xlsx, float threshold

phac-nml · Dec 12, 2024 · 4c18be3 · 4c18be3
1 parent 0699fca
commit 4c18be3
Show file tree

Hide file tree

Showing 2 changed files with 14 additions and 9 deletions.
diff --git a/bin/process_output.py b/bin/process_output.py
@@ -36,7 +36,7 @@ def main(argv=None):
         "--threshold",
         action="store",
         dest="threshold",
-        type=int,
+        type=float,
         help="distance threshold to be included in output",
         default=None,
         required=True,
@@ -47,22 +47,26 @@ def main(argv=None):
         action="store",
         dest="output",
         type=str,
-        help="output in query-reference format",
+        help="output prefix (without extension)",
         default=None,
         required=True,
     )
 
     args = parser.parse_args(argv)
 
     input = Path(args.input)
-    output = Path(args.output)
+    tsv_path = Path(args.output + ".tsv")
+    excel_path = Path(args.output + ".xlsx")
     threshold = args.threshold
 
     data = pd.read_csv(input, sep="\t")
     data = data[data['Distance'] <= threshold]
-    data.to_csv(output, sep="\t", index=False)
+    data.to_csv(tsv_path, sep="\t", index=False)
+    data.to_excel(excel_path)
 
-    print(f"Output written to [{output}]")
+    print("Output written to:")
+    print(tsv_path)
+    print(excel_path)
 
     return 0
 

diff --git a/modules/local/process_output/main.nf b/modules/local/process_output/main.nf
@@ -2,15 +2,16 @@ process PROCESS_OUTPUT {
     label 'process_single'
 
     container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
-        'https://depot.galaxyproject.org/singularity/pandas:2.2.1' :
-        'biocontainers/pandas:2.2.1' }"
+        'https://depot.galaxyproject.org/singularity/staramr:0.10.0--pyhdfd78af_0':
+        'biocontainers/staramr:0.10.0--pyhdfd78af_0' }"
 
     input:
     path distances
     val threshold
 
     output:
-    path "results.tsv", emit: results
+    path "results.tsv", emit: tsv
+    path "results.xlsx", emit: excel
     path "versions.yml", emit: versions
 
     when:
@@ -22,7 +23,7 @@ process PROCESS_OUTPUT {
     process_output.py \\
         $args \\
         --input $distances \\
-        --output results.tsv \\
+        --output results \\
         --threshold $threshold
 
     cat <<-END_VERSIONS > versions.yml