improve performance of find_similar_names

[maxbachmann]

on most platforms this improves the performance by 30-40x by using a highly optimized C extension.
On unsupported platforms falls back to a pure Python implementation, which still improves performance by 2x.

I am not really familiar with poetry. I used:

poetry add rapidfuzz
poetry remove pylev

However this removed files for a lot of dependencies, which seems wrong. What is the correct process to replace a dependency?

[radoering]

I am not really familiar with poetry. I used:

poetry add rapidfuzz
poetry remove pylev

However this removed files for a lot of dependencies, which seems wrong. What is the correct process to replace a dependency?

Seems you have experienced python-poetry/poetry#5970. Update to poetry 1.1.14, clear your cache (poetry cache clear pypi --all), restore the lock file and do the steps again (add and remove should be fine).

[maxbachmann]

Seems you have experienced python-poetry/poetry#5970. Update to poetry 1.1.14, clear your cache (poetry cache clear pypi --all), restore the lock file and do the steps again (add and remove should be fine).

thanks this worked

[maxbachmann]

Not sure whats up with the Python3.11 failures, but I don't think they are caused by the change.

[maxbachmann]

restarting the CI today solved the build issues, so they were unrelated.

[maxbachmann]

This is now ready for review.

[Secrus]

@maxbachmann would you mind writing a simple test for that? I see there are no tests for _utils module at the moment and this could be a good start.

[maxbachmann]

Sure I can add a test

[maxbachmann]

I added a simple test for find_similar_names