def pdf_to_text(pdf):
"""Return extracted text from PDF.
Warning: This function can be slow... up to 300ms per page
This function does not perform optical character recognition.
Args:
pdf: bytestring of PDF file
Returns:
str of text extracted from `pdf` contents.
"""
# make input and output buffers
in_buffer = StringIO.StringIO(pdf)
out_buffer = StringIO.StringIO()
# configure pdf parser
parser = pdfparser.PDFParser(in_buffer)
doc = pdfparser.PDFDocument()
parser.set_document(doc)
doc.set_parser(parser)
doc.initialize(password='')
rsrcmgr = pdfinterp.PDFResourceManager()
laparams = layout.LAParams()
# convert pdf to text
device = converter.TextConverter(rsrcmgr,
outfp=out_buffer,
codec='utf-8',
laparams=laparams)
interpreter = pdfinterp.PDFPageInterpreter(rsrcmgr, device)
for page in doc.get_pages():
interpreter.process_page(page)
return out_buffer.getvalue()
def parsePDFPorto(fileStream,OutType):
rsrcmgr = PDFResourceManager()
retstr = StringIO()
codec = 'utf-8'
laparams = LAParams()
if (OutType=="HTML"):
device = converter.HTMLConverter(rsrcmgr, retstr, codec=codec, laparams=laparams)
if (OutType=="Text"):
device = converter.TextConverter(rsrcmgr, retstr, codec=codec, laparams=laparams)
if (OutType=="XML"):
device = converter.XMLConverter(rsrcmgr, retstr, codec=codec, laparams=laparams)
interpreter = PDFPageInterpreter(rsrcmgr, device)
parser = PDFParser(fileStream)
for page in PDFPage.get_pages(fileStream):
interpreter.process_page(page)
fileStream.close()
device.close()
str = retstr.getvalue()
retstr.close()
return str.decode('utf-8')
def pdf_miner(from_file, to_txt):
log.debug('trying with pdfminer')
pdf = codecs.open(from_file.path, mode='rb')
output = codecs.open(to_txt.path, mode='wb')
try:
resourceman = pdfinterp.PDFResourceManager()
device = converter.TextConverter(
resourceman, output, laparams=layout.LAParams())
interpreter = pdfinterp.PDFPageInterpreter(resourceman, device)
for page in pdfpage.PDFPage.get_pages(pdf):
interpreter.process_page(page)
output.close()
device.close()
pdf.close()
except Exception, e:
log.critical(e)
return
def scan(self, file_object, options):
extract_text = options.get("extract_text", False)
file_limit = options.get("limit", 2000)
self.metadata["total"] = {"objects": 0, "extracted": 0}
extracted_objects = set()
try:
with io.BytesIO(file_object.data) as pdf_object:
parsed_pdf = pdfparser.PDFParser(pdf_object)
pdf_document = pdfdocument.PDFDocument(parsed_pdf)
self.metadata.setdefault("annotatedUris", [])
for xref in pdf_document.xrefs:
for object_id in xref.get_objids():
self.metadata["total"]["objects"] += 1
try:
object = pdf_document.getobj(object_id)
if isinstance(object, dict):
for (key, value) in object.items():
if key in ["AA", "OpenAction"]:
file_object.flags.append(f"{self.scanner_name}::auto_action")
if key in ["JS", "Javascript"]:
file_object.flags.append(f"{self.scanner_name}::javascript_embedded")
try:
if key == "A":
uri = value.get("URI")
if (uri is not None and
uri not in self.metadata["annotatedUris"]):
self.metadata["annotatedUris"].append(uri)
except AttributeError:
pass
if self.metadata["total"]["extracted"] >= file_limit:
continue
if isinstance(object, pdftypes.PDFStream):
try:
child_filename = f"{self.scanner_name}::object_{object_id}"
child_fo = objects.StrelkaFile(data=object.get_data(),
filename=child_filename,
depth=file_object.depth + 1,
parent_uid=file_object.uid,
root_uid=file_object.root_uid,
parent_hash=file_object.hash,
root_hash=file_object.root_hash,
source=self.scanner_name)
if object_id not in extracted_objects:
self.children.append(child_fo)
extracted_objects.add(object_id)
self.metadata["total"]["extracted"] += 1
except TypeError:
file_object.flags.append(f"{self.scanner_name}::type_error_{object_id}")
except struct.error:
file_object.flags.append(f"{self.scanner_name}::struct_error_{object_id}")
except ValueError:
file_object.flags.append(f"{self.scanner_name}::value_error_{object_id}")
except pdftypes.PDFObjectNotFound:
file_object.flags.append(f"{self.scanner_name}::object_not_found_{object_id}")
except pdftypes.PDFNotImplementedError:
file_object.flags.append(f"{self.scanner_name}::not_implemented_error_{object_id}")
except pdftypes.PSSyntaxError:
file_object.flags.append(f"{self.scanner_name}::ps_syntax_error_{object_id}")
if extract_text:
rsrcmgr = pdfinterp.PDFResourceManager(caching=True)
retstr = io.StringIO()
la_params = layout.LAParams(detect_vertical=True,
char_margin=1.0,
line_margin=0.3,
word_margin=0.3)
device = converter.TextConverter(rsrcmgr, retstr,
codec="utf-8",
laparams=la_params)
interpreter = pdfinterp.PDFPageInterpreter(rsrcmgr, device)
for page in pdfpage.PDFPage.get_pages(pdf_object, set()):
try:
interpreter.process_page(page)
except struct.error:
file_object.flags.append(f"{self.scanner_name}::text_struct_error")
pdf_object_text = retstr.getvalue()
child_filename = f"{self.scanner_name}::text"
child_fo = objects.StrelkaFile(data=pdf_object_text,
filename=child_filename,
depth=file_object.depth + 1,
parent_uid=file_object.uid,
root_uid=file_object.root_uid,
parent_hash=file_object.hash,
root_hash=file_object.root_hash,
source=self.scanner_name)
self.children.append(child_fo)
file_object.flags.append(f"{self.scanner_name}::extracted_text")
device.close()
retstr.close()
except IndexError:
file_object.flags.append(f"{self.scanner_name}::index_error")
except pdfdocument.PDFEncryptionError:
file_object.flags.append(f"{self.scanner_name}::encrypted_pdf")
except pdfparser.PDFSyntaxError:
file_object.flags.append(f"{self.scanner_name}::pdf_syntax_error")
except psparser.PSEOF:
file_object.flags.append(f"{self.scanner_name}::ps_eof")
except psparser.PSSyntaxError:
file_object.flags.append(f"{self.scanner_name}::ps_syntax_error")
def scan(self, data, file, options, expire_at):
extract_text = options.get('extract_text', False)
file_limit = options.get('limit', 2000)
self.event['total'] = {'objects': 0, 'extracted': 0}
extracted_objects = set()
try:
with io.BytesIO(data) as pdf_io:
parsed = pdfparser.PDFParser(pdf_io)
pdf = pdfdocument.PDFDocument(parsed)
self.event.setdefault('annotated_uris', [])
for xref in pdf.xrefs:
for object_id in xref.get_objids():
self.event['total']['objects'] += 1
try:
object = pdf.getobj(object_id)
if isinstance(object, dict):
for (key, value) in object.items():
if key in ['AA', 'OpenAction']:
self.flags.append('auto_action')
if key in ['JS', 'Javascript']:
self.flags.append(
'javascript_embedded')
try:
if key == 'A':
uri = value.get('URI')
if uri not in self.event[
'annotated_uris']:
self.event[
'annotated_uris'].append(
uri)
except AttributeError:
pass
if self.event['total']['extracted'] >= file_limit:
continue
if isinstance(object, pdftypes.PDFStream):
try:
if object_id not in extracted_objects:
extract_file = strelka.File(
name=f'object_{object_id}',
source=self.name,
)
for c in strelka.chunk_string(
object.get_data()):
self.upload_to_cache(
extract_file.pointer,
c,
expire_at,
)
self.files.append(extract_file)
self.event['total']['extracted'] += 1
extracted_objects.add(object_id)
except TypeError:
self.flags.append('type_error_{object_id}')
except struct.error:
self.flags.append(
'struct_error_{object_id}')
except ValueError:
self.flags.append('value_error_{object_id}')
except pdftypes.PDFObjectNotFound:
self.flags.append('object_not_found_{object_id}')
except pdftypes.PDFNotImplementedError:
self.flags.append(
'not_implemented_error_{object_id}')
except psparser.PSSyntaxError:
self.flags.append('ps_syntax_error_{object_id}')
if extract_text:
rsrcmgr = pdfinterp.PDFResourceManager(caching=True)
retstr = io.StringIO()
la_params = layout.LAParams(
detect_vertical=True,
char_margin=1.0,
line_margin=0.3,
word_margin=0.3,
)
device = converter.TextConverter(
rsrcmgr,
retstr,
codec='utf-8',
laparams=la_params,
)
interpreter = pdfinterp.PDFPageInterpreter(rsrcmgr, device)
for page in pdfpage.PDFPage.get_pages(data, set()):
try:
interpreter.process_page(page)
except struct.error:
self.flags.append('text_struct_error')
extract_file = strelka.File(
name='text',
source=self.name,
)
for c in strelka.chunk_string(retstr.getvalue()):
self.upload_to_cache(
extract_file.pointer,
c,
expire_at,
)
self.files.append(extract_file)
self.flags.append('extracted_text')
device.close()
retstr.close()
except IndexError:
self.flags.append('index_error')
except pdfdocument.PDFEncryptionError:
self.flags.append('encrypted_pdf')
except pdfparser.PDFSyntaxError:
self.flags.append('pdf_syntax_error')
except psparser.PSEOF:
self.flags.append('ps_eof')
except psparser.PSSyntaxError:
self.flags.append('ps_syntax_error')
