( SEL **psel ); + +l_int32 +setMsgSeverity(l_int32 newsev); + """) +ffibuilder.set_source("ocrmypdf.lib._leptonica", None) + if __name__ == '__main__': ffibuilder.compile(verbose=True) diff --git a/src/ocrmypdf/pdfinfo.py b/src/ocrmypdf/pdfinfo/__init__.py similarity index 83% rename from src/ocrmypdf/pdfinfo.py rename to src/ocrmypdf/pdfinfo/__init__.py index 53279c33..0b783259 100644 --- a/src/ocrmypdf/pdfinfo.py +++ b/src/ocrmypdf/pdfinfo/__init__.py @@ -23,14 +23,16 @@ from math import hypot, isclose from pathlib import Path from unittest.mock import Mock import re -import xml.etree.ElementTree as ET - -from .exec import ghostscript -from .helpers import fspath from pikepdf import PdfMatrix import pikepdf +from . import ghosttext +from .layout import get_page_analysis, get_text_boxes + +from ..helpers import fspath + + Colorspace = Enum('Colorspace', 'gray rgb cmyk lab icc index sep devn pattern jpeg2000') @@ -38,17 +40,6 @@ Encoding = Enum('Encoding', 'ccitt jpeg jpeg2000 jbig2 asciihex ascii85 lzw flate ' + \ 'runlength') -# Forgive me for I have sinned -# I am using regular expressions to parse XML. However the XML in this case, -# generated by Ghostscript, is self-consistent enough to be parseable. -regex_remove_char_tags = re.compile(br""" - ] # anything single character but > - | \">\" # special case: trap ">" - )* - /> # terminate with '/>' -""", re.VERBOSE) - FRIENDLY_COLORSPACE = { '/DeviceGray': Colorspace.gray, '/CalGray': Colorspace.gray, @@ -104,7 +95,15 @@ InlineSettings = namedtuple('InlineSettings', ['iimage', 'shorthand', 'stack_depth']) ContentsInfo = namedtuple('ContentsInfo', - ['xobject_settings', 'inline_images', 'found_text', 'found_vector']) + ['xobject_settings', 'inline_images', 'found_vector']) + +TextboxInfo = namedtuple('TextboxInfo', + ['bbox', 'is_visible', 'is_corrupt']) + + +class VectorInfo: + def __init__(self): + pass def _normalize_stack(graphobjs): @@ -142,10 +141,10 @@ def _interpret_contents(contentstream, initial_shorthand=UNIT_SQUARE): ctm = PdfMatrix(initial_shorthand) xobject_settings = [] inline_images = [] - found_text, found_vector = False, False - text_operators = set("""Tj " ' TJ""".split()) - vector_operators = set('S s f F f* B B* b b*'.split()) - operator_whitelist = """q Q Do cm TJ Tj " ' BI ID EI S s f F f* B B* b b*""" + found_vector = False + vector_ops = set('S s f F f* B B* b b*'.split()) + image_ops = set('BI ID EI q Q Do cm'.split()) + operator_whitelist = ' '.join(vector_ops | image_ops) for n, graphobj in enumerate(_normalize_stack( pikepdf.parse_content_stream(contentstream, operator_whitelist))): @@ -176,15 +175,12 @@ def _interpret_contents(contentstream, initial_shorthand=UNIT_SQUARE): iimage=iimage, shorthand=ctm.shorthand, stack_depth=len(stack)) inline_images.append(inline) - elif operator in text_operators: - found_text = True - elif operator in vector_operators: + elif operator in vector_ops: found_vector = True return ContentsInfo( xobject_settings=xobject_settings, inline_images=inline_images, - found_text=found_text, found_vector=found_vector) @@ -252,11 +248,6 @@ def _get_dpi(ctm_shorthand, image_size): return dpi_w, dpi_h -class VectorInfo: - def __init__(self): - pass - - class ImageInfo: DPI_PREC = Decimal('1.000') @@ -394,7 +385,6 @@ def _find_regular_images(container, contentsinfo): that contains images. Generates images with their DPI at time of drawing. - """ for pdfimage, xobj in _image_xobjects(container): @@ -444,11 +434,11 @@ def _find_form_xobject_images(pdf, container, contentsinfo): # but in practice both Form XObjects and multiple drawing of the # same object are both very rare. ctm_shorthand = settings.shorthand - yield from _find_images( + yield from _process_content_streams( pdf=pdf, container=form_xobject, shorthand=ctm_shorthand) -def _find_images(*, pdf, container, shorthand=None): +def _process_content_streams(*, pdf, container, shorthand=None): """Find all individual instances of images drawn in the container Usually the container is a page, but it may also be a Form XObject. @@ -496,43 +486,6 @@ def _find_images(*, pdf, container, shorthand=None): yield from _find_form_xobject_images(pdf, container, contentsinfo) -def _page_get_textblocks(infile, pageno, xmltext): - """Smarter text detection""" - - root = xmltext - if not hasattr(xmltext, 'findall'): - return [] - - def blocks(): - for span in root.findall('.//span'): - bbox_str = span.attrib['bbox'] - font_size = span.attrib['size'] - pts = [int(pt) for pt in bbox_str.split()] - pts[1] = pts[1] - int(float(font_size) + 0.5) - bbox = tuple(pts) - yield bbox - - def joined_blocks(): - prev = None - for bbox in blocks(): - if prev is None: - prev = bbox - if bbox[1] == prev[1] and bbox[3] == prev[3]: - gap = prev[2] - bbox[0] - height = bbox[3] - bbox[1] - if gap < height: - # Join boxes - prev = (prev[0], prev[1], bbox[2], bbox[3]) - continue - # yield previously joined bboxes and start anew - yield prev - prev = bbox - if prev is not None: - yield prev - - return [block for block in joined_blocks()] - - def _page_has_text(text_blocks, page_width, page_height): """Smarter text detection that ignores text in margins""" @@ -540,18 +493,19 @@ def _page_has_text(text_blocks, page_width, page_height): margin_ratio = 0.125 interior_bbox = ( - margin_ratio * pw, margin_ratio * ph, - (1 - margin_ratio) * pw, (1 - margin_ratio) * ph + margin_ratio * pw, # left + (1 - margin_ratio) * ph, # top + (1 - margin_ratio) * pw, # right + margin_ratio * ph # bottom (first quadrant: bottom < top) ) def rects_intersect(a, b): """ Where (a,b) are 4-tuple rects (left-0, top-1, right-2, bottom-3) https://stackoverflow.com/questions/306316/determine-if-two-rectangles-overlap-each-other - Negative signs to account for our coordinates being in the fourth quadrant - and the formula assuming the first + Formula assumes all boxes are in first quadrant """ - return a[0] < b[2] and a[2] > b[0] and -a[1] > -b[3] and -a[3] < -b[1] + return a[0] < b[2] and a[2] > b[0] and a[1] > b[3] and a[3] < b[1] has_text = False for bbox in text_blocks: @@ -561,22 +515,43 @@ def _page_has_text(text_blocks, page_width, page_height): return has_text +def simplify_textboxes(miner): + """Extract only limited content from text boxes + + We do this to save memory and ensure that our objects are pickleable. + """ + for box in get_text_boxes(miner): + first_line = box._objs[0] + first_char = first_line._objs[0] + + visible = (first_char.rendermode != 3) + corrupt = (first_char.get_text() == '\ufffd') + yield TextboxInfo(box.bbox, visible, corrupt) + + def _pdf_get_pageinfo(pdf, pageno: int, infile, xmltext): pageinfo = {} pageinfo['pageno'] = pageno pageinfo['images'] = [] page = pdf.pages[pageno] - - pageinfo['textinfo'] = _page_get_textblocks( - fspath(infile), pageno, xmltext=xmltext) - mediabox = [Decimal(d) for d in page.MediaBox.as_list()] width_pt = mediabox[2] - mediabox[0] height_pt = mediabox[3] - mediabox[1] + if xmltext: + bboxes = ghosttext.page_get_textblocks( + fspath(infile), pageno, xmltext=xmltext, height=height_pt) + pageinfo['bboxes'] = bboxes + else: + pscript5_mode = str(pdf.metadata.get('/Creator')).startswith('PScript5') + miner = get_page_analysis(infile, pageno, pscript5_mode) + pageinfo['textboxes'] = list(simplify_textboxes(miner)) + bboxes = (box.bbox for box in pageinfo['textboxes']) + pageinfo['has_text'] = _page_has_text( - pageinfo['textinfo'], width_pt, height_pt) + bboxes, width_pt, height_pt + ) userunit = page.get('/UserUnit', Decimal(1.0)) if not isinstance(userunit, Decimal): @@ -591,16 +566,16 @@ def _pdf_get_pageinfo(pdf, pageno: int, infile, xmltext): pageinfo['rotate'] = 0 userunit_shorthand = (userunit, 0, 0, userunit, 0, 0) - pageinfo['images'] = [im for im in - _find_images(pdf=pdf, container=page, - shorthand=userunit_shorthand)] + contentsinfo = [ci for ci in + _process_content_streams(pdf=pdf, container=page, + shorthand=userunit_shorthand)] pageinfo['has_vector'] = False - if any(isinstance(im, VectorInfo) for im in pageinfo['images']): + if any(isinstance(ci, VectorInfo) for ci in contentsinfo): pageinfo['has_vector'] = True - pageinfo['images'] = [im for im in pageinfo['images'] - if not isinstance(im, VectorInfo)] + pageinfo['images'] = [im for im in contentsinfo + if isinstance(im, ImageInfo)] if pageinfo['images']: xres = Decimal(max(image.xres for image in pageinfo['images'])) yres = Decimal(max(image.yres for image in pageinfo['images'])) @@ -613,37 +588,20 @@ def _pdf_get_pageinfo(pdf, pageno: int, infile, xmltext): return pageinfo -def _pdf_get_all_pageinfo(infile, log=None): +def _pdf_get_all_pageinfo(infile, detailed_page_analysis, log=None): if not log: log = Mock() pdf = pikepdf.open(infile) - - existing_text = ghostscript.extract_text(infile, pageno=None) - existing_text = regex_remove_char_tags.sub(b' ', existing_text) - - try: - root = ET.fromstringlist([ - b'\n', existing_text, b'\n' - ]) - page_xml = root.findall('page') - except ET.ParseError as e: - log.error( - "An error occurred while attempting to retrieve existing text in " - "the input file. Will attempt to continue assuming that there is " - "no existing text in the file. The error was:") - log.error(e) - page_xml = [None] * len(pdf.pages) - - page_count_difference = len(pdf.pages) - len(page_xml) - if page_count_difference != 0: - log.error("The number of pages in the input file is inconsistent.") - if page_count_difference > 0: - page_xml.extend([None] * page_count_difference) + if not detailed_page_analysis: + pages_xml = None + else: + pages_xml = ghosttext.extract_text_xml(infile, pdf, pageno=None, log=log) pages = [] for n in range(len(pdf.pages)): - page = PageInfo(pdf, n, infile, page_xml[n]) + page_xml = pages_xml[n] if pages_xml else None + page = PageInfo(pdf, n, infile, page_xml) pages.append(page) return pages, pdf @@ -663,6 +621,10 @@ class PageInfo: def has_text(self): return self._pageinfo['has_text'] + @property + def has_corrupt_text(self): + return any(tbox.is_corrupt for tbox in self._pageinfo['textboxes']) + @property def has_vector(self): return self._pageinfo['has_vector'] @@ -698,8 +660,26 @@ class PageInfo: def images(self): return self._pageinfo['images'] - def get_textareas(self): - yield from self._pageinfo['textinfo'] + def get_textareas(self, visible=None, corrupt=None): + def predicate(obj, want_visible, want_corrupt): + result = True + if want_visible is not None: + if obj.is_visible != want_visible: + result = False + if want_corrupt is not None: + if obj.is_corrupt != want_corrupt: + result = False + return result + + if 'textboxes' not in self._pageinfo: + if visible is not None and corrupt is not None: + raise NotImplementedError( + 'Ghostscript textboxes cannot be classified') + return self._pageinfo['bboxes'] + + return (obj.bbox for obj in self._pageinfo['textboxes'] + if predicate(obj, visible, corrupt)) + @property def xres(self): @@ -733,10 +713,12 @@ class PageInfo: class PdfInfo: """Get summary information about a PDF""" - def __init__(self, infile, log=None): + def __init__(self, infile, detailed_page_analysis=False, log=None): self._infile = infile - self._pages, pdf = _pdf_get_all_pageinfo(infile, log=log) + self._pages, pdf = _pdf_get_all_pageinfo( + infile, detailed_page_analysis, log=log) self._needs_rendering = pdf.root.get('/NeedsRendering', False) + self._has_acroform = '/AcroForm' in pdf.root @property def pages(self): @@ -751,6 +733,10 @@ class PdfInfo: def has_userunit(self): return any(page.userunit != 1.0 for page in self.pages) + @property + def has_acroform(self): + return self._has_acroform + @property def filename(self): if not isinstance(self._infile, (str, Path)): diff --git a/src/ocrmypdf/pdfinfo/ghosttext.py b/src/ocrmypdf/pdfinfo/ghosttext.py new file mode 100644 index 00000000..4b9e13a0 --- /dev/null +++ b/src/ocrmypdf/pdfinfo/ghosttext.py @@ -0,0 +1,96 @@ +# © 2018 James R. Barlow: github.com/jbarlow83 +# +# This file is part of OCRmyPDF. +# +# OCRmyPDF is free software: you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation, either version 3 of the License, or +# (at your option) any later version. +# +# OCRmyPDF is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with OCRmyPDF. If not, see . + +import re +import xml.etree.ElementTree as ET + +from ..exec import ghostscript + +# Forgive me for I have sinned +# I am using regular expressions to parse XML. However the XML in this case, +# generated by Ghostscript, is self-consistent enough to be parseable. +regex_remove_char_tags = re.compile(br""" + ] # anything single character but > + | \">\" # special case: trap ">" + )* + /> # terminate with '/>' +""", re.VERBOSE) + + +def page_get_textblocks(infile, pageno, xmltext, height): + """Get text boxes out of Ghostscript txtwrite xml""" + + root = xmltext + if not hasattr(xmltext, 'findall'): + return [] + + def blocks(): + for span in root.findall('.//span'): + bbox_str = span.attrib['bbox'] + font_size = span.attrib['size'] + pts = [int(pt) for pt in bbox_str.split()] + pts[1] = pts[1] - int(float(font_size) + 0.5) + bbox_topdown = tuple(pts) + bb = bbox_topdown + bbox_bottomup = (bb[0], height - bb[3], bb[2], height - bb[1]) + yield bbox_bottomup + + def joined_blocks(): + prev = None + for bbox in blocks(): + if prev is None: + prev = bbox + if bbox[1] == prev[1] and bbox[3] == prev[3]: + gap = prev[2] - bbox[0] + height = abs(bbox[3] - bbox[1]) + if gap < height: + # Join boxes + prev = (prev[0], prev[1], bbox[2], bbox[3]) + continue + # yield previously joined bboxes and start anew + yield prev + prev = bbox + if prev is not None: + yield prev + + return [block for block in joined_blocks()] + + +def extract_text_xml(infile, pdf, pageno=None, log=None): + existing_text = ghostscript.extract_text(infile, pageno=None) + existing_text = regex_remove_char_tags.sub(b' ', existing_text) + + try: + root = ET.fromstringlist([ + b'\n', existing_text, b'\n' + ]) + page_xml = root.findall('page') + except ET.ParseError as e: + log.error( + "An error occurred while attempting to retrieve existing text in " + "the input file. Will attempt to continue assuming that there is " + "no existing text in the file. The error was:") + log.error(e) + page_xml = [None] * len(pdf.pages) + + page_count_difference = len(pdf.pages) - len(page_xml) + if page_count_difference != 0: + log.error("The number of pages in the input file is inconsistent.") + if page_count_difference > 0: + page_xml.extend([None] * page_count_difference) + return page_xml diff --git a/src/ocrmypdf/pdfinfo/layout.py b/src/ocrmypdf/pdfinfo/layout.py new file mode 100644 index 00000000..9aef17b1 --- /dev/null +++ b/src/ocrmypdf/pdfinfo/layout.py @@ -0,0 +1,220 @@ +# © 2018 James R. Barlow: github.com/jbarlow83 +# +# This file is part of OCRmyPDF. +# +# OCRmyPDF is free software: you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation, either version 3 of the License, or +# (at your option) any later version. +# +# OCRmyPDF is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with OCRmyPDF. If not, see . + +import re +from math import copysign +from pathlib import Path +from unittest.mock import patch + +import pdfminer.encodingdb +import pdfminer.pdfdevice +import pdfminer.pdfinterp +from pdfminer.converter import PDFLayoutAnalyzer +from pdfminer.glyphlist import glyphname2unicode +from pdfminer.layout import (LAParams, LTChar, LTContainer, LTLayoutContainer, + LTPage, LTTextBox, LTTextLine) +from pdfminer.pdfdocument import PDFTextExtractionNotAllowed +from pdfminer.pdffont import (PDFCIDFont, PDFFont, PDFType3Font, + PDFUnicodeNotDefined) +from pdfminer.pdfpage import PDFPage +from pdfminer.utils import bbox2str, fsplit, matrix2str + +from ..exceptions import EncryptedPdfError + +STRIP_NAME = re.compile(r'[0-9]+') + +# +# Unconditional pdfminer patches +# + +def name2unicode(name): + """Fix pdfminer's regex in name2unicode function + + Font cids that are mapped to names of the form /g123 seem to be, by convention + characters with no corresponding Unicode entry. These can be subsetted fonts + or symbolic fonts. There seems to be no way to map /g123 fonts to Unicode, + barring a ToUnicode data structure. + """ + if name in glyphname2unicode: + return glyphname2unicode[name] + if name.startswith('g'): + raise KeyError(name) + m = STRIP_NAME.search(name) + if not m: + raise KeyError(name) + return chr(int(m.group(0))) +pdfminer.encodingdb.name2unicode = name2unicode + +original_PDFFont_init = PDFFont.__init__ +def PDFFont__init__(self, descriptor, widths, default_width=None): + original_PDFFont_init(self, descriptor, widths, default_width) + # PDF spec says descent should be negative + # A font with a positive descent implies it floats entirely above the + # baseline, i.e. it's not really a baseline anymore. I have fonts that + # claim a positive descent, but treating descent as positive always seems + # to misposition text. + if self.descent > 0: + self.descent = -self.descent +PDFFont.__init__ = PDFFont__init__ + +# +# pdfminer patches when creator is PScript5.dll +# + +def PDFType3Font__PScript5_get_height(self): + h = self.bbox[3]-self.bbox[1] + if h == 0: + h = self.ascent - self.descent + return h * copysign(1.0, self.vscale) + +def PDFType3Font__PScript5_get_descent(self): + return self.descent * copysign(1.0, self.vscale) + +def PDFType3Font__PScript5_get_ascent(self): + return self.ascent * copysign(1.0, self.vscale) + + +class LTStateAwareChar(LTChar): + """A subclass of LTChar that tracks text render mode at time of drawing""" + + __slots__ = ( + 'rendermode', '_text', 'matrix', 'fontname', 'adv', 'upright', 'size', + 'width', 'height', 'bbox', 'x0', 'x1', 'y0', 'y1' + ) + + def __init__(self, matrix, font, fontsize, scaling, rise, text, textwidth, + textdisp, textstate, *args): + super().__init__(matrix, font, fontsize, scaling, rise, text, textwidth, + textdisp, *args) + self.rendermode = textstate.render + + def is_compatible(self, obj): + """Check if characters can be combined into a textline + + We consider characters compatible if: + - the Unicode mapping is known, and both have the same render mode + - the Unicode mapping is unknown but both are part of the same font + """ + both_unicode_mapped = (isinstance(self._text, str) and + isinstance(obj._text, str)) + try: + if both_unicode_mapped: + return self.rendermode == obj.rendermode + font0, _ = self._text + font1, _ = obj._text + return font0 == font1 and self.rendermode == obj.rendermode + except (ValueError, AttributeError): + return False + + def get_text(self): + if isinstance(self._text, tuple): + return '�' + return self._text + + def __repr__(self): + return ('<%s %s matrix=%s rendermode=%r font=%r adv=%s text=%r>' % + (self.__class__.__name__, bbox2str(self.bbox), + matrix2str(self.matrix), self.rendermode, self.fontname, self.adv, + self.get_text())) + + +class TextPositionTracker(PDFLayoutAnalyzer): + """A page layout analyzer that pays attention to text visibility""" + + def __init__(self, rsrcmgr, pageno=1, laparams=None): + super().__init__(rsrcmgr, pageno, laparams) + self.textstate = None + self.result = None + + def begin_page(self, page, ctm): + super().begin_page(page, ctm) + self.cur_item = LTPage(self.pageno, page.mediabox) + + def end_page(self, page): + assert not self._stack, str(len(self._stack)) + assert isinstance(self.cur_item, LTPage), str(type(self.cur_item)) + if self.laparams is not None: + self.cur_item.analyze(self.laparams) + self.pageno += 1 + self.receive_layout(self.cur_item) + + def render_string(self, textstate, seq, *args): + self.textstate = textstate.copy() + super().render_string(self.textstate, seq, *args) + + def render_char(self, matrix, font, fontsize, scaling, rise, cid, *args): + try: + text = font.to_unichr(cid) + assert isinstance(text, str), str(type(text)) + except PDFUnicodeNotDefined: + text = self.handle_undefined_char(font, cid) + textwidth = font.char_width(cid) + textdisp = font.char_disp(cid) + item = LTStateAwareChar( + matrix, font, fontsize, scaling, rise, text, + textwidth, textdisp, self.textstate, *args) + self.cur_item.add(item) + return item.adv + + def handle_undefined_char(self, font, cid): + #log.info('undefined: %r, %r', font, cid) + return (font.fontname, cid) + + def receive_layout(self, ltpage): + self.result = ltpage + + def get_result(self): + return self.result + + +def get_page_analysis(infile, pageno, pscript5_mode): + rman = pdfminer.pdfinterp.PDFResourceManager(caching=True) + dev = TextPositionTracker(rman, laparams=LAParams()) + interp = pdfminer.pdfinterp.PDFPageInterpreter(rman, dev) + + if pscript5_mode: + patcher = patch.multiple( + 'pdfminer.pdffont.PDFType3Font', + spec=True, + get_ascent=PDFType3Font__PScript5_get_ascent, + get_descent=PDFType3Font__PScript5_get_descent, + get_height=PDFType3Font__PScript5_get_height + ) + patcher.start() + + with Path(infile).open('rb') as f: + page = PDFPage.get_pages(f, pagenos=[pageno], maxpages=0) + try: + interp.process_page(next(page)) + except PDFTextExtractionNotAllowed as e: + raise EncryptedPdfError() + finally: + if pscript5_mode: + patcher.stop() + + return dev.get_result() + + +def get_text_boxes(obj): + for child in obj: + if isinstance(child, (LTTextBox)): + yield child + else: + try: + yield from get_text_boxes(child) + except TypeError: + continue diff --git a/tests/conftest.py b/tests/conftest.py index 2b5a90f7..6bcbd4e1 100644 --- a/tests/conftest.py +++ b/tests/conftest.py @@ -18,13 +18,18 @@ import sys import os import platform - -pytest_plugins = ['helpers_namespace'] - -import pytest from pathlib import Path from subprocess import Popen, PIPE +pytest_plugins = ['helpers_namespace'] +import pytest + +try: + from pytest_cov.embed import cleanup_on_sigterm +except ImportError: + pass +else: + cleanup_on_sigterm() # pylint: disable=E1101 # pytest.helpers is dynamic so it confuses pylint diff --git a/tests/test_lept.py b/tests/test_lept.py index 9c03a6b8..85e2482e 100644 --- a/tests/test_lept.py +++ b/tests/test_lept.py @@ -18,9 +18,14 @@ import os import shutil -import pytest import sys +from pickle import dumps, loads + +import pytest +from PIL import Image, ImageChops + import ocrmypdf.leptonica as lept +from ocrmypdf.helpers import fspath def test_colormap_backgroundnorm(resources): @@ -28,3 +33,59 @@ def test_colormap_backgroundnorm(resources): # can handle that case pix = lept.Pix.open(resources / 'baiona_colormapped.png') pix.background_norm() + + +@pytest.fixture +def crom_pix(resources): + pix = lept.Pix.open(resources / 'crom.png') + im = Image.open(resources / 'crom.png') + return pix, im + + +def test_pix_basic(crom_pix): + pix, im = crom_pix + + assert pix.width == im.width + assert pix.height == im.height + assert pix.mode == im.mode + + +def test_pil_conversion(crom_pix): + pix, im = crom_pix + + # Check for pixel perfect + assert ImageChops.difference(pix.topil(), im).getbbox() is None + + +def test_pix_otsu(crom_pix): + pix, _ = crom_pix + im1bpp = pix.otsu_adaptive_threshold() + assert im1bpp.mode == '1' + + +def test_crop(resources): + pix = lept.Pix.open(resources / 'linn.png') + foreground = pix.crop_to_foreground() + assert foreground.width < pix.width + + +def test_clean_bg(resources): + pix = lept.Pix.open(resources / 'congress.jpg') + imbg = pix.clean_background_to_white() + + +def test_pickle(crom_pix): + pix, _ = crom_pix + pickled = dumps(pix) + pix2 = loads(pickled) + assert pix.mode == pix2.mode + + +def test_leptonica_compile(tmpdir): + from ocrmypdf.lib.compile_leptonica import ffibuilder + + # Compile the library but build it somewhere that won't interfere with + # existing compiled library. Also compile in API mode so that we test + # the interfaces, even though we use it ABI mode. + ffibuilder.compile(tmpdir=fspath(tmpdir), + target=fspath(tmpdir / 'lepttest.*')) diff --git a/tests/test_main.py b/tests/test_main.py index 8f18a8b2..b74c03aa 100644 --- a/tests/test_main.py +++ b/tests/test_main.py @@ -28,7 +28,7 @@ import PIL import pytest from ocrmypdf.pdfinfo import PdfInfo, Colorspace, Encoding -from ocrmypdf.exceptions import ExitCode +from ocrmypdf.exceptions import ExitCode, MissingDependencyError from ocrmypdf.exec import ghostscript, qpdf, tesseract from ocrmypdf.pdfa import file_claims_pdfa from ocrmypdf.leptonica import Pix @@ -190,11 +190,22 @@ def test_force_ocr(spoof_tesseract_cache, resources, outpdf): def test_skip_ocr(spoof_tesseract_cache, resources, outpdf): out = check_ocrmypdf(resources / 'graph_ocred.pdf', outpdf, '-s', - env=spoof_tesseract_cache) + env=spoof_tesseract_cache) pdfinfo = PdfInfo(out) assert pdfinfo[0].has_text +def test_redo_ocr(spoof_tesseract_cache, resources, outpdf): + in_ = resources / 'graph_ocred.pdf' + before = PdfInfo(in_, detailed_page_analysis=True) + out = check_ocrmypdf(in_, outpdf, '--redo-ocr', + env=spoof_tesseract_cache) + after = PdfInfo(out, detailed_page_analysis=True) + assert before[0].has_text and after[0].has_text + assert before[0].get_textareas() != after[0].get_textareas(), \ + "Expected text to be different after re-OCR" + + def test_argsfile(spoof_tesseract_noop, resources, outdir): path_argsfile = outdir / 'test_argsfile.txt' with open(str(path_argsfile), 'w') as argsfile: @@ -208,7 +219,7 @@ def test_argsfile(spoof_tesseract_noop, resources, outdir): @pytest.mark.parametrize('renderer', RENDERERS) def test_ocr_timeout(renderer, resources, outpdf): out = check_ocrmypdf(resources / 'skew.pdf', outpdf, - '--tesseract-timeout', '0.01', + '--tesseract-timeout', '0', '--pdf-renderer', renderer) pdfinfo = PdfInfo(out) assert not pdfinfo[0].has_text @@ -377,7 +388,7 @@ def test_tesseract_image_too_big(renderer, spoof_tesseract_big_image_error, def test_algo4(resources, spoof_tesseract_noop, outpdf): p, _, _ = run_ocrmypdf(resources / 'encrypted_algo4.pdf', outpdf, env=spoof_tesseract_noop) - assert p.returncode == ExitCode.ok + assert p.returncode == ExitCode.encrypted_pdf @pytest.mark.parametrize('renderer', RENDERERS) @@ -934,3 +945,16 @@ def test_livecycle(resources, no_outpdf): ) assert p.returncode == ExitCode.input_file, err + + +def test_version_check(): + from ocrmypdf.exec import get_version + + with pytest.raises(MissingDependencyError): + get_version('NOT_FOUND_UNLIKELY_ON_PATH') + + with pytest.raises(MissingDependencyError): + get_version('sh', version_arg='-c') + + with pytest.raises(MissingDependencyError): + get_version('echo') diff --git a/tests/test_pdfinfo.py b/tests/test_pdfinfo.py index c3ee0b6d..85f51311 100644 --- a/tests/test_pdfinfo.py +++ b/tests/test_pdfinfo.py @@ -145,13 +145,13 @@ def test_pickle(resources): # For multiprocessing we must be able to pickle our information - if # this fails then we are probably storing some unpickleabe pikepdf or # other external data around - filename = resources / 'formxobject.pdf' + filename = resources / 'graph_ocred.pdf' pdf = pdfinfo.PdfInfo(filename) pickle.dumps(pdf) def test_regex(): - rx = pdfinfo.regex_remove_char_tags + rx = pdfinfo.ghosttext.regex_remove_char_tags must_match = [ b'', @@ -175,3 +175,11 @@ def test_vector(resources): filename = resources / 'vector.pdf' pdf = pdfinfo.PdfInfo(filename) assert pdf[0].has_vector + assert not pdf[0].has_text + + +def test_ocr_detection(resources): + filename = resources / 'graph_ocred.pdf' + pdf = pdfinfo.PdfInfo(filename) + assert not pdf[0].has_vector + assert pdf[0].has_text diff --git a/tests/test_qpdf.py b/tests/test_qpdf.py new file mode 100644 index 00000000..c6387834 --- /dev/null +++ b/tests/test_qpdf.py @@ -0,0 +1,24 @@ +# © 2018 James R. Barlow: github.com/jbarlow83 +# +# This file is part of OCRmyPDF. +# +# OCRmyPDF is free software: you can redistribute it and/or modify +# it under the terms of the GNU General Public License as published by +# the Free Software Foundation, either version 3 of the License, or +# (at your option) any later version. +# +# OCRmyPDF is distributed in the hope that it will be useful, +# but WITHOUT ANY WARRANTY; without even the implied warranty of +# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with OCRmyPDF. If not, see . + +import pytest + +import ocrmypdf.exec.qpdf as qpdf + +def test_qpdf_error(resources): + assert qpdf.check(resources / 'blank.pdf') + assert not qpdf.check(__file__) diff --git a/tests/test_rotation.py b/tests/test_rotation.py index a829ef2e..ae3751e9 100644 --- a/tests/test_rotation.py +++ b/tests/test_rotation.py @@ -17,6 +17,7 @@ import logging from io import BytesIO +from unittest.mock import Mock from PIL import Image import pytest @@ -25,7 +26,7 @@ import pikepdf from ocrmypdf import leptonica from ocrmypdf.pdfinfo import PdfInfo -from ocrmypdf.exec import ghostscript +from ocrmypdf.exec import ghostscript, tesseract from ocrmypdf.helpers import fspath @@ -169,6 +170,8 @@ def test_rotate_deskew_timeout(resources, outdir): check_ocrmypdf( resources / 'rotated_skew.pdf', outdir / 'deskewed.pdf', + '--rotate-pages', + '--rotate-pages-threshold', '0', '--deskew', '--tesseract-timeout', '0', '--pdf-renderer', 'sandwich' @@ -226,3 +229,13 @@ def test_rotate_page_level(image_angle, page_angle, resources, outdir): assert p.returncode == 0, err assert check_monochrome_correlation(outdir, reference, 1, out, 1) > 0.2 + + +def test_tesseract_orientation(resources, tmpdir): + pix = leptonica.Pix.open(resources / 'crom.png') + pix_rotated = pix.rotate_orth(2) # 180 degrees clockwise + pix_rotated.write_implied_format(tmpdir / '000001.png') + + log = Mock() + tesseract.get_orientation( # Test results of this are unreliable + tmpdir / '000001.png', engine_mode='3', timeout=10, log=log) diff --git a/tests/test_tess4.py b/tests/test_tess4.py index 859b3cbb..0af3b7e6 100644 --- a/tests/test_tess4.py +++ b/tests/test_tess4.py @@ -16,8 +16,9 @@ # along with OCRmyPDF. If not, see . import pytest -from ocrmypdf.exceptions import ExitCode +from ocrmypdf.exceptions import ExitCode, MissingDependencyError from ocrmypdf.exec import tesseract +from ocrmypdf.helpers import fspath from ocrmypdf import pdfinfo import sys import os @@ -43,6 +44,7 @@ def _ensure_tess4(): tess4 = Path(os.environ['OCRMYPDF_TESS4']) assert tess4.is_file() env['PATH'] = tess4.parent + ':' + env['PATH'] + env['OCRMYPDF_TESS4'] = os.environ['OCRMYPDF_TESS4'] return env raise EnvironmentError("Can't find Tesseract 4") @@ -56,9 +58,12 @@ def ensure_tess4(): @contextmanager def modified_os_environ(env): old_env = os.environ.copy() - os.environ = env + os.environ.update(env) yield - os.environ = old_env + for key in env: + del os.environ[key] + if key in old_env: + os.environ[key] = old_env[key] def tess4_available(): @@ -149,3 +154,14 @@ def test_content_preservation(ensure_tess4, resources, outpdf): info = pdfinfo.PdfInfo(outpdf) page = info[0] assert len(page.images) > 1, "masks were rasterized" + + +def test_no_languages(ensure_tess4, tmpdir): + env = ensure_tess4 + (tmpdir / 'tessdata').mkdir() + env['TESSDATA_PREFIX'] = fspath(tmpdir) + + with modified_os_environ(env): + with pytest.raises(MissingDependencyError): + tesseract.languages.cache_clear() + tesseract.languages()