feat: Migrate from CRF++ to Ingredient Parser (a Python package)

[michael-genson]

What this PR does / why we need it:

(REQUIRED)

Replaces the CRF++ dependency with ingredient-parser (a Python library). Inspired largely from @hay-kot's recipes-api which is getting used for recipinned.

While the results are pretty comparable to the old CRF++ library, this gives several added benefits:

• reduces deployment complexity since we can just install this as a Python dependency now
• much easier to work on in the dev container (it just works now)

Which issue(s) this PR fixes:

(REQUIRED)

Fixes #4960
Fixes #4993
Fixes #5121

Special notes for your reviewer:

(fill-in or delete this section)

Not everything is 1:1 (e.g. "ground black pepper" is parsed as a food rather than "black pepper" as the food and "ground" as the note), but it seems to be about 50/50 when it comes to improvements vs regressions (and they're minor). At least now we can easily add some more tests since it works in the dev container without a hassle.

This package theoretically supports additional languages, but it doesn't work out of the box. Something we can look into in the future. For now it really only works with English (same as the old parser).

Also had to fix a totally unrelated issue regarding stale data because it's making tests fail. No clue why this is happening now and wasn't happening before

Testing

(fill-in or delete this section)

Threw lots of examples at it and got comparable outputs vs the old parser.

[Kuchenpirat]

Looks good, just the one comment.

[Kuchenpirat]

LGTM 👍
Might have to see if i can add some german to the package :)