tests: migrate to data-driven fixtures for feature presence

[williballenthin]

closes #2743

This PR refactors our test suite with the primary goal of making the feature tests data-driven; that is, we now have a collection of JSON files that describe where backends should find various features. This enables non-Python implementations of capa to demonstrate that they extract the features we'd expect.

Along the way, I also made some other cleanups. In particular, I tried to simplify fixtures.py a bit. I removed a lot of the symbolic extractors (e.g., pma01_03_extractor) since many were used only once or twice in other modules; I inlined this logic with more explicit code. I figure, more obvious code -> easier to mechanically port to another language. Pytest fixtures as magic test function parameter names is not that easy to reason about...

To the reviewers, I request a thorough review of not just the code style, but of the underlying implementation, design, decisions, etc. I think you should set aside an hour for this.

Here's how I'd recommend that you review these changes:

• read the new test_viv_extractor.py to see how a backend test now looks
• read tests/fixtures/features/README.md and static.json to see how the features are declared
• read fixtures.py for familiarity. don't read the diff, read the final.
• read various diffs when they're reasonable, but i'm afraid there's a bunch of churn, so it may not be easy to parse. focus on the end result, imho, and possibly ask Gemini for any hotspots.

To be clear, there's moderate risk here: I've translated the existing feature test cases to a new format, and I might have missed or corrupted some of them. In some scenarios, I tried to consolidate feature test cases (like when IDA only worked on some samples, and Ghidra on a different set) and it's possible that I mangled the underlying intent. However, I did my best and the test suite probably high quality. I believe we should accept this risk, but you should think about this, too.

Checklist

• This submission includes AI-generated code and I have provided details in the description.

[github-actions]

Please add bug fixes, new features, breaking changes and anything else you think is worthwhile mentioning to the master (unreleased) section of CHANGELOG.md. If no CHANGELOG update is needed add the following to the PR description: [x] No CHANGELOG update needed

CHANGELOG updated or no update needed, thanks! 😄

[mr-tz]

nice work, great step into a data-driven test direction - a few comments to discuss inline

[mr-tz]

looking like it's on a good track!

[williballenthin]

awaiting CI outcome. i have no further pending TODOs, so recommend a full review and decision on whether to merge.

[mr-tz]

Ha! Just went through it so assuming CI passes, we're good to go IMHO.