fix: stabilize spearman distance correlation sampling

[GeorgePearse]

Summary

Makes sample_size subsample points (rows) before computing Spearman correlation over pairwise distances.

Why

Sampling arbitrary pairs can be noisy and can over/under-emphasize regions depending on the distribution of pairs; subsampling points tends to give a more stable estimate for large datasets.

Testing

• Not run (pure Python change).

[openhands-ai]

Looks like there are a few issues preventing this PR from being merged!

• GitHub Actions are failing:
  • Performance Benchmarks

If you'd like me to help, just leave a comment, like

@OpenHands please fix the failing actions on PR #7 at branch `fix/spearman-subsample`

Feel free to include any additional details that might help me get this PR into a better state.

_{^{You can manage your notification settings}}