feat: add public sample_count_het() and cohort_heterozygosity() methods (#775)

[Tanisha127]

Summary

This PR addresses the remaining work from issue #775 by improving the
heterozygosity API with two new public methods and unit tests.

Changes

New Methods

sample_count_het()

• Adds a public counterpart to the existing private _sample_count_het() method
• Returns a pandas DataFrame with one row per window, containing:
  sample_id, window_start, window_stop, and heterozygosity
• Makes heterozygosity data directly accessible to users without
  relying on private methods

cohort_heterozygosity()

• Enables cohort-level comparisons of heterozygosity across a genome region
• Accepts the standard cohorts parameter (e.g. "taxon", "admin1_year")
  consistent with other cohort-level methods in the API
• Returns a DataFrame with cohort, n_samples, and mean_heterozygosity

Tests

• Added test_sample_count_het() and test_cohort_heterozygosity() to
  tests/anoph/test_heterozygosity.py
• All 24 tests pass across ag3, af1, adir1, and amin1 simulated datasets

Related Issue

Closes #775