fix: Fix PMD command handle errors and output csv

src/metaDMG/PMD.py

@@ -0,0 +1,62 @@
+#
+
+import shlex
+import subprocess
+from pathlib import Path
+
+import pandas as pd
+
+
+#%%
+
+
+def iterate_command(command: str):
+
+    p = subprocess.Popen(
+        shlex.split(command),
+        stdout=subprocess.PIPE,
+        stderr=subprocess.DEVNULL,
+    )
+    for line in iter(p.stdout.readline, b""):  # type: ignore
+        if line:
+            line = line.decode("utf-8")
+            if line.endswith("\n"):
+                line = line[:-1]
+            yield line
+
+    # waits for the process to finish and returns the returncode
+    yield p.wait()
+
+
+def compute_PMDs(bam_file: Path, metaDMG_cpp: str):
+    """Run the PMD command from metaDMG-cpp and get a DataFrame with the results.
+
+    Parameters
+    ----------
+    bam_file
+        Alignment file to compute the PMD scores on
+    metaDMG_cpp
+        The metaDMG binary to use
+    """
+
+    command = f"{metaDMG_cpp} pmd {bam_file}"
+
+    returncode = 1
+    reads = []
+    PMDs = []
+    for i, line in enumerate(iterate_command(command)):
+
+        # if finished, check returncode
+        if isinstance(line, int):
+            returncode = line
+
+        else:
+            read, PMD = line.split("\tPMD:")
+            reads.append(read)
+            PMDs.append(float(PMD))
+
+    if returncode != 0:
+        raise Exception(f"Error running {command}")
+
+    df = pd.DataFrame({"read": reads, "PMD": PMDs})
+    return df

src/metaDMG/cli/cli.py

@@ -618,20 +618,34 @@ def mismatch_to_mapDamage(
 
 
 @cli_app.command("PMD", rich_help_panel="Utils")
-def compute_PMD(
-    config_file: Path = typer.Argument(
+def PMD(
+    filename: Path = typer.Argument(
         ...,
         file_okay=True,
-        help="Path to the config-file.",
+        help="Path to the config-file or a single BAM file.",
+    ),
+    csv_out: Path = typer.Option(
+        ...,
+        "--output",
+        "-o",
+        file_okay=True,
+        help="Output CSV file.",
+    ),
+    metaDMG_cpp: str = typer.Option(
+        "./metaDMG-cpp",
+        "--metaDMG-cpp",
+        "-m",
+        help="The command needed to run the metaDMG-cpp program.",
     ),
 ):
-    """Compute the PMD scores for the samples in the config file."""
+    """Compute the PMD scores for the chosen BAM file and save to csv."""
+
+    from metaDMG.PMD import compute_PMDs
 
-    from metaDMG.utils import make_configs, run_PMD
+    df_PMDs = compute_PMDs(filename, metaDMG_cpp)
 
-    configs = make_configs(config_file)
-    for config in configs:
-        run_PMD(config)
+    csv_out.parent.mkdir(parents=True, exist_ok=True)
+    df_PMDs.to_csv(csv_out, index=False)
 
 
 #%%