#476 Use native factorization for object string categories

Summary

Fixes #471

• Add a safe native factorization path for homogeneous object-backed
  string categories.

• Preserve existing fallback behavior for mixed objects, null-like values,
  bytes, and custom string semantics.

• Keep the existing low-cardinality probe as the fast-path gate.

• Preserve category ordering, codes, and count semantics.

• Add regression coverage proving only unique labels cross the Python
  label path.

Testing

• tests/test_custom_ramps_and_palette.py: 92 passed, 1 skipped
• tests/test_components.py: 103 passed
• Ruff check and format passed
• git diff --check passed

Summary by CodeRabbit

• Bug Fixes

  • Improved categorical handling for object arrays containing only exact Python strings, keeping category order, codes, counts, and distinct labels consistent with fixed-width string arrays.
  • Preserved distinct categories containing NUL characters and fallback behavior for mixed or unsupported values.
  • Limited distinct-count probing and sampling when assessing near-unique categories.

• Tests

  • Added coverage for string-only, NUL-containing, near-unique, and mixed-value categories.