Release Version 0.0.1a15

What's New:

• Added xbarray.backends.get_backend_from_tensor utility function to infer the appropriate ComputeBackend from a given tensor, improving interoperability with user code.

Release Version 0.0.1a14

What's new:

• Added xbarray.transformations.pointcloud.pointcloud_jitter function to apply jitter noise to pointclouds for data augmentation.
• Rename PytorchComputeBackend to PyTorchComputeBackend for consistent capitalization with UniEnv.

Release Version 0.0.1a13

Changes:

• Added default_index_dtype property to the ComputeBackend interface to specify the default dtype for indexing operations.
• Make extrinsic_matrix parameters optional in the pointcloud transformation functions, defaulting to identity matrices when not provided.
• Minor documentation improvements and bugfixes.

Release Version 0.0.1a12

What's new:

• For pointcloud-based transformations, fixed a few bugs that caused the depth unprojection to output flipped x and y coordinates in the produced pointcloud
• Fix special case max recursion error when a string or byte data is inside a backend PyTree

Release Version 0.0.1a11

• Added point cloud sampling methods

Release Version 0.0.1a10

First release under the UniEnvOrg organization!