DATA-5365: Fix case/punctuation-sensitive committee matching for USA events

[sinchana00]

Replaced the formatting-guessing heuristic with normalized comparison on both sides:

1. Expand & → and
2. Strip all punctuation
3. Collapse whitespace
4. Uppercase
   This normalization is applied identically to the incoming scraped name (in Python) and to Organization.name + every entry in Organization.other_names (via a Postgres regexp_replace/jsonb_array_elements SQL expression), so matches no longer depend on guessing exact formatting conventions.

Testing

Added test_related_committee_event_other_names_case_mismatch unit test.
All tests passed (351/351)
Manually verified against real scraped event data for TX and IL (different scrapers, different naming conventions), confirming committees resolve correctly despite case, punctuation, Oxford comma, and &/and differences - verified directly against a local Postgres DB via psql.

[jessemortenson]

Interesting solution! I need to verify some local testing results with real data, but looks promising

[jessemortenson]

LGTM - we'll need to go over openstates-core release procedure, but if you want to take a shot on Monday you can. It's a few steps. Basically:

1. Follow steps in openstates-core/RELEASE.md
2. Those steps will create a new release of the openstates package with a new version number. It is "officially" released when that new version is available on pypi. Some steps of the CI will fail (that have to do with trying to push a separate "metadata" package - that's been broken for a while :( ). But if the new version is on PyPi and you can upgrade a local copy of openstates-scrapers to that new version of openstates, then it's good enough.
3. Modify openstates-scrapers to use that new version of openstates and make and merge a PR. Once that second CI process completes, then the next round of scraper runs will actually pull an image that uses the new openstates-core version's code.