Fix parallel transport for LEM and Log-Cholesky metrics

[bruAristimunha]

Summary

• Fix LEM and Log-Cholesky parallel transport to be non-trivial in the ambient SPD tangent space, replacing the incorrect identity transport. LEM now uses Fréchet derivatives of matrix log/exp (Thanwerdas & Pennec 2023), and Log-Cholesky uses Cholesky pullback/pushforward (Lin 2019, Proposition 7).
• Add frechet.py module with frechet_derivative_log and frechet_derivative_exp via the Daleckii-Krein theorem, using the existing Loewner matrix pattern.
• Fix linkcode_resolve in Sphinx docs to use inspect.getsourcefile with line numbers instead of the broken src/ prefix approach.
• Expand test coverage with self-transport, roundtrip, linearity, inner product preservation, gradient flow, non-identity assertion, and parametrized batch tests for all three transport methods.

Test plan

• All 74 tests in test_transport.py pass
• Verify Sphinx docs build with correct source links
• Run the plot_parallel_transport.py example to confirm updated output

[github-actions]

📚 Documentation Preview

📦 Download Documentation Artifact

Download the documentation-html artifact from the workflow run to view the docs locally.

💡 To enable live previews, add a SURGE_TOKEN secret to this repository. See surge.sh for setup instructions.

[bruAristimunha]

many thanks for pointing out the issues @qbarthelemy