Max Bachmann

@SolebaySharp why do you try to install both `Levenshtein` and `python-Levenshtein`? It would be enough to install `Levenshtein`, which probably provides a wheel for your platform.

@SolebaySharp did both approaches fail?

This is caused by the new C module `dtw` which is part of rhasspywake_raven. This module is compiled and installed next to the Python files. However when running `bin/rhasspy-wake-raven` it...

Yes, I guess this should work around the problem.

You can directly use the implementation in [`RapidFuzz`](https://github.com/rapidfuzz/RapidFuzz/tree/main) which is the basis for this library. The wikipedia describes both the damerau levenshtein distance and the optimal string alignment distance. Both...

`python-Levenshtein` is not maintained anymore. I maintain a fork, which includes wheels: https://github.com/maxbachmann/Levenshtein

You can use the Levenshtein fork for this ``` pip install Levenshtein ``` Which allows the following: ``` from Levenshtein import distance distance(s1, s2, weights=(insertion, deletion, substitution)) ``` A future...

> Is there any reference on which of these functions are fastest vs. slowest? We need to sort through about 10,000 search items very quickly. Or I can experiment with...

> It's only one query over 10k items. But fairly long items and has to be The original library python-Levenshtein Python 3.5 would be a good start. I'm not sure...

Closing this, since it is tracked as part of #241