[DA-3868] - Try to decompressing data streams first

[Itisfilipe]

This change will force zlib decompressing data streams that won’t fit into memory first and then fall back to the usual decompress method.

It is intended to fix those events https://sentry.drchrono.com/drchrono/drchrono-web/issues/245739 with the message Error -5 while decompressing data: incomplete or truncated stream and it is based on py-pdf#88

You can get a broken PDF by going to Sentry event and getting the ConsentFormSignature id from the stack trace and then under prod or staging djsp you get the signed URL for the file.

cfs = ConsentFormSignature.objects.get(pk=XXXXXX)
cf = cfs.consent_form
print cf.uri.url

Download the file to your computer and use this script to test it

from PyPDF2 import PdfFileReader, PdfFileWriter

cf = PdfFileReader("broken-pdf.pdf")

output = PdfFileWriter()

num_pages = cf.getNumPages()

for i in range(num_pages):
    output.addPage(cf.getPage(i))

output_stream = open("fixedFile3.pdf", "wb")
output.write(output_stream)
output_stream.close()

It Should blow up with this error without this change error: Error -5 while decompressing data: incomplete or truncated stream but then work with this change.