io-ts VS pydantic

Qué rico. Si tenés chance meté un proceso de code review fuerte, y para el tema de I/O probá a usar https://github.com/Effect-TS/schema ó https://github.com/gcanti/io-ts que les da una solución obvia al tema de "tipos para lo que devuelva el backend", aunque es en realidad mucho más capaz que eso.

My fave is still io-ts (https://github.com/gcanti/io-ts/blob/master/docs/index.md) as I find it more flexible than zod at the ingress. The author is also working on the Effect ecosystem which also looks interesting.

I’ve been using io-ts for this and been very happy with it. [1] It’s similar to Swift’s Coding protocol in case you’re familiar.

[1] https://gcanti.github.io/io-ts/

Yeah, there are a few new concepts and it's not the easiest to pick up right away. The best introduction is here on the main documentation page.

We use https://github.com/gcanti/io-ts/blob/master/Decoder.md which has a very similar interface. It can even be used to mutate the data using https://github.com/gcanti/io-ts/blob/master/Decoder.md#the-parse-combinator.

A good way to significantly improve the reliability of your app is via improving type-safety by moving away from using any to unknown. One relevant example could be when you type your backend responses and when stringifying JSON to using unknown combined with some sort of runtime type checking. It can be done either by using built-in functionality like type guards or using an external library like io-ts, zod or yup.

Usage of TypeBox is similar with io-ts and zod, but it is much powerful and faster than them. Also, TypeBox can generate JSON schema very easily. Therefore, if you're looking for a validator library for new project and not suffering from legacy codes, I think TypeBox would be much better choice than io-ts and zod. TypeBox can totally replace them.

This check can be done with different libraries like: io-ts, typebox, or zod. These libraries allow you to create objects that represent your typescript definitions. Then, these objects can be used at runtime to validate the received data, in addition, you can also convert this object to a Typescript definition to have all the benefits of using typescript. These objects can be called schema validations because they are responsible for the data validation.

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.