Python Bindings with pybind11 or nanobind

[JonasHablitzelAVLD]

Hi,

thanks for the great work!

I wanted to ask if there is any plan to inlcude python bindings for the tool in general or if it should stay as simple as possible and the bindings should be handeled externally? I think other tools like scenariogeneration could benefit from the faster and nice processing of geometries in this libary.

Best regards,

Jonas

[pageldev]

Hi, thanks, appreciate it!

I do notice a bigger interest in Python bindings. Unfortunately, I have no experience with Python bindings and currently don't have plans / time to work on it. But I am of course open to adding it. I would like to keep this project lean, but I imagine that only the build needs to be extended a bit and some glue code needs to be written.

Anybody feel free to make a pull request to add this functionality.

[jpahm]

Hi, just wanted to write here to notify that I'm currently working on a Cython wrapper for this library as part of a project for an organization I'm in. It's still extremely WIP and isn't usable yet, but I'll be working on it consistently and should have something usable within a month or so ideally.

Thank you @pageldev for this awesome library!

[chrisjcc]

@pageldev @JonasHablitzelAVLD, I've created PR #108 to address this. A few highlights from the changes:

Improved parsing logic: The PR resolves inconsistencies when handling geometries with small deltas or when transitioning between road segments, which was at the root of the NaN propagation issue.

Validation added: Includes additional checks to catch invalid or malformed geometry definitions early during parsing.

Unit tests: Added a couple targeted tests helping ensure regression coverage going forward.

Let me know if you'd like any further refinements or if there's a preferred approach to integrate these fixes.