Merge branch 'feature/jbig2thresh' into v15

misc/completion/ocrmypdf.bash

@@ -38,6 +38,7 @@ __ocrmypdf_arguments()
 --jpeg-quality                  (JPEG quality [0..100])
 --png-quality                   (PNG quality [0..100])
 --jbig2-lossy                   (enable lossy JBIG2 (see docs))
+--jbig2-threshold               (set JBIG2 threshold (see docs))
 --pages                         (apply OCR to only the specified pages)
 --max-image-mpixels             (image decompression bomb threshold)
 --pdf-renderer                  (select PDF renderer options)

misc/completion/ocrmypdf.fish

@@ -84,6 +84,7 @@ complete -c ocrmypdf -x -l skip-big -d "skip OCR on pages larger than this many
 complete -c ocrmypdf -x -l jpeg-quality -d "JPEG quality [0..100]"
 complete -c ocrmypdf -x -l png-quality -d "PNG quality [0..100]"
 complete -c ocrmypdf -x -l jbig2-lossy -d "enable lossy JBIG2 (see docs)"
+complete -c ocrmypdf -x -l jbig2-threshold -d "JBIG2 compression threshold (see docs)"
 complete -c ocrmypdf -x -l max-image-mpixels -d "image decompression bomb threshold"
 complete -c ocrmypdf -x -l pages -d "apply OCR to only the specified pages"
 complete -c ocrmypdf -x -l tesseract-config -d "set custom tesseract config file"

src/ocrmypdf/_exec/jbig2enc.py

@@ -23,15 +23,17 @@ def available():
     return True
 
 
-def convert_group(*, cwd, infiles, out_prefix):
+def convert_group(*, cwd, infiles, out_prefix, threshold):
     args = [
         'jbig2',
         '-b',
         out_prefix,
-        '-s',  # symbol mode (lossy)
+        '--symbol-mode',  # symbol mode (lossy)
+        '-t',
+        str(threshold),  # threshold
         # '-r', # refinement mode (lossless symbol mode, currently disabled in
         # jbig2)
-        '-p',
+        '--pdf',
     ]
     args.extend(infiles)
     proc = run(args, cwd=cwd, stdout=PIPE, stderr=PIPE)
@@ -40,11 +42,13 @@ def convert_group(*, cwd, infiles, out_prefix):
 
 
 def convert_group_mp(args):
-    return convert_group(cwd=args[0], infiles=args[1], out_prefix=args[2])
+    return convert_group(
+        cwd=args[0], infiles=args[1], out_prefix=args[2], threshold=args[3]
+    )
 
 
-def convert_single(*, cwd, infile, outfile):
-    args = ['jbig2', '-p', infile]
+def convert_single(*, cwd, infile, outfile, threshold):
+    args = ['jbig2', '--pdf', '-t', str(threshold), infile]
     with open(outfile, 'wb') as fstdout:
         proc = run(args, cwd=cwd, stdout=fstdout, stderr=PIPE)
     proc.check_returncode()
@@ -52,4 +56,6 @@ def convert_single(*, cwd, infile, outfile):
 
 
 def convert_single_mp(args):
-    return convert_single(cwd=args[0], infile=args[1], outfile=args[2])
+    return convert_single(
+        cwd=args[0], infile=args[1], outfile=args[2], threshold=args[3]
+    )

src/ocrmypdf/builtin_plugins/optimize.py

@@ -86,6 +86,16 @@ def add_options(parser):
         # Adjust number of pages to consider at once for JBIG2 compression
         help=argparse.SUPPRESS,
     )
+    optimizing.add_argument(
+        '--jbig2-threshold',
+        type=numeric(float, 0.4, 0.9),
+        default=0.85,
+        metavar='T',
+        help=(
+            "Adjust JBIG2 symbol code classification threshold "
+            "(default 0.85), range 0.4 to 0.9."
+        ),
+    )
 
 
 @hookimpl

src/ocrmypdf/optimize.py

@@ -367,6 +367,7 @@ def _produce_jbig2_images(
                 fspath(root),  # =cwd
                 (img_name(root, xref, ext) for xref, ext in xref_exts),  # =infiles
                 prefix,  # =out_prefix
+                options.jbig2_threshold,
             )
 
     def jbig2_single_args(root, groups: dict[int, list[XrefExt]]):
@@ -379,6 +380,7 @@ def _produce_jbig2_images(
                     fspath(root),
                     img_name(root, xref, ext),
                     root / f'{prefix}.{n:04d}',
+                    options.jbig2_threshold,
                 )
 
     if options.jbig2_page_group_size > 1:
@@ -737,6 +739,7 @@ def main(infile, outfile, level, jobs=1):
             self.png_quality = png_quality
             self.jbig2_page_group_size = 0
             self.jbig2_lossy = jb2lossy
+            self.jbig2_threshold = 0.85
             self.quiet = True
             self.progress_bar = False
 

tests/test_optimize.py

@@ -94,6 +94,8 @@ def test_jbig2_lossy(lossy, resources, outpdf):
         '20',
         '--plugin',
         'tests/plugins/tesseract_noop.py',
+        '--jbig2-threshold',
+        '0.7',
     ]
     if lossy:
         args.append('--jbig2-lossy')