Refactor extract_read_features_from_bam to support pysam and samtools backends

[jkmckenna]

Motivation

• Bring extract_read_features_from_bam in line with other bam_functions utilities by supporting both a Python (pysam) and CLI (samtools) backend.
• Make the function more robust to missing qualities and provide consistent numeric outputs for downstream QC steps.

Description

• Changed function signature to extract_read_features_from_bam(bam_file_path: str | Path, samtools_backend: str | None = "auto") -> Dict[str, List[float]] and added type hints and a docstring.
• Implemented a Python backend using _require_pysam() that reads via pysam.AlignmentFile and computes median base quality, mapped length via get_blocks(), and reference lengths from bam_file.references/lengths.
• Implemented a CLI backend that parses samtools view -H for @SQ lengths and streams samtools view -F 4 output, computing mapped length from CIGAR strings and converting ASCII qualities to Phred scores.
• Added defensive handling for missing qualities/sequences and normalized returned metric types to floats.
• Added unit tests for both the python and cli backend paths in tests/unit/informatics/test_tool_backends.py to cover the new behavior.

Testing

• Added unit tests test_extract_read_features_from_bam_python_backend and test_extract_read_features_from_bam_cli_backend to cover both backends; these tests were committed with the change.
• Ran pytest tests/unit/informatics/test_tool_backends.py -q which failed during test collection with importlib.metadata.PackageNotFoundError: No package metadata was found for smftools, so automated test execution in this environment did not complete successfully.
• The new tests are self-contained and mock pysam / subprocess calls; they can be executed locally or in CI once the package is installable in the test environment (e.g. pip install -e . before running pytest).

---

Codex Task