The question is in!

To offer some more information while we wait: The pipelines will for sure be available. And I am pushing for the next 2 months. Regarding RT sources, we have a system to upload it already: dm.upload_rt_source()
However, most scenarios were introduced before this system. So, the data is not available for most of them. We'd have to upload it, which is not hard. Then the plan would be to add an "available RT source" flag in the scenarios database so researchers can see which scenarios have the available source.

Bottomline: We'll do both. The plan is clear. When is the question. This is what I'll aim to let you know in few days.

@jmoraispk This is amazing. I was actually able to generate the Mitusba .xml and mesh/*.ply files based on params.json, objects.json, materials.mat and vertices.mat. I then assigned custom RadioMaterial based on Sionna's definition and I am successfully able to load the scene in sionna. The only downside is I am not able to get the roads properly for example, let me paste the image that I see when I do dm.plot_summary('city_0_newyork_3p5_s') and the scene.preview() from Sionna.

Oh nice!

The roads have been a common pain point. It has to do with a flawed early design choice of DeepMIMO that we are yet to correct - lossy scenes translation. As you can see, they even pollute the scene visualization in DeepMIMO. I'd suggest skipping those for now if you can.

This should be fixed for the 4.0 release, without API changes.

Thanks for the clarification. Does the road/materials of the road have an impact in the ray traced channels? If we remove it would it make sense?

They do, but not much. You can change the terrain material to match the material of the road. Or just remove the roads without changing anything else and keep it like that for consistency - most of DeepMIMO scenarios don't have roads.