def wrap_object(obj, width, margin):
''' Wrap an xobj in its own page object.
'''
fmt = 'q %s 0 0 %s %s %s cm /MyImage Do Q'
contents = PdfDict(indirect=True)
subtype = obj.Subtype
if subtype == PdfName.Form:
contents._stream = obj.stream
contents.Length = obj.Length
contents.Filter = obj.Filter
contents.DecodeParms = obj.DecodeParms
resources = obj.Resources
mbox = obj.BBox
elif subtype == PdfName.Image: # Image
xoffset = margin[0]
yoffset = margin[1]
cw = width - margin[0] - margin[2]
iw, ih = float(obj.Width), float(obj.Height)
ch = 1.0 * cw / iw * ih
height = ch + margin[1] + margin[3]
p = tuple(('%.9f' % x).rstrip('0').rstrip('.') for x in (cw, ch, xoffset, yoffset))
contents.stream = fmt % p
resources = PdfDict(XObject=PdfDict(MyImage=obj))
mbox = PdfArray((0, 0, width, height))
else:
raise TypeError("Expected Form or Image XObject")
return PdfDict(
indirect=True,
Type=PdfName.Page,
MediaBox=mbox,
Resources=resources,
Contents=contents,
)
def _pdf_image(self, psem):
with TemporaryDirectory(prefix="djpdf-") as temp_dir:
# JBIG2Globals are only used in symbol mode
# In symbol mode jbig2 writes output to files otherwise
# it's written to stdout
symbol_mode = self.jbig2_threshold != 1
images_with_shared_globals = []
if symbol_mode and SHARE_JBIG2_GLOBALS:
# Find all Jbig2Images that share the same symbol directory
for obj in self._factory._cache:
if (isinstance(obj, Jbig2Image) and
self.compression == obj.compression and
self.jbig2_threshold == obj.jbig2_threshold):
images_with_shared_globals.append(obj)
else:
# The symbol directory is not shared with other Jbig2Images
images_with_shared_globals.append(self)
# Promise all handled Jbig2Images the finished image
image_futures = []
my_image_future = None
for image in images_with_shared_globals:
future = asyncio.Future()
asyncio.ensure_future(image._cache.get(future))
image_futures.append(future)
if image is self:
my_image_future = future
# All futures are in place, the lock can be released
self._factory._cache_lock.release()
self._cache_lock_acquired = False
# Prepare everything in parallel
@asyncio.coroutine
def get_jbig2_images(psem):
# Convert images with ImageMagick to bitonal png in parallel
yield from asyncio.gather(*[
run_command_async([
CONVERT_CMD,
"-alpha", "remove",
"-alpha", "off",
"-colorspace", "gray",
"-threshold", "50%",
path.abspath(image.filename),
path.abspath(path.join(temp_dir,
"input.%d.png" % i))], psem)
for i, image in enumerate(images_with_shared_globals)])
cmd = [JBIG2_CMD, "-p"]
if symbol_mode:
cmd.extend(["-s", "-t",
format_number(self.jbig2_threshold, 4)])
for i, _ in enumerate(images_with_shared_globals):
cmd.append(path.abspath(path.join(temp_dir,
"input.%d.png" % i)))
jbig2_images = []
jbig2_globals = None
if symbol_mode:
yield from run_command_async(cmd, psem, cwd=temp_dir)
jbig2_globals = PdfDict()
jbig2_globals.indirect = True
with open(path.join(temp_dir, "output.sym"), "rb") as f:
jbig2_globals.stream = f.read().decode("latin-1")
for i, _ in enumerate(images_with_shared_globals):
with open(path.join(temp_dir,
"output.%04d" % i), "rb") as f:
jbig2_images.append(f.read())
else:
jbig2_images.append(
(yield from run_command_async(cmd, psem,
cwd=temp_dir)))
return jbig2_images, jbig2_globals
@asyncio.coroutine
def get_image_mask(image, psem):
if image._mask is None:
return None
return (yield from image._mask.pdf_image(psem))
((jbig2_images, jbig2_globals),
image_masks) = yield from asyncio.gather(
get_jbig2_images(psem),
asyncio.gather(*[get_image_mask(image, psem)
for image in images_with_shared_globals]))
for image, jbig2_image, image_mask, image_future in zip(
images_with_shared_globals, jbig2_images, image_masks,
image_futures):
(width, height, xres, yres) = struct.unpack(
'>IIII', jbig2_image[11:27])
pdf_image = PdfDict()
pdf_image.indirect = True
pdf_image.Type = PdfName.XObject
pdf_image.Subtype = PdfName.Image
pdf_image.Width = width
pdf_image.Height = height
if image._image_mask:
pdf_image.ImageMask = PdfBool(True)
else:
pdf_image.ColorSpace = PdfName.DeviceGray
if image_mask is not None:
pdf_image.Mask = image_mask
pdf_image.BitsPerComponent = 1
pdf_image.Filter = [PdfName.JBIG2Decode]
if symbol_mode:
pdf_image.DecodeParms = [{
PdfName.JBIG2Globals: jbig2_globals}]
pdf_image.stream = jbig2_image.decode("latin-1")
image_future.set_result(pdf_image)
return my_image_future.result()
def _pdf_image(self, psem):
with TemporaryDirectory(prefix="djpdf-") as temp_dir:
# JBIG2Globals are only used in symbol mode
# In symbol mode jbig2 writes output to files otherwise
# it's written to stdout
symbol_mode = self.jbig2_threshold != 1
images_with_shared_globals = []
if symbol_mode and SHARE_JBIG2_GLOBALS:
# Find all Jbig2Images that share the same symbol directory
for obj in self._factory._cache:
if (isinstance(obj, Jbig2Image)
and self.compression == obj.compression
and self.jbig2_threshold == obj.jbig2_threshold):
images_with_shared_globals.append(obj)
else:
# The symbol directory is not shared with other Jbig2Images
images_with_shared_globals.append(self)
# Promise all handled Jbig2Images the finished image
image_futures = []
my_image_future = None
for image in images_with_shared_globals:
future = asyncio.Future()
asyncio.ensure_future(image._cache.get(future))
image_futures.append(future)
if image is self:
my_image_future = future
# All futures are in place, the lock can be released
self._factory._cache_lock.release()
self._cache_lock_acquired = False
# Prepare everything in parallel
@asyncio.coroutine
def get_jbig2_images(psem):
# Convert images with ImageMagick to bitonal png in parallel
yield from asyncio.gather(*[
run_command_async([
CONVERT_CMD, "-alpha", "remove", "-alpha", "off",
"-colorspace", "gray", "-threshold", "50%",
path.abspath(image.filename),
path.abspath(path.join(temp_dir, "input.%d.png" % i))
], psem)
for i, image in enumerate(images_with_shared_globals)
])
cmd = [JBIG2_CMD, "-p"]
if symbol_mode:
cmd.extend(
["-s", "-t",
format_number(self.jbig2_threshold, 4)])
for i, _ in enumerate(images_with_shared_globals):
cmd.append(
path.abspath(path.join(temp_dir, "input.%d.png" % i)))
jbig2_images = []
jbig2_globals = None
if symbol_mode:
yield from run_command_async(cmd, psem, cwd=temp_dir)
jbig2_globals = PdfDict()
jbig2_globals.indirect = True
with open(path.join(temp_dir, "output.sym"), "rb") as f:
jbig2_globals.stream = f.read().decode("latin-1")
for i, _ in enumerate(images_with_shared_globals):
with open(path.join(temp_dir, "output.%04d" % i),
"rb") as f:
jbig2_images.append(f.read())
else:
jbig2_images.append((yield from
run_command_async(cmd,
psem,
cwd=temp_dir)))
return jbig2_images, jbig2_globals
@asyncio.coroutine
def get_image_mask(image, psem):
if image._mask is None:
return None
return (yield from image._mask.pdf_image(psem))
((jbig2_images, jbig2_globals),
image_masks) = yield from asyncio.gather(
get_jbig2_images(psem),
asyncio.gather(*[
get_image_mask(image, psem)
for image in images_with_shared_globals
]))
for image, jbig2_image, image_mask, image_future in zip(
images_with_shared_globals, jbig2_images, image_masks,
image_futures):
(width, height, xres,
yres) = struct.unpack('>IIII', jbig2_image[11:27])
pdf_image = PdfDict()
pdf_image.indirect = True
pdf_image.Type = PdfName.XObject
pdf_image.Subtype = PdfName.Image
pdf_image.Width = width
pdf_image.Height = height
if image._image_mask:
pdf_image.ImageMask = PdfBool(True)
else:
# NOTE: DefaultGray color space is required for PDF/A
pdf_image.ColorSpace = PdfName.DeviceGray
if image_mask is not None:
pdf_image.Mask = image_mask
pdf_image.BitsPerComponent = 1
pdf_image.Filter = [PdfName.JBIG2Decode]
if symbol_mode:
pdf_image.DecodeParms = [{
PdfName.JBIG2Globals:
jbig2_globals
}]
pdf_image.stream = jbig2_image.decode("latin-1")
image_future.set_result(pdf_image)
return my_image_future.result()