Improve Python usage harness

[jribbens]

The current Python usage harness:

• is very, very, very slow;
• matches the patterns case-insensitively, which is not what the other languages' harnesses do, and is not how the patterns appear to be intended to be used;
• is missing the matching_crawlers() function that the Go harness provides.

This patch fixes all these issues. In some simple tests, the updated is_crawler() function is well over a hundred times faster than the current version.

[monperrus]

Thanks a lot for the contribution. Would you also be able to contribute with a test case run in CI? That would be awesome!

[jribbens]

Ok I've added some simple tests, updated pyproject.toml a bit (dev dependencies are now in it rather than a separate requirements.txt, and it defines a project homepage), and changed the CI workflow so it uses pip install -e .[dev] so that the test file can import crawleruseragents in order to test it.

[jribbens]

If you're interested, I also reworked the Python harness to be completely type-annotated and other fixes so it can be fully type-checked and linted with 'mypy --strict', 'ruff check' and 'ruff format': https://github.com/jribbens/crawler-user-agents/compare/master...jribbens:crawler-user-agents:full-typing

This involves considerably more changes to the code though, of course, and it means it's only compatible with Python 3.8 or later (not sure what version you were targeting previously).

[monperrus]

thanks a lot @jribbens