Refactor language search autocomplete and add inner word matching

[sbwhitt]

Closes #8146

Refactor the current autocomplete_languages util function to include a limit parameter as well as add prefix matching of inner language name words. For example, '/languages/_autocomplete?q=greek' will match with 'Ancient Greek', 'Modern Greek', and 'Greek' rather than just 'Greek' as it does now.

Technical

Consolidates existing match checks into a more generic matching function that allows for optional translation id.

A limit parameter was also added which controls how many results will be added to the returned iterator. Previously autocomplete_languages would return the total number of matches then was truncated down to a default limit of 5 when called in autocomplete.py. This default limit is preserved, but now it is possible to request more than 5 language matches.

Additionally, a small change was made to the supporting get_languages function which adds an optional limit parameter that defaults to the previous hard-coded value of 1000. This should clarify that a limit is being used within that function without changing any existing behavior.

Testing

Go to /languages/_autocomplete and attempt to search with queries like ?q=greek&limit=10. Verify that at least one inner word of the returned language names match the prefix given (q=greek). Also verify that the limit param accurately controls how many languages are returned.

Screenshot

Stakeholders

@cdrini @tfmorris

[cdrini]

General logic lgtm! A few code reorg suggestions.

[cclauss]

language is a dict...

[hornc]

@cdrini I think this is good to merge now. I thought there was a problem with case matching, but I think that was just some form of response caching when I tested it locally. Testing this with the full language data in the staging env would be good.

[cdrini]

Ok lgtm! Team effort on this one thank you both @sbwhitt and @hornc ! Thanks for your patience, I was having trouble determining the refactor needed here, but I think I got it.

I did a refactor to remove the _matches_lang_name method since that was making this rather difficult for me to follow. The method was introducing a confusing bit of misdirection in the logic. I instead introduced a more general method, word_prefix_match and a more specific method, get_names_to_try, for getting the various language names to try.

Appears to be working like a charm!