OPENNLP-1935: Regex removal (10/10): Forbid java.util.regex in production code

[krickert]

Part 10 of the regex removal series, kept as a draft until the remaining classes are converted. Continues #1282 on an apache branch.

A forbiddenapis execution checks the production classes of every module against dev/forbidden-regex.txt: java.util.regex, String.split, String.matches, String.replaceAll, String.replaceFirst and Scanner.useDelimiter. It runs at process-classes, before the tests, and its message names the StringUtil replacements. Tests are not checked, so test code can keep using regular expressions.

Exceptions: a member with a documented public contract that is a user-supplied regular expression is marked with @SuppressForbidden(reason). Only RegexNameFinder and RegexNameFinderFactory have it. A private pattern is converted, not annotated, so the report of the check is the to-fix list for the rest of the series. CONTRIBUTING.md has the rule.

Also in this PR: StringUtil has one non-empty split loop behind splitNonEmpty, splitOnUnicodeWhitespace and the mode-aware splitOnWhitespace, replacing the copies in EventFields, MascIdentifiers, MecabDictionary and HunspellDictionary.

Report on main today: 104 findings in 13 classes. The tokenizer factories and TokenizerME, the parser family, NameSample, FeatureGeneratorUtil, SpellCheckingCharSequenceNormalizer, and the rest of ConlluStream, DownloadUtil and TokenSampleStream is #1280.

OPENNLP-1935

[krickert]

Rebased main, updated description.

AI assisted summary (reviewed and edited).

Rebased onto main after #1313 and #1280. The report is down to 104 findings in 13 classes: the tokenizer family, the parser family, NameSample, FeatureGeneratorUtil, SpellCheckingCharSequenceNormalizer, and the rest of ConlluStream, DownloadUtil and TokenSampleStream, of which #1314 takes five.

[krickert]

Jumped the gun on making this green - after 547 is done, this will be green.

[jzonthemtn]

Jumped the gun on making this green - after 547 is done, this will be green.

Is this blocked by #1236?

[krickert]

@jzonthemtn yup.. I can make it 10/10 a draft for now. It needs 547 (and it's ticket stack I'd rebase right after that would help 547 - it's made to be stacked w/it)...

So order would be:

#1236
#1323 (after rebase - which would happen immediately after 1236)
#1348

[krickert]

Jumped the gun on making this green - after 547 is done, this will be green.

Is this blocked by #1236?

Also - NBD but there's still about 30 or so places to clean up in the codebase for the regex stuff - not needed for the 3.0 release (the surface was a lot bigger than I thought).

It's just something we should just chip away at. The ones we have here are the hot path for most. The other tickets expose some minor API changes made to remove Pattern from constructors.

[krickert]

@jzonthemtn going to put this in draft temporarily - #1360 has a nit where we would need to add to this list after it merges too. To minimize churn, I'll keep this in draft until #1236 and #1360 both complete.

[rzo1]

lgtm. Think it is a major step forward. Thx @krickert and team.

[krickert]

Thank you! I've also identified the remaining regex brats, we're down to about 3ish PRs for us to be regex free :) I have branches for these, but still cleaning to draft them into PR