please review

CodSpeed Performance Report

Merging #9595 will not alter performance

_{Comparing K-dash:add-experimental-pipline-test (b687734) with main (6055019)}

Summary

✅ 13 untouched benchmarks

@sydney-runkle
Thank you for merging the PR!

What did you think about this API? We're looking for feedback on this new experimental feature 🚀 !

I genuinely think this is a great feature.

In the past, when defining complex constraints and transformations for model fields in Pydantic, we had to use separate Field Validators and Model Validators. However, with the introduction of this pipeline API, I believe it has enhanced type safety and improved readability with more concise syntax.

If there are any other benefits or effective ways to use this pipeline API, I would love to learn about them. :)

I understand this is currently an experimental feature, but I look forward to incorporating it once it is officially released. Thank you for implementing such a useful feature!

@K-dash,

Wonderful, thanks so much for your feedback! We're super excited to get your feedback. @adriangb did all of the initial work on this awesome feature, and even wrote a PEP to help support this! https://peps.python.org/pep-0746/

I think the main benefits are:

1. Type safety
2. Readability
3. Ease of reuse
4. Customizing constraint orders (like for string constraints, which isn't currently easily doable without this feature)

I'm sure there are even more things, but these were the main motivating factors!

@sydney-runkle
Thank you for your response.
I have looked at PEP 746 created by Adrian.
I see, in order to validate the consistency between the type annotations specified in Annotated and the associated metadata, the metadata needs to implement the __supports_type__ protocol.
I hadn't understood that part.

3. Ease of reuse
4. Customizing constraint orders (like for string constraints, which isn't currently easily doable without this feature)

Indeed, I also think the above two benefits are significant. Thank you.
Through the implementation of these test codes, I have learned a lot not only about the pipeline API but also about type annotations. I am very grateful!