Implement —pdfa-image-compression to control Ghostscript’s compression

Fixes #163

ocrmypdf/__main__.py

@@ -242,6 +242,18 @@ def complain(message):
     '--rotate-pages-threshold', default=14.0, type=float, metavar='CONFIDENCE',
     help="only rotate pages when confidence is above this value (arbitrary "
          "units reported by tesseract)")
+advanced.add_argument(
+    '--pdfa-image-compression', choices=['auto', 'jpeg', 'lossless'],
+    default='auto',
+    help="specify how to compress images in the output PDF/A. 'auto' lets "
+         "OCRmyPDF decide.  'jpeg' changes all grayscale and color images to "
+         "JPEG compression.  'lossless' uses PNG-style lossless compression "
+         "for all images.  Monochrome images are always compressed using a "
+         "lossless codec.  In 'auto' mode OCRmyPDF lets Ghostscript's auto "
+         "filter decide how to compress the image.  Compression settings "
+         "are applied to all pages, including those for which OCR was "
+         "skipped.  Not supported for --output-type=pdf ; that setting "
+         "preserves the original compression of all images.")
 
 debugging = parser.add_argument_group(
     "Debugging",
@@ -346,6 +358,12 @@ def check_options_advanced(options, log):
         raise MissingDependencyError(
             "--pdf-renderer tess4 requires Tesseract 4.x "
             "commit 3d9fb3b or later")
+    if options.pdfa_image_compression != 'auto' and \
+            options.output_type != 'pdfa':
+        log.warning(
+            "--pdfa-image-compression argument has no effect when "
+            "--output-type is not 'pdfa'"
+        )
 
 
 def check_options_metadata(options, log):

ocrmypdf/exec/ghostscript.py

@@ -65,7 +65,28 @@ def rasterize_pdf(input_file, output_file, xres, yres, raster_device, log,
             raise SubprocessOutputError()
 
 
-def generate_pdfa(pdf_pages, output_file, log, threads=1):
+def generate_pdfa(pdf_pages, output_file, compression, log, threads=1):
+    compression_args = []
+    if compression == 'jpeg':
+        compression_args = [
+            "-dAutoFilterColorImages=false",
+            "-dColorImageFilter=/DCTEncode",
+            "-dAutoFilterGrayImages=false",
+            "-dGrayImageFilter=/DCTEncode",
+        ]
+    elif compression == 'lossless':
+        compression_args = [
+            "-dAutoFilterColorImages=false",
+            "-dColorImageFilter=/FlateEncode",
+            "-dAutoFilterGrayImages=false",
+            "-dGrayImageFilter=/FlateEncode",
+        ]
+    else:
+        compression_args = [
+            "-dAutoFilterColorImages=true",
+            "-dAutoFilterGrayImages=true",
+        ]
+
     with NamedTemporaryFile(delete=True) as gs_pdf:
         args_gs = [
             get_program("gs"),
@@ -76,11 +97,8 @@ def generate_pdfa(pdf_pages, output_file, log, threads=1):
             "-sDEVICE=pdfwrite",
             "-dAutoRotatePages=/None",
             "-sColorConversionStrategy=/RGB",
-            "-sProcessColorModel=DeviceRGB",
-            "-dAutoFilterColorImages=false",
-            "-sColorImageFilter=",
-            "-dAutoFilterGrayImages=false",
-            "-sGrayImageFilter=",
+            "-sProcessColorModel=DeviceRGB"
+        ] + compression_args + [
             "-dJPEGQ=95",
             "-dPDFA=2",
             "-dPDFACompatibilityPolicy=1",

ocrmypdf/pipeline.py

@@ -521,7 +521,7 @@ def select_visible_page_image(
     if pageinfo['images'] and \
             all(im['enc'] == 'jpeg' for im in pageinfo['images']):
         log.debug('{:4d}: JPEG input -> JPEG output'.format(
-            page_number(page_pdf)))
+            page_number(image)))
         # If all images were JPEGs originally, produce a JPEG as output
         im = Image.open(image)
 
@@ -807,7 +807,9 @@ def input_file_order(s):
 
     pdf_pages = sorted(input_files, key=input_file_order)
     log.debug("Final pages: " + "\n".join(pdf_pages))
-    ghostscript.generate_pdfa(pdf_pages, output_file, log, options.jobs or 1)
+    ghostscript.generate_pdfa(
+        pdf_pages, output_file, options.pdfa_image_compression,
+        log, options.jobs or 1)
 
 
 def merge_pages_qpdf(

tests/test_main.py

@@ -885,9 +885,10 @@ def test_no_contents(spoof_tesseract_noop, resources, outpdf):
 @pytest.mark.parametrize('image', [
     'baiona.png',
     'baiona_gray.png',
+    'congress.jpg'
     ])
-def test_lossless_to_lossless(spoof_tesseract_noop, ocrmypdf_exec,
-                              resources, image, outpdf):
+def test_compression_preserved(spoof_tesseract_noop, ocrmypdf_exec,
+                               resources, image, outpdf):
     from PIL import Image
 
     input_file = str(resources / image)
@@ -897,7 +898,8 @@ def test_lossless_to_lossless(spoof_tesseract_noop, ocrmypdf_exec,
 
     # Runs: ocrmypdf - output.pdf < testfile
     with open(input_file, 'rb') as input_stream:
-        p_args = ocrmypdf_exec + ['--image-dpi', '150', '-', output_file]
+        p_args = ocrmypdf_exec + [
+            '--image-dpi', '150', '--output-type', 'pdf', '-', output_file]
         p = Popen(
             p_args, close_fds=True, stdout=PIPE, stderr=PIPE,
             stdin=input_stream, env=spoof_tesseract_noop)
@@ -906,11 +908,62 @@ def test_lossless_to_lossless(spoof_tesseract_noop, ocrmypdf_exec,
         assert p.returncode == ExitCode.ok
 
     pdfinfo = pdf_get_all_pageinfo(output_file)
-    assert pdfinfo[0]['images'][0]['enc'] != 'jpeg', \
-        "Lossless compression changed to lossy!"
+
+    pdfimage = pdfinfo[0]['images'][0]
+
+    if input_file.endswith('.png'):
+        assert pdfimage['enc'] != 'jpeg', \
+            "Lossless compression changed to lossy!"
+    elif input_file.endswith('.jpg'):
+        assert pdfimage['enc'] == 'jpeg', \
+            "Lossy compression changed to lossless!"
+    if im.mode.startswith('RGB') or im.mode.startswith('BGR'):
+        assert pdfimage['color'] == 'rgb', \
+            "Colorspace changed"
+    elif im.mode.startswith('L'):
+        assert pdfimage['color'] == 'gray', \
+            "Colorspace changed"
+
+
+@pytest.mark.parametrize('image,compression', [
+    ('baiona.png', 'jpeg'),
+    ('baiona_gray.png', 'lossless'),
+    ('congress.jpg', 'lossless')
+    ])
+def test_compression_changed(spoof_tesseract_noop, ocrmypdf_exec,
+                             resources, image, compression, outpdf):
+    from PIL import Image
+
+    input_file = str(resources / image)
+    output_file = str(outpdf)
+
+    im = Image.open(input_file)
+
+    # Runs: ocrmypdf - output.pdf < testfile
+    with open(input_file, 'rb') as input_stream:
+        p_args = ocrmypdf_exec + [
+            '--image-dpi', '150', '--output-type', 'pdfa',
+            '--pdfa-image-compression', compression,
+            '-', output_file]
+        p = Popen(
+            p_args, close_fds=True, stdout=PIPE, stderr=PIPE,
+            stdin=input_stream, env=spoof_tesseract_noop)
+        out, err = p.communicate()
+
+        assert p.returncode == ExitCode.ok
+
+    pdfinfo = pdf_get_all_pageinfo(output_file)
+
+    pdfimage = pdfinfo[0]['images'][0]
+
+    if compression == 'jpeg':
+        assert pdfimage['enc'] == 'jpeg'
+    elif compression == 'lossless':
+        assert pdfimage['enc'] == 'image'
+
     if im.mode.startswith('RGB') or im.mode.startswith('BGR'):
-        assert pdfinfo[0]['images'][0]['color'] == 'rgb', \
+        assert pdfimage['color'] == 'rgb', \
             "Colorspace changed"
     elif im.mode.startswith('L'):
-        assert pdfinfo[0]['images'][0]['color'] == 'gray', \
+        assert pdfimage['color'] == 'gray', \
             "Colorspace changed"