wherobots / geotorchai Goto Github PK

Space partitioning functions raise an error in shapely version >= 2.0

There are some API changes in the newer versions of the image processing library scikit-image. These changes need to be handled so that including handcrafted features to the raster models do not break.

I think it will be good if you can set up GitHub Action for CI. This way, people can create PRs and your CI can automatically verify the correctness of new PRs.

There are three popular tensor representations used by various grid-based models in the literature. Support all three representations for all grid-based datasets.

Add following two functions:

1. This function aggregates features in a Sedona DataFrame into a spatiotemporal grid format
2. This function aggregates features in a Sedona DataFrame into a spatial grid format

When I am running following code
from geotorch.models.raster import DeepSatV2
I am getting ModuleNotFoundError
ModuleNotFoundError: No module named 'geotorch.models'

Dear GeoTorchAI team
I would like to know if there will be any support for preprocessing H3 hexagons.
Thanks!

There is clear documentation on how a contributor can embed new models.

is there a way GeoTorch / dependencies can be simply installed using pip install. That will make it easy for a lot of people to use it.

No module named 'geotorch.models' how to solve that error

Can you please add a link like this one from Sedona

to the README and website? Therefore, the users can try GeoTorchAI immediately.