predict_proba functionality for kNN. Small bugfix. More concise and e…

[skywardfire1]

This PR adds:

1. Bugfix: Robust class selection. Fixed initialization of let mut max_c = 0f64. If something went wrong, the result will always be a "class 0". Now correctly initializes with the first computed probability.
2. Refactoring & DRY. predict_for_row is now a thin wrapper around predict_proba_for_row, the latter is somewhat of "main routine" for 2 or 3 functions, so we avoid repeating the code.
3. predict_proba functionality for kNN. Returns class probability distributions for input samples. Works with both search algorithms.
4. Extensive test suite. Two tests, contains ~15-20 distinct assertion checks, covering:
   • Validity of probability distributions (sum to 1.0, range [0, 1])
   • Consistency between predict() and predict_proba()
   • Equivalence of results across search algorithms
   • Edge cases: zero-weight sums, extreme k values, multiclass labels
   • Behavior differences between Uniform and Distance weighting
   • Batch prediction correctness
5. Backward compatibility No breaking changes to the public API. All existing tests pass.

I also made a real world test.

This is the predictions of a prev functionality

 [0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 5, 0, 1, 1, 5, 0, 1, 1, 2, 0, 2, 2, 0, 0, 4, 3, 3, 3, 3, 3, 0, 1, 0, 4, 4, 4, 4, 4, 4, 4, 5, 5, 1, 5, 5, 5, 5, 5, 0, 0, 7]
 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 1, 0, 0, 1, 5, 0, 1, 3, 1, 2, 2, 2, 2, 2, 0, 3, 3, 3, 3, 3, 0, 0, 0, 4, 2, 4, 4, 4, 4, 4, 1, 5, 0, 5, 5, 5, 5, 5, 5, 0, 7]
 [0, 0, 0, 0, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 3, 1, 1, 4, 1, 1, 0, 1, 2, 4, 2, 4, 2, 0, 0, 3, 3, 3, 3, 0, 4, 0, 4, 0, 4, 0, 4, 4, 4, 4, 5, 5, 5, 5, 5, 5, 5, 0, 5, 0]
 [0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 1, 0, 5, 1, 1, 1, 1, 1, 2, 4, 2, 2, 4, 1, 2, 3, 3, 3, 3, 3, 4, 0, 4, 4, 0, 4, 3, 4, 4, 4, 5, 5, 1, 5, 5, 0, 5, 5, 5, 0]
 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 0, 0, 1, 1, 1, 1, 1, 1, 2, 0, 0, 2, 0, 0, 4, 3, 3, 3, 0, 0, 0, 4, 0, 0, 3, 4, 4, 0, 4, 0, 0, 5, 5, 5, 5, 4, 5, 5, 2]

Those are new ones

 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 5, 0, 1, 1, 5, 0, 1, 1, 2, 0, 0, 2, 0, 0, 4, 3, 3, 3, 3, 3, 0, 1, 0, 4, 0, 4, 4, 4, 4, 4, 5, 5, 1, 0, 1, 5, 5, 5, 0, 2, 7]
 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 1, 0, 0, 1, 0, 0, 1, 3, 1, 2, 2, 2, 2, 2, 0, 3, 3, 3, 3, 3, 0, 0, 0, 4, 2, 4, 4, 0, 0, 4, 1, 0, 0, 5, 0, 5, 5, 5, 5, 0, 0]
 [0, 0, 0, 0, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 3, 1, 0, 0, 1, 1, 0, 1, 2, 4, 2, 1, 0, 0, 0, 3, 3, 3, 3, 0, 4, 0, 4, 0, 4, 0, 4, 4, 4, 0, 5, 5, 0, 5, 5, 5, 0, 0, 1, 7]
 [0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 0, 0, 0, 1, 1, 1, 0, 1, 1, 0, 0, 2, 0, 4, 0, 0, 3, 3, 3, 3, 3, 4, 0, 4, 4, 0, 0, 3, 4, 4, 4, 5, 5, 1, 5, 5, 0, 5, 5, 5, 0]
 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 2, 0, 0, 2, 4, 2, 4, 0, 3, 3, 0, 0, 0, 4, 0, 0, 0, 4, 4, 0, 4, 0, 0, 1, 5, 5, 5, 1, 5, 5, 1]

Pls note, that they are not identical. What we should also remember is that every result shown is actually a result of 18 kNN models voting, where every one kNN model is created using smartcore.

The difference, in my opinion, comes from float computing issues. And I did my research to ensure our new kNN realization is slightly more concise and exact. I have made some computations more mathematically correct.

[codecov]

Codecov Report

❌ Patch coverage is 62.06897% with 11 lines in your changes missing coverage. Please review.
✅ Project coverage is 44.35%. Comparing base (70d8a0f) to head (9b94198).
⚠️ Report is 11 commits behind head on development.

Files with missing lines	Patch %	Lines
src/neighbors/knn_classifier.rs	62.06%	11 Missing ⚠️

Additional details and impacted files

@@               Coverage Diff               @@
##           development     #362      +/-   ##
===============================================
- Coverage        45.59%   44.35%   -1.24%     
===============================================
  Files               93       95       +2     
  Lines             8034     8017      -17     
===============================================
- Hits              3663     3556     -107     
- Misses            4371     4461      +90     

☔ View full report in Codecov by Sentry.
📢 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.