perf: improve subtraction and em performance

* Remove some unused code.
* Use Rust for paring down eliminated reads during subtraction.
* Improve efficiency of subtraction code.
* Replace all SAM files with BAM in subtraction elimination.

Author: igboyes

example/to_isolates.bam

@@ -47,6 +47,6 @@ def p_score_cutoff():
 
 
 @fixture
-def subtracted_sam_path(work_path: Path) -> Path:
-    """The path to the SAM file after subtraction reads have been eliminated."""
-    return work_path / "subtracted.sam"
+def subtracted_bam_path(work_path: Path) -> Path:
+    """The path to the BAM file after subtraction reads have been eliminated."""
+    return work_path / "subtracted.bam"

fixtures.py

@@ -20,8 +20,33 @@ class PathoscopeResults:
     coverage: dict[str, list[int]]
 
 def run_expectation_maximization(
-    sam_path: str,
+    alignment_path: str,
     p_score_cutoff: float,
     ref_lengths: dict[str, int],
 ) -> PathoscopeResults:
-    """Run Pathoscope expectation maximization algorithm using Rust."""
+    """Run Pathoscope expectation maximization algorithm using Rust on SAM/BAM files."""
+
+def parse_isolate_scores(
+    alignment_path: str,
+    p_score_cutoff: float,
+) -> dict[str, float]:
+    """Parse isolate alignment file (SAM or BAM) and extract high scores for each read."""
+
+def find_candidate_otus(
+    alignment_path: str,
+    p_score_cutoff: float,
+) -> set[str]:
+    """Extract candidate OTU reference IDs from an alignment file (SAM/BAM)."""
+
+def find_candidate_otus_from_bytes(
+    sam_bytes: bytes,
+    p_score_cutoff: float,
+) -> set[str]:
+    """Extract candidate OTU reference IDs from SAM text data."""
+
+def calculate_coverage_from_em_results(
+    alignment_path: str,
+    p_score_cutoff: float,
+    ref_lengths: dict[str, int],
+) -> dict[str, list[int]]:
+    """Calculate coverage directly from EM results and alignment data."""

python/workflow_pathoscope/rust.pyi

@@ -218,20 +218,32 @@ def write_report(
 
 
 def run_pathoscope(
-    sam_path: Path,
+    alignment_path: Path,
     p_score_cutoff: float,
     ref_lengths: dict[str, int],
 ):
-    """Run Pathoscope on the SAM file at ``sam_path`` with the given ``p_score_cutoff``.
+    """Run Pathoscope on an alignment file.
 
     Returns PathoscopeResults containing EM results and coverage data.
 
-    :param sam_path: The path to the SAM file.
+    :param alignment_path: The path to the SAM or BAM file.
     :param p_score_cutoff: The minimum allowed ``p_score`` for an alignment.
     :param ref_lengths: Dictionary mapping reference IDs to their lengths.
     """
     return run_expectation_maximization(
-        str(sam_path),
+        str(alignment_path),
         p_score_cutoff,
         ref_lengths,
     )
+
+
+# Backward compatibility alias - DEPRECATED
+def run_pathoscope_sam(
+    sam_path: Path, p_score_cutoff: float, ref_lengths: dict[str, int]
+):
+    """
+    Deprecated: Use run_pathoscope instead.
+
+    This function is kept for backward compatibility.
+    """
+    return run_pathoscope(sam_path, p_score_cutoff, ref_lengths)