fix: resolve 'Couldn't locate file' error with input BAM with extra delimiters

bio/cnvkit/batch/environment.linux-64.pin.txt

@@ -138,6 +138,7 @@ https://conda.anaconda.org/conda-forge/linux-64/pandas-2.2.3-py310h5eaa309_1.con
 https://conda.anaconda.org/conda-forge/linux-64/pango-1.54.0-h4c5309f_1.conda#7df02e445367703cd87a574046e3a6f0
 https://conda.anaconda.org/conda-forge/noarch/rlpycairo-0.2.0-pyhd8ed1ab_0.conda#71d8c309e4e5c13d8f4113b5a394e633
 https://conda.anaconda.org/conda-forge/linux-64/scipy-1.14.1-py310hfcf56fc_1.conda#d9b1b75a227dbc42f3fe0e8bc852b805
+https://conda.anaconda.org/bioconda/noarch/snakemake-wrapper-utils-0.8.0-pyhdfd78af_0.conda#1650e521333852f45468d97b1b2fdcce
 https://conda.anaconda.org/conda-forge/linux-64/matplotlib-base-3.9.2-py310h68603db_2.conda#4a8c5dd75815384ddbc3aee99ccd7b50
 https://conda.anaconda.org/conda-forge/linux-64/pomegranate-0.14.8-py310hc4a4660_0.tar.bz2#5318c4d1baef8fd825b2c61b88d5f117
 https://conda.anaconda.org/bioconda/noarch/pyfaidx-0.8.1.3-pyhdfd78af_0.tar.bz2#b7235b21bc5dd12cd480a23201cc8886

bio/cnvkit/batch/environment.yaml

@@ -4,3 +4,4 @@ channels:
   - nodefaults
 dependencies:
   - cnvkit =0.9.12
+  - snakemake-wrapper-utils =0.8.0

bio/cnvkit/batch/wrapper.py

@@ -3,16 +3,14 @@
 __email__ = "[email protected]"
 __license__ = "MIT"
 
-import logging
 from os import listdir
-from os.path import basename
-from os.path import dirname
 from os.path import join
 from tempfile import TemporaryDirectory
-import shutil
 from snakemake.shell import shell
+from snakemake_wrapper_utils.snakemake import move_files
 
-log = snakemake.log_fmt_shell(stdout=False, stderr=True)
+
+log = snakemake.log_fmt_shell(stdout=True, stderr=True, append=True)
 
 input_bam_files = f"{snakemake.input.bam}"
 
@@ -50,9 +48,7 @@
 with TemporaryDirectory() as tmpdirname:
     if create_reference:
         output = f"--output-reference {join(tmpdirname, 'reference.cnn')}"
-        output_list = [snakemake.output.reference]
     else:
-        output_list = [value for key, value in snakemake.output.items()]
         output = f"-d {tmpdirname} "
     shell(
         "(cnvkit.py batch {input_bam_files} "
@@ -65,15 +61,33 @@
         "{extra}) {log}"
     )
 
-    for output_file in output_list:
-        filename = basename(output_file)
-        parent = dirname(output_file)
-
-        try:
-            shutil.copy2(join(tmpdirname, filename), output_file)
-        except FileNotFoundError as e:
-            temp_files = listdir(tmpdirname)
-            logging.error(
-                f"Couldn't locate file {basename} possible files are {[basename(f) for f in temp_files]}"
-            )
-            raise e
+    # actual generated files
+    temp_files = listdir(tmpdirname)
+
+    # mapping of file suffixes to snakemake.output attributes
+    file_map = {
+        "antitargetcoverage.cnn": "antitarget_coverage",
+        "bintest.cns": "bins",
+        ".cnr": "regions",
+        "call.cns": "segments_called",
+        "targetcoverage.cnn": "target_coverage",
+        ".cns": "segments",
+        "reference.cnn": "reference",
+    }
+
+    mapping = {}
+
+    # find matches btw generated files and snakemake output
+    for file in temp_files:
+        for suffix, attr in file_map.items():
+            if file.endswith(suffix):
+                # Skip ambiguous matches
+                if attr == "segments" and any(
+                    x in file for x in ["call.cns", "bintest.cns"]
+                ):
+                    continue
+                mapping[attr] = join(tmpdirname, file)
+                break  # stop after first match
+
+    for file in move_files(snakemake, mapping):
+        shell("{file} {log}")