fix(text): unicode support and word splitting according to case

[guy-borderless]

The current implementation of splitToWords() only supports Latin letters. This PR adds a unicode-compatible implementation + support for WordSpliting according to case.

[iuioiua]

Nice. Just a few nits on the first pass. Thank you.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 96.29%. Comparing base (a8a637f) to head (90c7286).
Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #5447      +/-   ##
==========================================
- Coverage   96.30%   96.29%   -0.01%     
==========================================
  Files         465      465              
  Lines       37705    37711       +6     
  Branches     5561     5560       -1     
==========================================
+ Hits        36312    36315       +3     
- Misses       1351     1354       +3     
  Partials       42       42              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[kt3k]

By this change, the result of toSnakeCase("Camel1Case") becomes camel_1_case (It is camel1_case in main), while the result of toSnakeCase("camel1Case") is kept as camel1_case. Is this change reasonable/acceptable?

[guy-borderless]

By this change, the result of toSnakeCase("Camel1Case") becomes camel_1_case (It is camel1_case in main), while the result of toSnakeCase("camel1Case") is kept as camel1_case. Is this change reasonable/acceptable?

Looking at a more semantic example like iAte2Cookies the change does seem reasonable to me.

[kt3k]

Looking at a more semantic example like iAte2Cookies the change does seem reasonable to me.

I agree with this.

In that case, probably we should return has_2_cookies for toSnakeCase("has2Cookies") for consistency (It currently returns has2_cookies). I'll work on this.

[guy-borderless]

I made some changes to better handle Acronyms followed by a capitalized word, like in URLPattern (before it would have been ["URLP", "aattern"], now return ["URL", "Pattern"].
But went ahead and simplified the implementation to only use one input.match() call

I think an added benefit of doing everything with one input.match may be that if we will want to export splitToWords it'll be simple to allow the user to extend the word regexp (including with non-alphanumeric patterns like a URL or an email address) and also it now returns an iterator.

[kt3k]

LGTM

[kt3k]

I made some changes to better handle Acronyms followed by a capitalized word, like in URLPattern (before it would have been ["URLP", "aattern"], now return ["URL", "Pattern"].

This looks a great improvement to me. Nice!