zlib.error: Error -3 while decompressing data: incorrect data check

[zegrep]

We have do deal with a huge amount of broken PDF files. The creator is "jsPDF 1.x-master". These files are not totally corrupted. It would be nice to get the readable content.
I found a solution on stackoverflow and it works fine for our needs.

PyPDF2/filters.py

    def decompress(data):
        try:
            return zlib.decompress(data)
        except zlib.error:
            return decompress_corrupted(data)

    def decompress_corrupted(data):
        d = zlib.decompressobj(zlib.MAX_WBITS | 32)
        f = StringIO(data)
        result_str = b''
        buffer = f.read(1)
        try:
            while buffer:
                result_str += d.decompress(buffer)
                buffer = f.read(1)
        except zlib.error:
            pass
        return result_str

[zegrep]

zen_of_python_corrupted.pdf

[Pragabhava]

This fix helped me deal with an Error -5 while decompressing data: incomplete or truncated stream for what I think is an improper end of line handling of a byte stream (Windows \r\n vs. Linux \n). I only had to replace f = StringIO(data) for f = BytesIO(data)

[MartinThoma]

It seems as if this issue is gone with PyPDF2==2.3.0:

>>> from PyPDF2 import PdfReader
>>> reader = PdfReader("zen_of_python_corrupted.pdf")
>>> reader.metadata
{'/Producer': 'GPL Ghostscript 9.18', '/CreationDate': "D:20180530133107+02'00'", '/ModDate': "D:20180530133107+02'00'", '/Title': 'zen_of_python.txt', '/Author': '', '/Creator': 'a2ps version 4.14'}

Thank you for reporting it!