Fix: catch RuntimeError alongside ImportError for the optional numpy import

[stumpylog]

Fixes #1908

The optional numpy import in image_parsing.py only caught ImportError. On CPUs numpy's manylinux wheel doesn't support (below its x86-64-v2 baseline as of numpy 2.4.0), import numpy raises RuntimeError instead, which propagated out of import fpdf entirely and crashed every caller, even ones that never touch the numpy-guarded code paths. Both existing call sites (pack_codes_into_bytes(), _has_alpha()) already guard on numpy is not None, so treating RuntimeError the same as ImportError is enough to restore the intended "numpy unavailable, fall back" behavior.

Checklist:

• A unit test is covering the code added / modified by this PR - N/A, this only widens an except clause around an import statement; reproducing it requires either genuinely unsupported hardware or monkeypatching builtins.__import__, none of which seem feasible or really useful
• In case of a new feature, docstrings have been added, with also some documentation in the docs/ folder
• A mention of the change is present in CHANGELOG.md
• This PR is ready to be merged

By submitting this pull request, I confirm that my contribution is made under the terms of the GNU LGPL 3.0 license.

[andersonhc]

@allcontributors please add @stumpylog for bug, code

[allcontributors]

@andersonhc

I've put up a pull request to add @stumpylog! 🎉

[andersonhc]

Good to merge. Thank you.

[stumpylog]

Thanks for the merge. If you happen to get time to cut a release, this would be great to include for our users soon.

[shamoon]

Hi @andersonhc im one of the other paperless devs, thanks for your work on this project. Is there any chance you would be planning (or willing) to make a release with this change soon? Thanks again

[andersonhc]

Hi @andersonhc im one of the other paperless devs, thanks for your work on this project. Is there any chance you would be planning (or willing) to make a release with this change soon? Thanks again

I am preparing a new release today. Probably a couple hours it should be in pypi.

[shamoon]

Oh wow thank you!!