Move enum validation and serialization to Rust

[samuelcolvin]

Change Summary

Sequel to pydantic/pydantic-core#1235, should give a 4x+ speedup to enum validation.

This will fail until we make a new pydantic-core release.

Checklist

• The pull request title is a good summary of the changes - it will be used in the changelog
• Unit tests for the changes exist
• Tests pass on CI
• Documentation reflects the changes where applicable
• My PR is ready to review, please add a comment including the phrase "please review" to assign reviewers

[cloudflare-pages]

Deploying pydantic-docs with    Cloudflare Pages

Latest commit:	9d42245
Status:	✅  Deploy successful!
Preview URL:	https://e2d623f1.pydantic-docs2.pages.dev
Branch Preview URL:	https://rust-enum-validation.pydantic-docs2.pages.dev

View logs

[samuelcolvin]

we probably need a test to check that tuple values generate the right schema.

[codspeed-hq]

CodSpeed Performance Report

Merging #9064 will improve performances by 30.01%

_{Comparing rust-enum-validation (9d42245) with main (d6ca333)}

Summary

⚡ 3 improvements
✅ 7 untouched benchmarks

Benchmarks breakdown

	Benchmark	main	rust-enum-validation	Change
⚡	test_north_star_dump_json	202.4 ms	180.9 ms	+11.9%
⚡	test_north_star_validate_json	335 ms	258.3 ms	+29.7%
⚡	test_north_star_validate_json_strict	334.6 ms	257.4 ms	+30.01%

[sydney-runkle]

Left a few thoughts! Looks good overall! Excited for the significant speedup :)

[sydney-runkle]

The only failing fastapi test is due to a msg changing re an enum member error, and our version policy states that we can indeed expect msg changes in minor versions, so I think this is ok (https://docs.pydantic.dev/latest/version-policy/).

Perhaps we could just open a heads up with a test fix on the fastapi end? This is due to the fact that we removed the logic in the to_enum function which raised an error with a specific format for the allowed members...

[davidhewitt]

Looking solid! We've done a clear win for the JSON north star. The Python north start is less signficant than I hoped; I'll have to dig around with that another time and see if there's any possible wins...

[sydney-runkle]

The only failing fastapi test is due to a msg changing re an enum member error, and our version policy states that we can indeed expect msg changes in minor versions, so I think this is ok (https://docs.pydantic.dev/latest/version-policy/).

Update, the issue is with the url being included now, which we can fix with a simple one liner PR where we add match_pydantic_error_url to the failing test. For now, we're skipping said failing test.

[sydney-runkle]

Looks great!