py: expose icp algorithm to pyo3

[edgarriba]

This pull request introduces several changes to add ICP (Iterative Closest Point) functionality to the kornia-py package, including new modules, classes, and tests. The most important changes include adding new dependencies, implementing the ICP algorithm in Rust and exposing it to Python, and creating corresponding unit tests.

New Features:

• kornia-py/Cargo.toml: Added dependencies for kornia-icp and kornia-3d crates.
• kornia-py/src/icp.rs: Implemented the PyICPConvergenceCriteria and PyICPResult classes and the icp_vanilla function to expose the ICP algorithm to Python.
• kornia-py/src/pointcloud.rs: Added FromPyPointCloud trait and PyPointCloud type to facilitate point cloud data conversion between Python and Rust.

Code Integration:

• kornia-py/src/lib.rs: Integrated the new ICP functionality by importing the necessary modules and adding the new classes and functions to the Python module. [1] [2] [3]

Testing:

• kornia-py/tests/test_icp.py: Added unit tests for the new ICP functionality, including tests for ICPConvergenceCriteria, ICPResult, and the icp_vanilla function.

Miscellaneous:

• kornia-py/.gitignore: Added uv.lock to the .gitignore file.
• crates/kornia-icp/src/icp_vanilla.rs: Added Clone trait to ICPConvergenceCriteria struct.