Document that silhouette_samples ignores kwargs for metric="precomputed"

[ApparentlyTejas]

Summary

silhouette_samples(..., metric="precomputed") returns before metric_params, n_jobs, verbose, or **kwds are ever consulted, so they're silently ignored — no distance function is called on a precomputed matrix, matching sklearn's own silhouette_samples. That behavior wasn't documented anywhere on the function itself.

This PR documents it rather than raising, for parity with sklearn and with silhouette_score's existing precomputed handling (see discussion on #713).

• Docstring notes on metric_params, n_jobs, verbose, and **kwds stating they're ignored when metric="precomputed".
• Inline comment at the early-return explaining why.
• Regression test asserting metric_params/n_jobs/verbose passed alongside metric="precomputed" don't change the result.

Closes #714.

Test plan

• Manually traced the precomputed branch against the docstring changes and the existing test comments at tests/test_clustering.py:335-337, which already assumed this behavior.
• Added test_silhouette_samples assertion comparing a baseline precomputed call against one with metric_params/n_jobs/verbose set, asserting equal output.
• Couldn't run the suite locally due to an unrelated broken Python/libexpat install in my environment — would appreciate CI running it.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 95.21%. Comparing base (a4eb51a) to head (2d81d3c).

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #719   +/-   ##
=======================================
  Coverage   95.21%   95.21%           
=======================================
  Files          83       83           
  Lines        7902     7905    +3     
=======================================
+ Hits         7524     7527    +3     
  Misses        378      378           

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.