def _replace_image(self, img_xref, resampled_image):
image_meta = self._pdf.lookup(img_xref).content
alpha_xref = image_meta.get(PDFName("/SMask"))
new_image_obj = PDFObject.create_image(img_xref.objid, img_xref.gennum, resampled_image, alpha_xref = alpha_xref)
self._pdf.replace_object(new_image_obj)
if resampled_image.alpha:
new_alpha_obj = PDFObject.create_image(alpha_xref.objid, alpha_xref.gennum, resampled_image.alpha)
self._pdf.replace_object(new_alpha_obj)
def _create_object(self, content, raw_stream=None):
objid = self._pdf.get_free_objid()
obj = PDFObject.create(objid=objid, gennum=0, content=content)
if raw_stream is not None:
obj.set_stream(EncodedObject.create(raw_stream))
self._pdf.replace_object(obj)
return PDFXRef(objid, 0)
def _add_xmp_metadata(self):
info_node_xref = self._pdf.trailer[PDFName("/Info")]
info_node = self._pdf.lookup(info_node_xref)
metadata_date = Timestamp.localnow()
modify_date = Timestamp.frompdf(info_node.content[PDFName("/ModDate")].decode("ascii")) if (PDFName("/ModDate") in info_node.content) else metadata_date
create_date = Timestamp.frompdf(info_node.content[PDFName("/CreationDate")].decode("ascii")) if (PDFName("/CreationDate") in info_node.content) else metadata_date
xmp_metadata = {
"creator_tool": self._pdf.get_info("Creator"),
"producer": self._pdf.get_info("Producer"),
"modify_date": modify_date.format_xml(),
"create_date": create_date.format_xml(),
"metadata_date": metadata_date.format_xml(),
"description": self._pdf.get_info("Subject"),
"title": self._pdf.get_info("Title"),
"creator": self._pdf.get_info("Author"),
"keywords": self._pdf.get_info("Keywords"),
"document_uuid": str(uuid.uuid4()),
"instance_uuid": str(uuid.uuid4()),
"llpdf_version": "llpdf " + llpdf.VERSION,
}
xmp_metadata_template = pkgutil.get_data("llpdf.resources", "xmp_metadata.xml").decode("utf-8")
stream = (xmp_metadata_template % xmp_metadata).encode("utf-8")
content = {
PDFName("/Type"): PDFName("/Metadata"),
PDFName("/Subtype"): PDFName("/XML"),
}
objid = self._pdf.get_free_objid()
pdf_object = PDFObject.create(objid, gennum = 0, content = content, stream = EncodedObject.create(stream, compress = False))
self._pdf.replace_object(pdf_object)
return pdf_object.xref
def get_fontfile_object(self, objid):
content = {
PDFName("/Length1"): len(self._cleardata),
PDFName("/Length2"): len(self._cipherdata),
PDFName("/Length3"): len(self._trailerdata),
}
stream = EncodedObject.create(self._cleardata + self._cipherdata + self._trailerdata, compress = True)
obj = PDFObject.create(objid, 0, content, stream)
return obj
def run(self):
reformatter = ImageReformatter(lossless = True, scale_factor = 1)
for image_obj in self._pdf.image_objects:
if PDFName("/SMask") in image_obj.content:
current_image = self._pdf.get_image(image_obj.xref)
flattened_image = reformatter.flatten(current_image, background_color = self._args.background_color)
flattened_image_obj = PDFObject.create_image(image_obj.xref.objid, image_obj.xref.gennum, flattened_image)
self._pdf.replace_object(flattened_image_obj)
def serialize_xref_object(self, trailer_dict, objid):
offset_width = self._get_offset_width()
content = dict(trailer_dict)
content.update({
PDFName("/Type"): PDFName("/XRef"),
PDFName("/Index"): [ 0, self._max_objid + 1 ],
PDFName("/Size"): self._max_objid + 1,
PDFName("/W"): [ 1, offset_width, 1 ],
})
data = self._serialize_xref_data(offset_width)
return PDFObject.create(objid = objid, gennum = 0, content = content, stream = EncodedObject.create(data))
def _add_color_intent(self, color_profile_xref):
content = [{
PDFName("/Type"): PDFName("/OutputIntent"),
PDFName("/DestOutputProfile"): color_profile_xref,
PDFName("/Info"): b"sRGB IEC61966-2.1",
PDFName("/OutputCondition"): b"sRGB",
PDFName("/OutputConditionIdentifier"): b"Custom",
PDFName("/RegistryName"): b"",
PDFName("/S"): PDFName("/GTS_PDFA1"),
}]
objid = self._pdf.get_free_objid()
pdf_object = PDFObject.create(objid, gennum=0, content=content)
self._pdf.replace_object(pdf_object)
return pdf_object.xref
def _add_color_profile(self):
if self._args.color_profile is None:
profile_data = pkgutil.get_data("llpdf.resources", "sRGB_IEC61966-2-1_black_scaled.icc")
else:
with open(self._args.color_profile, "rb") as f:
profile_data = f.read()
content = {
PDFName("/N"): 3,
PDFName("/Range"): [ 0, 1, 0, 1, 0, 1 ],
}
objid = self._pdf.get_free_objid()
pdf_object = PDFObject.create(objid, gennum = 0, content = content, stream = EncodedObject.create(profile_data))
self._pdf.replace_object(pdf_object)
return pdf_object.xref
def run(self): # Relink the content dictionaries relinked_objects = [ ] for obj in self._pdf: relinked_content = self._relink(obj.content) relinked_xref = self._old_to_new.get(obj.xref, obj.xref) relinked_object = PDFObject.create(relinked_xref.objid, relinked_xref.gennum, relinked_content, obj.stream) relinked_objects.append(relinked_object) # Then delete all old objects for delete_obj_xref in self._old_to_new: self._pdf.delete_object(delete_obj_xref.objid, delete_obj_xref.gennum) # And insert the relinked ones for relinked_object in relinked_objects: self._pdf.replace_object(relinked_object)
def get_font_descriptor_object(self, objid, fontfile_xref):
bbox = self.get_font_bbox()
content = {
PDFName("/Type"): PDFName("/FontDescriptor"),
PDFName("/ItalicAngle"): 0,
PDFName("/FontFile"): fontfile_xref,
PDFName("/FontName"): self.get_font_name(),
PDFName("/Flags"): int(FontDescriptorFlag.Symbolic),
PDFName("/FontBBox"): bbox,
PDFName("/Ascent"): bbox[3],
PDFName("/CapHeight"): bbox[3],
PDFName("/Descent"): bbox[1],
PDFName("/CharSet"): self.charset_string,
PDFName("/MissingWidth"): self.get_missing_width(),
# PDFName("/StemV"): 30, # TODO: Do we need this?
}
obj = PDFObject.create(objid, 0, content)
return obj
def serialize(self, serializer):
header = [ ]
data = bytearray()
for obj in self._contained_objects:
obj_data = serializer.serialize(obj.content)
offset = len(data)
header.append(obj.objid)
header.append(offset)
data += obj_data + b"\n"
header = " ".join(str(value) for value in header)
header = header.encode("utf-8") + b"\n"
full_data = header + data
content = {
PDFName("/Type"): PDFName("/ObjStm"),
PDFName("/N"): self.objects_inside_count,
PDFName("/First"): len(header),
}
return PDFObject.create(objid = self.objid, gennum = 0, content = content, stream = EncodedObject.create(full_data))
def get_font_object(self, objid, fontdescriptor_xref):
widths_dict = self.get_widths_dict()
first_char = min(widths_dict.keys())
last_char = max(widths_dict.keys())
default_width = self.get_missing_width()
widths_array = [ widths_dict.get(i, default_width) for i in range(first_char, last_char + 1) ]
content = {
PDFName("/Type"): PDFName("/Font"),
PDFName("/Subtype"): PDFName("/Type1"),
PDFName("/FirstChar"): first_char,
PDFName("/LastChar"): last_char,
PDFName("/Widths"): widths_array,
PDFName("/Encoding"): {
PDFName("/Type"): PDFName("/Encoding"),
PDFName("/Differences"): build_encoding_array("latin1"),
},
PDFName("/BaseFont"): self.get_font_name(),
PDFName("/FontDescriptor"): fontdescriptor_xref,
}
obj = PDFObject.create(objid, 0, content)
return obj
def run(self):
with open(self._args.embed_payload, "rb") as f:
payload = f.read()
objid = self._pdf.get_free_objid()
self._log.debug(
"Embedding %d bytes payload from file \"%s\" into PDF file as objid %d",
len(payload), self._args.embed_payload, objid)
mtime = os.stat(self._args.embed_payload).st_mtime
mtime_str = datetime.datetime.utcfromtimestamp(mtime).strftime(
"%Y-%m-%dT%H:%M:%SZ")
content = {
PDFName("/PDFMinify.OriginalFilename"):
os.path.basename(self._args.embed_payload).encode(),
PDFName("/PDFMinify.MTime"):
mtime_str.encode(),
PDFName("/PDFMinify.Version"):
llpdf.VERSION.encode(),
}
obj = PDFObject.create(objid=objid, gennum=0, content=content)
obj.set_stream(EncodedObject.create(payload, compress=False))
self._pdf.replace_object(obj)
def run(self):
# Put an ID into the PDF
self._pdf.trailer[PDFName("/ID")] = [os.urandom(16), os.urandom(16)]
# Do not interpolate any image objects
for image_obj in self._pdf.image_objects:
image_obj.content[PDFName("/Interpolate")] = False
# No pages may be transparency groups
for page in self._pdf.pages:
if PDFName("/Group") in page.content:
del page.content[PDFName("/Group")]
# No transparency groups in Form XObjects
for obj in self._pdf:
if (obj.getattr(PDFName("/Type"))
== PDFName("/XObject")) and (obj.getattr(
PDFName("/Subtype")) == PDFName("/Form")) and (
obj.getattr(PDFName("/Group")) is not None):
del obj.content[PDFName("/Group")]
# Add color profile data
color_profile_xref = self._add_color_profile()
# Add color intent object
color_intent_xref = self._add_color_intent(color_profile_xref)
# Add XMP metadata
metadata_xref = self._add_xmp_metadata()
# Set output intent and metadata reference for all catalogs
for obj in self._pdf:
if obj.getattr(PDFName("/Type")) == PDFName("/Catalog"):
obj.content[PDFName("/OutputIntents")] = color_intent_xref
obj.content[PDFName("/Metadata")] = metadata_xref
# Set all annotations with annotation flag "printable" (4)
for obj in self._pdf:
if obj.getattr(PDFName("/Type")) == PDFName("/Annot"):
obj.content[PDFName("/F")] = 4
fixed_descriptors = set()
for obj in list(self._pdf):
if obj.getattr(PDFName("/Type")) == PDFName("/Font"):
font_obj = obj
if font_obj.getattr(
PDFName("/Subtype")) == PDFName("/CIDFontType2"):
# Type2 fonts need to have a CIDtoGIDMap
font_obj.content[PDFName("/CIDToGIDMap")] = PDFName(
"/Identity")
if PDFName("/FontDescriptor") in font_obj.content:
font_descriptor_xref = font_obj.content[PDFName(
"/FontDescriptor")]
if font_descriptor_xref in fixed_descriptors:
continue
fixed_descriptors.add(font_descriptor_xref)
font_descriptor_obj = self._pdf.lookup(
font_descriptor_xref)
if font_obj.getattr(
PDFName("/Subtype")) == PDFName("/Type1"):
# Update Type1 font descriptors with missing CharSet entries
font_file_obj = self._pdf.lookup(
font_descriptor_obj.content[PDFName("/FontFile")])
t1_font = T1Font.from_fontfile_obj(font_file_obj)
font_descriptor_obj.content[PDFName(
"/CharSet")] = t1_font.charset_string
elif font_obj.getattr(
PDFName("/Subtype")) == PDFName("/CIDFontType2"):
# Type2 font descriptors need to have a CIDSet
glyph_count = self.type2_font_glyph_count(
font_obj.content[PDFName("/W")])
full_bytes = glyph_count // 8
set_bits = glyph_count % 8
last_byte = ((1 << set_bits) - 1) << (8 - set_bits)
self._log.debug(
"Assuming CIDSet for %d glyphs of %d full 0xff bytes and a final value of 0x%x.",
glyph_count, full_bytes, last_byte)
cidset_objid = self._pdf.get_free_objid()
stream = (bytes([0xff]) * full_bytes) + bytes(
[last_byte])
pdf_object = PDFObject.create(
cidset_objid,
gennum=0,
content={},
stream=EncodedObject.create(stream))
self._pdf.replace_object(pdf_object)
font_descriptor_obj.content[PDFName(
"/CIDSet")] = pdf_object.xref
