flama VS pydantic-core

There has been a considerable effort in the last few years to try and standardise the way in which these type of APIs are implemented via different frameworks. However, over the last few years, a new type of functionality has become more and more popular: machine learning (ML) models; and the existing frameworks for building APIs are not well suited for this type of functionality. In this series of posts, we will learn how to build APIs using a Framework for the development of Lightweight Applications and Machine-learning Automation, also known as Flama.

*some options that were mentioned in the thread I linked to include - Starlite (made to bypass perceived shortcomings of FastAPI) - Flama (Starlette w/ Marshmellow instead of Pydantic) - APIFlask (Flask w/ Pydantic) - Django Ninja - Quart-Schema

Flama (Starlette, Marshmallow instead of Pydantic)

Funny that you ask... https://github.com/pydantic/pydantic-core Unfortunately it seems that the functionality you ask for is not (yet) part of this ...

I encourage you to checkout the official benchmarks for more realistic and detailed examples, and, as always, YMMV.

> to also be constrained by a separate set of data types which are legal in rust.

This isn't really how writing rust/python iterop works. You tend to have opaque handles you call python methods on. Here's a decent example I found skimming the code.

https://github.com/pydantic/pydantic-core/blob/main/src/inpu...

Thanks for pointing out to that, I did not know about it. Also attaching repo in case someone would be interested as well - https://github.com/pydantic/pydantic-core

pydantic-core: The core validation logic for pydantic, a Python data parsing and validation library.

pydandic-core [0] will hopefully solve this issue (written in Rust)

[0] -- https://github.com/pydantic/pydantic-core

LinkedIn influencers are weird lol. Rust v Python is apples and oranges. Rust would be glued together by python just like it does with C/C++ and Java/Spark today. We’re already seeing some packages go this direction, like pydantic v2 is rewriting its core validation in rust.