Update global search for API changes

[niqjohnson]

The search.gov API that powers cf.gov global site search has changed in a couple of ways recently:

1. They've removed custom facets, which we were using to get page language and render better search result descriptions than the default snippet
2. Something has changed that has re-introduced the " | Consumer Financial Protection Bureau" from page <title>s into search result titles.

This PR removes our handling custom facets and strips the " | CFPB" suffix from search result titles.

See GHE/Design-Development/cfgov/issues/4707 for more details. Also see #8993, #8889, #8865, #8803, #8815, and #8809 for where most of this was added.

---

Additions

• Strip out the " | Consumer Financial Protection Bureau" from search result titles

Removals

• Everything related to custom search.gov facets, which are now no longer indexed or returned in API results

How to test this PR

1. Do some searches, like http://localhost:8000/search/?q=mortgage and http://localhost:8000/search/?q=bureau
2. Make sure result titles don't have " | Consumer Financial Protection Bureau" on the end of them as they do in production
3. Make sure best bets, result descriptions/snippets, result links, result count, and pagination work as they do in production

Notes and todos

• See the inline note I added to the bit that removes the " | Consumer Financial Protection Bureau" suffix. Is there a better way to do this and still account for all the possibilities?

Checklist

• PR has an informative and human-readable title
  • PR titles are used to generate the change log in releases; good ones make that easier to scan.
  • Consider prefixing, e.g., "Mega Menu: fix layout bug", or "Docs: Update Docker installation instructions".
• Changes are limited to a single goal (no scope creep)
• Code follows the standards laid out in the CFPB development guidelines
• Future todos are captured in comments and/or tickets

[niqjohnson]

It was trickier than I thought to strip the | Consumer Financial Protection Bureau from the end of search result titles. We used to do it like res["title"] = res["title"][:-39] (see #8993). That mostly worked, but I realized it didn't cover if someone searches for, say, "bureau", which would come back in the API results as | Consumer Financial Protection \ue000Bureau\ue001 and thus leave | in the result title. This regex seems to work fine unless we for some reason happen to have pages with a | somewhere else in their title. We shouldn't have any page titles like that, but if there's a way to improve this regex, I'm all for it. I thought about using a negative lookahead to only capture the last instance, but I wasn't sure if the extra complexity was worth it.

[wpears]

Seems reasonable to me! Definitely better than the previous way.

[wpears]

Nice work!