yujunjun/OCRmyPDF
1
0
Fork 0
mirror of https://github.com/ocrmypdf/OCRmyPDF.git synced 2025-12-07 12:31:25 +00:00
Code Issues Packages Projects Releases Wiki Activity

Refactor JBIG2 path for non-CCITT monochrome images

Browse Source
This commit is contained in:
James R. Barlow 2018-05-15 15:32:15 -07:00
parent 6171de41bf
commit 08bf651ef2
1 changed files with 26 additions and 15 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

39
src/ocrmypdf/_optimize.py
View File

@ -39,16 +39,20 @@ JPEG_QUALITY = 75
PNG_QUALITY = (65, 75) PNG_QUALITY = (65, 75)
def img_name(root, xref, ext):
return str(root / '{:08d}{}'.format(xref, ext))
def png_name(root, xref): def png_name(root, xref):
return str(root / '{:08d}.png'.format(xref)) return img_name(root, xref, '.png')
def jpg_name(root, xref): def jpg_name(root, xref):
return str(root / '{:08d}.jpg'.format(xref)) return img_name(root, xref, '.jpg')
def tif_name(root, xref): def tif_name(root, xref):
return str(root / '{:08d}.tif'.format(xref)) return img_name(root, xref, '.tif')
def extract_image(*, doc, pike, root, log, image, xref, jbig2s, def extract_image(*, doc, pike, root, log, image, xref, jbig2s,
@ -75,11 +79,14 @@ def extract_image(*, doc, pike, root, log, image, xref, jbig2s,
if pim.bits_per_component == 1 \ if pim.bits_per_component == 1 \
and filtdp != '/JBIG2Decode' \ and filtdp != '/JBIG2Decode' \
and jbig2enc.available(): and jbig2enc.available():
with Path(tif_name(root, xref)).open('wb') as f: try:
result = pim.write_stream(f) imgname = Path(root / '{:08d}'.format(xref))
if not result: with imgname.open('wb') as f:
ext = pim.extract(f)
imgname.rename(imgname.with_suffix(ext))
except pikepdf.UnsupportedImageTypeError:
return False return False
jbig2s.append(xref) jbig2s.append((xref, ext))
elif filtdp[0] == '/DCTDecode' \ elif filtdp[0] == '/DCTDecode' \
and options.optimize >= 2: and options.optimize >= 2:
# This is a simple heuristic derived from some training data, that has # This is a simple heuristic derived from some training data, that has
@ -98,9 +105,12 @@ def extract_image(*, doc, pike, root, log, image, xref, jbig2s,
# iccbytes = icc.read_bytes() # iccbytes = icc.read_bytes()
# with Image.open(stream) as im: # with Image.open(stream) as im:
# im.save(jpg_name(root, xref), icc_profile=iccbytes) # im.save(jpg_name(root, xref), icc_profile=iccbytes)
with Path(jpg_name(root, xref)).open('wb') as f: try:
result = pim.write_stream(f) imgname = Path(root / '{:08d}'.format(xref))
if not result: with imgname.open('wb') as f:
ext = pim.extract(f)
imgname.rename(imgname.with_suffix(ext))
except pikepdf.UnsupportedImageTypeError:
return False return False
jpegs.append(xref) jpegs.append(xref)
elif pim.indexed \ elif pim.indexed \
@ -186,12 +196,12 @@ def convert_to_jbig2(pike, jbig2_groups, root, log, options):
with concurrent.futures.ThreadPoolExecutor( with concurrent.futures.ThreadPoolExecutor(
max_workers=options.jobs) as executor: max_workers=options.jobs) as executor:
futures = [] futures = []
for group, xrefs in jbig2_groups.items(): for group, xref_exts in jbig2_groups.items():
prefix = 'group{:08d}'.format(group) prefix = 'group{:08d}'.format(group)
future = executor.submit( future = executor.submit(
jbig2enc.convert_group, jbig2enc.convert_group,
cwd=str(root), cwd=str(root),
infiles=(png_name(root, xref) for xref in xrefs), infiles=(img_name(root, xref, ext) for xref, ext in xref_exts),
out_prefix=prefix out_prefix=prefix
) )
futures.append(future) futures.append(future)
@ -199,12 +209,13 @@ def convert_to_jbig2(pike, jbig2_groups, root, log, options):
proc = future.result() proc = future.result()
log.debug(proc.stderr.decode()) log.debug(proc.stderr.decode())
for group, xrefs in jbig2_groups.items(): for group, xref_exts in jbig2_groups.items():
prefix = 'group{:08d}'.format(group) prefix = 'group{:08d}'.format(group)
jbig2_globals_data = (root / (prefix + '.sym')).read_bytes() jbig2_globals_data = (root / (prefix + '.sym')).read_bytes()
jbig2_globals = pikepdf.Stream(pike, jbig2_globals_data) jbig2_globals = pikepdf.Stream(pike, jbig2_globals_data)
for n, xref in enumerate(xrefs): for n, xref_ext in enumerate(xref_exts):
xref, ext = xref_ext
jbig2_im_file = root / (prefix + '.{:04d}'.format(n)) jbig2_im_file = root / (prefix + '.{:04d}'.format(n))
jbig2_im_data = jbig2_im_file.read_bytes() jbig2_im_data = jbig2_im_file.read_bytes()
im_obj = pike._get_object_id(xref, 0) im_obj = pike._get_object_id(xref, 0)