Error when trying to load sbf file with python code [SOLVED]

[SebastianGut]

Hello,

i'm trying to do the training process in python code in a jupyter lab environment. Reason being that i want to train on a pretty big pointcloud.
In the moment i am stuck with the step of loading the .sbf file with the below error message.
The sbf file was created with cloudcompare and features have been computed beforehand.
Does anyone have an idea what the problem ist? Or have a workaround?
Could the sbf-file be faulty?
If anyone could help, i would be very thankful

[paul.leroy]

The problem here is not 3DMASC, the problem is the python code, read the error message.

Either you try to understand the error, you correct it, and you open a pull request on the lidar_platform repository, either you wait for some else (me ;-)) to correct this for you! You should post your request on the lidar_platform repos. Anyway, someone has already raised a similar issue, you can put a comment to motivate the developer to solve the issue:

https://github.com/p-leroy/lidar_platform/issues/18

Will try to work on that asap.

[paul.leroy]

I did a modification on the lidar_platform library, could you check it?

[SebastianGut]

Thank you very much for your quick answer and your work!
I did manage to get the training to work now :) Though i did need to add two lines to the cc_3dmasc.py to properly load all required functions, i'm not sure if this is a problem with my anaconda environment or in general. (from ..tools.sbf import sbf and from sklearn import ensemble). If wanted,I'll look into opening a pull request on Git Hub.
(Also, i think the requirements file could get updated to include the pandas=1.5 and plotly packages?)

[paul.leroy]

If you think that something could be improved, please open a pull request, then we will review it to decide whether we should merge it or not. Before opening it, be sure to have the last version of the repository merged into your branch.

But I think that the import is made in the right way. Your PYTHON_PATH may not be properly set, it should contain the directory which contains the lidar_platform library.