Add entity synonym fuzzy match

[odannyc]

Proposed changes:

• Add fuzzy matching to ner_synonyms

Status (please check what you already did):

• made PR ready for code review
• added some tests for the functionality
• updated the documentation
• updated the changelog

[CLAassistant]

All committers have signed the CLA.

[akelad]

Thanks for submitting this PR, we'll give it a review as soon as possible

[MetcalfeTom]

Linking this to the original issue: #1626

[tmbo]

We can not depend on fuzzywuzzy because as far as I can see it is released under the GPL license, which would require us to relicense NLU under GPL as well

[MetcalfeTom]

@odannyc Any chance you could get this working with a library under an Apache 2.0 license?

[odannyc]

@tmbo @MetcalfeTom

Before I dive too deep into implementing another library, I found this which has the MIT license.

Is that fine?

[odannyc]

@MetcalfeTom

I have updated this PR to use a package that doesn't have the GPL license.
Let me know if it looks good now.

Thanks!

[odannyc]

@akelad @MetcalfeTom

Still waiting on a response for this.

Thanks!

[akelad]

@odannyc we appreciate your hard work, but this requires some further discussion from our side. We'll get back to you once we've discussed it and ran some experiments ourselves.

[codeclimate]

Code Climate has analyzed commit 9982fd7 and detected 0 issues on this pull request.

View more on Code Climate.

[odannyc]

@akelad

Any news?

[akelad]

not yet sorry, we haven't had time to properly look into it yet

[tmbo]

What we still need to validate: earlier tests showed that fuzzy matching is quite slow. To avoid introducing a component that can't be used in production, we need to validate whether we can really on the chosen matching library to deliver good performance.

[tmbo]

I'll close this for now as it is quite outdated. But happy to reopen if @Ghostvv thinks this makes sense from a entity matching accuricy standpoint.