Support manually specifying case labels for union validators

[dmontagu]

Produces error locs for union cases following the description in this comment: pydantic/pydantic#6897 (comment)

Also adds the ability to (optionally) specify the label to use for each union case.

I won't be offended if this gets rejected, or gets asked for some kind of rework.

Selected Reviewer: @davidhewitt

[codecov]

Codecov Report

Merging #841 (40a5f7e) into main (7d67b91) will increase coverage by 0.06%.
Report is 2 commits behind head on main.
The diff coverage is 100.00%.

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #841      +/-   ##
==========================================
+ Coverage   93.85%   93.91%   +0.06%     
==========================================
  Files         104      104              
  Lines       15387    15414      +27     
  Branches       25       25              
==========================================
+ Hits        14441    14476      +35     
+ Misses        940      932       -8     
  Partials        6        6              

Files Changed	Coverage Δ
python/pydantic_core/core_schema.py	96.77% <100.00%> (ø)
src/serializers/type_serializers/union.rs	82.44% <100.00%> (+0.69%)	⬆️
src/validators/union.rs	85.80% <100.00%> (+4.10%)	⬆️

... and 6 files with indirect coverage changes

---

Continue to review full report in Codecov by Sentry.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 7d67b91...40a5f7e. Read the comment docs.

[codspeed-hq]

CodSpeed Performance Report

Merging #841 will degrade performances by 24.34%

_{Comparing improved-locs (40a5f7e) with main (7d67b91)}

Summary

❌ 1 regressions
✅ 137 untouched benchmarks

⚠️ Please fix the performance issues or acknowledge them on CodSpeed.

Benchmarks breakdown

	Benchmark	main	improved-locs	Change
❌	test_decimal_from_string_core	53.4 µs	70.5 µs	-24.34%

[adriangb]

My only complaint is the performance hit. I wonder if a Vec<(CombinedValidator, Option<String>)> would be better?

[dmontagu]

@adriangb what do you think now? That updated-to-40% threshold might be helping... 😈 But looking at the raw report it doesn't seem that obviously bad. Hard to tell I guess 🤷‍♂️.

[davidhewitt]

I will run a full local benchmark tomorrow and try to upload something useful.

[davidhewitt]

Had a quick chat with @samuelcolvin this morning and the impression I got was that he has opinions, I'm going to reassign this.

[dmontagu]

please review

[davidhewitt]

Looks good to me 👍