openapi-code-generator VS poem

Whilst it's not as expressive/flexible as typespec, and in my experience it's not always well supported by tooling, you can do $ref's across files in openapi specifications.

Eg: https://github.com/mnahkies/openapi-code-generator/blob/main...

I've become interested in SBOM recently, and found there were great tools like https://dependencytrack.org/ for CycloneDX SBOMs, but all I have is SPDX SBOMs generated by GitHub.

I decided to have a go at writing my own dependency track esque tool aiming to integrate with the APIs GitHub provides.

It's pretty limited in functionality so far, but can give a high level summary of the types of licenses your repository dependencies use, and let you drill down into potentially problematic ones.

Written in NextJS + mui + sqlite, and using another project of mine to generate most of the API boilerplate/glue (https://github.com/mnahkies/openapi-code-generator)

Congratulations on launching, you have some interesting ideas in there.

Using a LLM to generate missing operation ids isn't something I've tried, instead I simply combine http method plus path segments which at least guarantees uniqueness [1]. I do a similar thing for extracting and naming inline schemas based on the operation and media types [2].

How do you prevent naming collisions? And do you find the resulting names to be significantly better than a deterministic approached like I described?

I'll definitely checkout the curated specifications - always useful to get more high quality (and hopefully varied) specifications to test my code generator with, and the lint rules is a great idea - I've had to explain what patterns lend themselves well to code generation many times.

I'm on mobile so may have missed it, but looking at one of your typescript examples I couldn't see any runtime response body validation, is this something you're thinking about?

- [1] https://github.com/mnahkies/openapi-code-generator/blob/main...

Yeah I'm also on the schema first side of the debate.

I think for me it comes down to a few key points:

- APIs are forever, the choice of language/framework is an implementation detail

- Constraining yourself to what can be represented in the specification is better than generating a specification from implementation that may not be capable of expressing the full details

- When working with diverse languages it provides a common ground/language for discussing API changes. Eg: if you have java backend, kotlin android, swift iOS, react/whatever web you can bring everyone together with the spec

- Subjective, but a good spec will include a bunch of documentation and examples that tend to create a lot of noise in the code. I personally prefer to keep this in the spec and the implementation smaller

I think the main counterpoint to this is that you can generate the spec and then take that and change your mind if you later change language/framework etc - it's not a one-way door.

My biggest bug bear is that regardless of spec first or implementation first, you should have something you write once and generate the rest of the glue from (eg: docs, client sdks). Writing each piece manually/independently always leads to drift and bugs.

(I'm working on my own little openapi -> typescript code generator over here https://github.com/mnahkies/openapi-code-generator - eventually plan to support more than typescript, and adding typespec support is something I'm currently considering)

TypeSpec is great, but if you're working with Rust and you're about to write a new project that will require an OpenApi spec sooner or later, I'd like to recommend a web framework that has spec generation baked in:

https://github.com/poem-web/poem (see poem_openapi)

All you need to do is derive a trait on your response structs and in return you get an almost perfectly generated spec. Unions, objects, enums are first class citizens.

Also, if you're from coming from PHP, the controllers feel very much like symfony controllers.

P.s. Please do recommend an ORM that would feel closer to doctrine. I miss doctrine.

I've used FastAPI/Pydantic before as well and it was a very good experience. It has been a while since I used them though, so I forget exactly how they compare, but I do recall it wrote most of the OpenAPI spec. for you. I will also agree when I saw utoipa and aide they seemed to require more boilerplate even though I have not done an actual comparison. Poem, on the other hand, seems very neat and clean and doesn't have any duplication that I can see. Here is the [TODO example](https://github.com/poem-web/poem/blob/master/examples/openapi/todos/src/main.rs).

The example at https://github.com/poem-web/poem/tree/master/examples/openapi/auth-apikey works without issues for me.

'Poem', it turns to nicer to play with for me than 'Axum'. Example of web socket chat https://github.com/poem-web/poem/blob/master/examples/poem/websocket-chat/src/main.rs

If you need built-in OpenAPI support, I can recommend Poem.

If OpenAPI is important, then you may to consider https://github.com/poem-web/poem instead of axum. It isn't nearly as widely used, but it should still be compatible with everything else.

Poem: More rigid and less popular and aims to be easier https://github.com/poem-web/poem

Did you check out Poem? It seems promising, with openapi generation built in already.

https://github.com/poem-web/poem is a pretty nice framework

I recently been working with warp and stumbled upon poem framework (https://github.com/poem-web/poem), any thoughts about that? Considering Axum as well but staying a bit on cautionary side until 0.6 full release.