ENH: Add basic support for JBIG2 by using jbig2dec

[stefan6419846]

Closes #1989.

[stefan6419846]

This requires at least version 0.15 of jbig2dec, which is not properly documented in the changelog and required me to look through the commits to find the offending change and its tags: ArtifexSoftware/jbig2dec@c0353a0

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 96.69%. Comparing base (77238e0) to head (83d81e0).
Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3163      +/-   ##
==========================================
+ Coverage   96.68%   96.69%   +0.01%     
==========================================
  Files          53       53              
  Lines        8977     9024      +47     
  Branches     1664     1674      +10     
==========================================
+ Hits         8679     8726      +47     
  Misses        176      176              
  Partials      122      122              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[j-t-1]

Good new functionality.

The rationale for having decode_parms is so external code has a uniform interface for filters. Previously I found this very confusing, partly due to not knowing the PDF specification well. This has two negatives:

1. We are relying on comments to explain an argument not being used (rather than this being fully in code by not having the argument, and then not needing the comment).
2. Coding to external library users who may not even use a uniform interface, complicates our interfaces by having an unused argument. Sometimes this is mentioned in code reviews.

Because this function has kwargs, decode_parms could be swallowed by kwargs, that is remove decode_parms argument when not used.

[stefan6419846]

I have more or less copied the existing filters and used the same approach here. For now, I would stay with the current approach and maybe discuss this in a dedicated issue, although there probably will not be much input about this and there is no real urgency to change the current approach in this regard.

[j-t-1]

Appreciate all the added functionality.

I have more or less copied the existing filters and used the same approach here. For now, I would stay with the current approach and maybe discuss this in a dedicated issue, although there probably will not be much input about this and there is no real urgency to change the current approach in this regard.

Agree low urgency, and understandable good approach.

Just opened the specification and of the ten filters, four do not have parameters:

ASCIIHexDecode
ASCII85Decode
RunLengthDecode
JPXDecode

So I was mistaken about JBIG2Decode parameters, and think that keeping the argument is the right thing to do here for this PR (even though it is not used).

So now we have three cases:

1. decode_parms used.
2. decode_parms not used (ASCIIHexDecode, ASCII85Decode, RunLengthDecode, JPXDecode).
3. decode_parms used by PDF specification and not used by pypdf (JBIG2Decode).

I may raise an issue as suggested, as think that removing the argument when not used by the PDF specification (2) will simplify this for pypdf developers and follow the design principle of moving semantics out of comments into code.

[DaanSelen]

What is the status?

[stefan6419846]

As mentioned in #1989 (comment), I am rather busy at the moment, thus bigger changes being a bit delayed. I hope to get this done this month.

[DaanSelen]

As mentioned in #1989 (comment), I am rather busy at the moment, thus bigger changes being a bit delayed. I hope to get this done this month.

Hello @stefan6419846 "this month" is wonderfully to hear, not long-term, for now I'll just make a catch and skip those images. Thank you for your work. If I can help in a way. Let me know.

[stefan6419846]

As usual: You are of course invited to install pypdf from my branch and check if everything works as intended or if there are bigger issues.

[DaanSelen]

As usual: You are of course invited to install pypdf from my branch and check if everything works as intended or if there are bigger issues.

I hope to get around to that, I must admit that this library is totally new so my feedback might come too late. Thanks in advance.

[stefan6419846]

Currently, this fails with /JBIG2Globals containing shared/global segments. I have a local patch, but still need to prepare test files I can share.