File causes loop method call between functions extract_xform_text and _extract_text

[VBobCat]

While reading a certain file, my program exits without any exception being raised.

I investigated the issue and it seems the cause is functions extract_xform_text and _extract_text in _page.py call each other in a neverending loop.

Environment

Which environment were you using when you encountered the problem?

Python 3.10.5
Windows-10-10.0.19044-SP0
PyPDF2 2.1.0

Code

This is a minimal, complete example that shows the issue:

My code (that uses PyPDF2) is this:

def pdf_to_text(filename):
    with open(filename, 'rb') as pdf_file_object:
        # try:
            reader = PyPDF2.PdfFileReader(pdf_file_object, strict=False)
            num_pages = reader.numPages
            if num_pages:
                page_texts = []
                for i in range(num_pages):
                    page = reader.getPage(i)
                    page_text = page.extract_text()
                    page_texts.append(page_text)
                return ' '.join(page_texts)

I put a breakpoint in extract_xform_text and it receives these three parameters (self, xform, space_width):

{
    'self': {
        '/Contents': IndirectObject(2, 0),
        '/CropBox': [0, 0, 595.56, 842.04],
        '/Group': {
            '/CS': '/DeviceRGB',
            '/S': '/Transparency',
            '/Type': '/Group'
        },
        '/MediaBox': [0, 0, 595.56, 842.04], '/Parent': IndirectObject(34, 0),
        '/Resources': {
            '/ExtGState': {
                '/GS0': IndirectObject(14, 0)
            },
            '/Font': {
                '/TT0': IndirectObject(54, 0),
                '/TT1': IndirectObject(57, 0)
            },
            '/ProcSet': ['/PDF', '/Text'],
            '/XObject': {
                '/Fm0': IndirectObject(7, 0)
            }
        },
        '/Rotate': 0,
        '/StructParents': 1,
        '/Tabs': '/S',
        '/Type': '/Page'
    },
    'xform': {
        '/ADBE_FillSign': {
            '/Subtype': '/page',
            '/Type': '/FillSignData'
        },
        '/BBox': [195.809, 527.097, 407.639, 553.999], 
        '/FormType': 1, 
        '/Matrix': [1, 0, 0, 1, 0, 0],
        '/Resources': {
            '/XObject': {
                '/Fm0': IndirectObject(6, 0)
            }
        }, 
        '/Subtype': '/Form', 
        '/Type': '/XObject'
    },
    'space_width': 200.0
}

PDF

I am sorry but I'm unable to share the very PDF file, because it contains sensitive information.

[pubpub-zz]

@VBobCat
To try to de-sensibilize the information, can you try to just cut off the page. if still not possible can you use _debug_for_extract() to extract the document structure (remove the TJ/Tj will remove the text)

[MartinThoma]

I think https://corpora.tika.apache.org/base/docs/govdocs1/998/998167.pdf might fall into the same issue:

  File "/home/moose/Github/py-pdf/PyPDF2/PyPDF2/_page.py", line 1295, in extract_xform_text
    return self._extract_text(xform, self.pdf, space_width, None)
  File "/home/moose/Github/py-pdf/PyPDF2/PyPDF2/_page.py", line 1256, in _extract_text
    text = self.extract_xform_text(xobj[operands[0]], space_width)
  File "/home/moose/Github/py-pdf/PyPDF2/PyPDF2/_page.py", line 1295, in extract_xform_text
    return self._extract_text(xform, self.pdf, space_width, None)
  File "/home/moose/Github/py-pdf/PyPDF2/PyPDF2/_page.py", line 1256, in _extract_text
    text = self.extract_xform_text(xobj[operands[0]], space_width)
  File "/home/moose/Github/py-pdf/PyPDF2/PyPDF2/_page.py", line 1295, in extract_xform_text
    return self._extract_text(xform, self.pdf, space_width, None)
  File "/home/moose/Github/py-pdf/PyPDF2/PyPDF2/_page.py", line 1256, in _extract_text
    text = self.extract_xform_text(xobj[operands[0]], space_width)
  File "/home/moose/Github/py-pdf/PyPDF2/PyPDF2/_page.py", line 1295, in extract_xform_text
    return self._extract_text(xform, self.pdf, space_width, None)
  File "/home/moose/Github/py-pdf/PyPDF2/PyPDF2/_page.py", line 1131, in _extract_text
    content = ContentStream(content, pdf, "bytes")
  File "/home/moose/Github/py-pdf/PyPDF2/PyPDF2/generic.py", line 1122, in __init__
    self.__parseContentStream(stream_bytes)
  File "/home/moose/Github/py-pdf/PyPDF2/PyPDF2/generic.py", line 1153, in __parseContentStream
    operands.append(read_object(stream, None, self.forced_encoding))
  File "/home/moose/Github/py-pdf/PyPDF2/PyPDF2/generic.py", line 1259, in read_object
    elif idx == 2:
KeyboardInterrupt

[pubpub-zz]

https://corpora.tika.apache.org/base/docs/govdocs1/998/998167.pdf
is a good sample I think the issue is because of we are looking for a xobject within the xobject with the same name. I was not looking in the good place :)

[pubpub-zz]

this should be closed by #969 (https://github.com/py-pdf/PyPDF2/releases/tag/2.2.0)
@VBobCat can you please confirm your file gets through

[MartinThoma]

I'm rather certain that the issue was solved. Please let us know if that is not the case!