openapi-python-client VS Poetry

On the backend, developers either need to manually document the entire API or rely on auto-generation tools that don’t fully meet their needs. Consumers face the same choice, write code by hand or workaround the bugs in their SDK generator (stated, lovingly, as the maintainer of an OpenAPI client generator). On top of this, these solutions result in inconsistent understandings of the API. Reproducing errors becomes time-consuming and frustrating, which feels like a battle instead of a collaboration. What we need is a shared language to describe how the API works—one that doesn’t add unnecessary layers of abstraction or manual work.

Has anyone tried Kiota, specifically the Python support? How does it compare to https://github.com/openapi-generators/openapi-python-client ?

I think we use these - https://github.com/openapi-generators/openapi-python-client

Thanks for the link, but not necessarily.

How WSDL and the code generation around it worked, was that you'd have a specification of the web API (much like OpenAPI attempts to do), which you could feed into any number of code generators, to get output code which has no coupling to the actual generator at runtime, whereas Pyotr is geared more towards validation and goes into the opposite direction: https://pyotr.readthedocs.io/en/latest/client/

The best analogy that i can think of is how you can also do schema first application development - you do your SQL migrations (ideally in an automated way as well) and then just run a command locally to generate all of the data access classes and/or models for your database tables within your application. That way, you save your time for 80% of the boring and repetitive stuff while minimizing the risks of human error and inconsistencies, while nothing preventing you from altering the generated code if you have specific needs (outside of needing to make it non overrideable, for example, a child class of a generated class). Of course, there's no reason why this can't be applied to server code either - write the spec first and generate stubs for endpoints that you'll just fill out.

Similarly there shouldn't be a need for a special client to generate stubs for OpenAPI, the closest that Python in particular has for now is this https://github.com/openapi-generators/openapi-python-client

However, for some reason, model driven development never really took off, outside of niche frameworks, like JHipster: https://www.jhipster.tech/

Furthermore, for whatever reason formal specs for REST APIs also never really got popular and aren't regarded as the standard, which to me seems silly: every bit of client code that you write will need a specific version to work against, which should be formalized.

Tallying up the results, we get 7/8 "MUST" requirements met. I think that Paperclip + actix-web seems like the most promising candidate. I'm really not opposed to writing the OpenAPI v3 construction myself as I've worked with the structure a fair bit in my openapi-python-client project (shameless plug).

Automatic documentation via OpenAPI, which lets you do things like generate Python code that knows how to talk to your API.

You can manage dependencies in Python with the package manager pip, which comes pre-installed with Python. Pip allows you to install and uninstall Python packages, and it uses a requirements.txt file to keep track of which packages your project depends on. However, pip does not have robust dependency resolution features or isolate dependencies for different projects; this is where tools like pipenv and poetry come in. These tools create a virtual environment for each project, separating the project's dependencies from the system-wide Python environment and other projects.

Poetry provides packaging and dependency management for Python. If you haven't already, install poetry via pip:

Poetry

The Semantify repository provides an example Astro.js project. Ensure you have poetry installed, then build the project from the root of the repository:

Has anyone else been paying attention to how hilariously hard it is to package PyTorch in poetry?

https://github.com/python-poetry/poetry/issues/6409

Based on this comment 5 days ago[0], it's working? I'm not sure didn't dig in too far but based on that comment it seems fair to say that it's not fully Poetry's fault because torch removed hashes (which poetry needs to be effective) for a while only recently adding it back in.

Not sure where I would stand if I fully investigated it tho.

[0] https://github.com/python-poetry/poetry/issues/6409#issuecom...

We will be running this project in Python 3.10 on Mac/Linux, and we will use Poetry to manage our dependencies. Later, we will bundle our app into a container using docker for deployment.

Here are the two main packaging issues I run into, specifically when using Poetry:

1) Lack of support for building extension modules (as mentioned by the article). There is a workaround using an undocumented feature [0], which I've tried, but ultimately decided it was not the right approach. I still use Poetry, but build the extension as a separate step in CI, rather than kludging it into Poetry.

2) Lack of support for offline installs [1], e.g. being able to download the dependencies, copy them to another machine, and perform the install from the downloaded dependencies (similar to using "pip --no-index --find-links=."). Again, you can work around this (by using "poetry export --with-credentials" and "pip download" for fetching the dependencies, then firing up pypiserver [2] to run a local PyPI server on the offline machine), but ideally this would all be a first class feature of Poetry, similar to how it is in pip.

I don't have the capacity to create Pull Requests for addressing these issues with Poetry, and I'm very grateful for the maintainers and those who do contribute. Instead, on the linked issues I share my notes on the matter, in the hope that it may at least help others and potentially get us closer to a solution.

Regardless, I'm sticking with Poetry for now. Though to be fair, the only other Python packaging tools I've used extensively are Pipenv and pip/setuptools. It's time consuming to thoroughly try out these other packaging tools, and is generally lower priority than developing features/fixing bugs, so it's helpful to read about the author's experience with these other tools, such as PDM and Hatch.

[0] https://github.com/python-poetry/poetry/issues/2740

[1] https://github.com/python-poetry/poetry/issues/2184

[2] https://pypi.org/project/pypiserver/

We believe that poetry is currently the best tool for this purpose, besides of being the most popular one at the moment. This is why we will use poetry to manage the dependencies of our project throughout this series of posts. Poetry allows you to declare the libraries your project depends on, and it will manage (install/update) them for you. Poetry also allows you to package your project into a distributable format and publish it to a repository, such as PyPI. We strongly recommend you to learn more about this tool by reading the official documentation.

I started using poetry. The problem is poetry will not install if there is dependency conflict and there is no way to ignore: github