melange VS pydantic

Hey HN, at Ahrefs we have been working on an online book that hopefully helps React developers get up and running with Melange, an OCaml to JavaScript compiler. You can read more about Melange here: https://melange.re/.

There are still a few chapters that we'd like to add before considering it "complete", but it might be already helpful for some folks out there, that's why we decided to publish it early.

The book uses Reason syntax to implement React components using ReasonReact components. You can read more about both in:

https://reasonml.github.io/

In my previous post on trying to use the NextJS App Router and Reason I described some of the problems and limitations of their compatibility with one another. With the release of Melange 2 I decided to see if the new features of Melange 2 could help to increase the compatibility of Reason and the NextJS App Router. I have also documented some of the things learnt after trying Melange (v1) with Astro and Remix.

So it's Reason, not ReasonML which the umbrella project's name, and Rescript is a imcompatible syntax split from the Bucklescript team (that previously transpiled Reason to JS). Bucklescript's new name is... Rescript.

But not everyone agrees with the split and work is being done on Melange to replace Bucklescript : https://github.com/melange-re/melange

Ultimately JsOfOcaml can directly transpile Ocaml to JS.

In reality, most folks that developed BuckleScript frontends with ReasonML switched to ReScript syntax and are happy with it. Some felt more friction because of their reliance on PPXes or FP-heavy libraries (like Relude) and those people tend to use the Melange fork of BuckleScript or they switched to js_of_ocaml.

There is a fork of ReScript that supports ReasonML syntax and with the goal of maintaining Ocaml compatibility: https://github.com/melange-re/melange.

There's also a fork of BuckleScript/ReScript called Melange that guts its build system so that instead of using ninja, it works with more standard tools for the ecosystem, specifically dune and esy. In doing so they managed to also finally get the compiler off of OCaml 4.06: now it can use a newer OCaml compiler and take advantage of four years worth of language and compiler improvements.

For OCaml and ReasonML your options are js_of_ocaml (mentioned here in ReasonML docs) or a fairly new fork of BuckleScript called melange. They differ in implementation and output, with JSOO taking intermediate bytecode generated by ocamlc and turning it into unreadable JS, vs Melange being a patched compiler that builds more human-readable JS.

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.