MAINT: Throw PdfReadError if Trailer can't be read

[ediamondscience]

… Added myself to contributor list.

In reference to #1279 I've modified _reader.py to throw a PdfReadError if the trailer doesn't exist, can't be read, or does not point to the object containing the metadata for the document.

Three pdf files were uploaded in relation to aforementioned issue, pca_var.pdf contained no trailer, and the other two had corrupted declarations that PyPDF2 could not read. I'll include them here for those interested.

From Rotary Potentiometer.pdf:
trailer^M <</Size 267/Prev 2815238/Root 176 0 R/Info 174 0 R/ID[<8CC42D96833D2A438E62CB60D00F610F><5C480BCE388CB6408C83C951FDD0DA22>]>>^M startxref^M 0^M %%EOF^M

From Status_v1_Reviewers-Guide.pdf:
trailer
/Size
447
/Root
4
0
R
/Info
5
0
R
startxref
6128518
(sorry for using multiple code blocks, but I think it's important to show that there are line breaks for each of these lines in my terminal)

We should probably figure out why these trailers are displayed like this. If I had to guess it might be because of a different encoding for line breaks. If we can update the way that we read the trailer to account for this we might be able to read useful information out of more files.

All three files from the issue should now throw an error when the test code is run. This should close #1279

[codecov]

Codecov Report

Merging #1298 (4f31b64) into main (4e7602a) will increase coverage by 0.00%.
The diff coverage is 100.00%.

@@           Coverage Diff           @@
##             main    #1298   +/-   ##
=======================================
  Coverage   95.02%   95.02%           
=======================================
  Files          30       30           
  Lines        4986     4988    +2     
  Branches     1025     1026    +1     
=======================================
+ Hits         4738     4740    +2     
  Misses        141      141           
  Partials      107      107           

Impacted Files	Coverage Δ
PyPDF2/_reader.py	91.23% <100.00%> (+0.02%)	⬆️

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

[MartinThoma]

Can you please add a test that covers this? You mentioned that https://github.com/p-sampson-II/SEEDLab-TAs-2022/blob/378e6f583ed9de274a8bda9240dfa7aa6872a733/Alex%20Curtis/Datasheets/Rotary%20Potentiometer.pdf shows the issue. You can use the file via:

url = "https://github.com/p-sampson-II/SEEDLab-TAs-2022/blob/378e6f583ed9de274a8bda9240dfa7aa6872a733/Alex%20Curtis/Datasheets/Rotary%20Potentiometer.pdf"
name = "github-Rotary-Potentiometer.pdf"
reader = PdfReader(BytesIO(get_pdf_from_url(url, name=name)))

[MartinThoma]

Thank you for your work on this one!

I've just noticed that you've added resources/unreadablemetadata.pdf which has 2MB. In order to keep this repository small (and thus easy to clone, also with bad connections), I would like this file to be in https://github.com/py-pdf/sample-files/ .

Could you please add the file there? (I can also do it + use the "Contributed-by" feature of Github to attribute it to you)

[MartinThoma]

Thank you for your work on this one!

I've just noticed that you've added resources/unreadablemetadata.pdf which has 2MB. In order to keep this repository small (and thus easy to clone, also with bad connections), I would like this file to be in https://github.com/py-pdf/sample-files/ .

Could you please add the file there? (I can also do it + use the "Contributed-by" feature of Github to attribute it to you)

[ediamondscience]

@MartinThoma I've added a pull request to get that file into the correct repository and updated the file location in test_reader.py

The commit may fail some tests in this repository before the sample files are merged, but I've tested it on my own rig here with pytest and it should work perfectly after the merge.

[MartinThoma]

I was just blocked at work and added the updated submodule to PyPDF2 :-)

[MartinThoma]

Could you please do git rm resources/unreadablemetadata.pdf?

[ediamondscience]

Looks like the the reader causes some issues if the metadata is unreadable. Is it the right move to throw a PdfReadError for this situation?

[MartinThoma]

Sorry, my mistake. I'll fix the test tomorrow 🙈