typing VS pydantic

Iterable[str] is unfortunately evil as it matches str which is often unintended. (see: https://github.com/python/typing/issues/256) One would need both NOT-type and AND-type in order to properly handle these.

The basic motivation behind them is that functions can have arguments that can only take a specific set of values, and those functions return values/types change based on that input. Common examples are (you can find more here):

While yes 26 people liked the idea here: https://github.com/python/typing/issues/193

but looking at relevant issues on GitHub it seems this has been shot down repeatedly. python/typing#18, python/typing#213

See https://github.com/python/typing/issues/256 for a discussion.

Lower bounds are not supported for TypeVars, unfortunately.

Mypy allows that because initial versions of PEP-484 allowed that. This has changed; here's the current wording on the PEP:

> This is no longer the recommended behavior. Type checkers should move towards requiring the optional type to be made explicit.

https://www.python.org/dev/peps/pep-0484/#id29

Dunno, other, larger projects than the one I'm working on seem to run up against this from time to time. (rasa_core, to pick one example from near the top of a Google search; also Telethon, Blender, TensorFlow, Pandas. Guido also filed a bug on the typing module in an early version of Python 3.5 because of unexpected implications of this particular issue, so the problem isn't exactly purely theoretical.) That's aside from the wish for conceptual purity in the call signatures of classes and their subclasses, which is not always and automatically a bad wish to have; and the notion that a language that prides itself on its introspective faculties might want to make introspection of classes from the top of a class hierarchy possible, at least in theory? Perhaps to facility learning about the language and/or visualizing large class hierarchies easily, for instance?

First, note the method prefix_allowed_tokens_fn. This method applies a Pydantic model to constrain/guide how the LLM generates tokens. Next, see how that constrain can be applied to txtai's LLM pipeline.

utype is a concise alternative of pydantic with simplified parameters and usages, supporting both sync/async functions and generators parsing, and capable of using native logic operators to define logical types like AND/OR/NOT, also provides custom type parsing by register mechanism that supports libraries like pydantic, attrs and dataclasses

I like some of the changes from v1 to v2. But then you have something like this [0] removed from the library without proper documentation or replacement, resulting in ugly workarounds in the link that wont' work properly.

[0]: https://github.com/pydantic/pydantic/discussions/6337

The easiest implementation I found was to use a PyDantic class for my target schema — and use that as a parameter for the method call to “ChatCompletion.create()”. Here’s a fragment of the GinAI Python classes used.

Also, FastStream uses Pydantic to parse input JSON-encoded data into Python objects, making it easy to work with structured data in your applications, so you can serialize your input messages just using type annotations.

Pydantic Validation: Leverage Pydantic's validation capabilities to serialize and validate incoming messages

Not sure if it is exactly related, but there is an open issue on Github for that exact message.

Well the performance increase is so huge because pydantic1 is really really slow. And for using rust, I'd have expected more tbh…

I've been benchmarking pydantic v2 against typedload (which I write) and despite the rust, it still manages to be slower than pure python in some benchmarks.

The ones on the website are still about comparing to v1 because v2 was not out yet at the time of the last release.

pydantic's author will refuse to benchmark any library that is faster (https://github.com/pydantic/pydantic/pull/3264 https://github.com/pydantic/pydantic/pull/1525 https://github.com/pydantic/pydantic/pull/1810) and keep boasting about amazing performances.

On pypy, v2 beta was really really really slow.