feat: MultiQC temp folder, specify outdir/filename, infer options, and bump version

bio/multiqc/environment.linux-64.pin.txt

@@ -82,6 +82,7 @@ https://conda.anaconda.org/conda-forge/linux-64/pyyaml-5.4.1-py311hd4cff14_4.tar
 https://conda.anaconda.org/conda-forge/noarch/setuptools-69.0.3-pyhd8ed1ab_0.conda#40695fdfd15a92121ed2922900d0308b
 https://conda.anaconda.org/conda-forge/linux-64/simplejson-3.19.2-py311h459d7ec_0.conda#d6478cbce002db6303f0d749860f3e22
 https://conda.anaconda.org/conda-forge/noarch/six-1.16.0-pyh6c4a22f_0.tar.bz2#e5f25f8dbc060e9a8d912e432202afc2
+https://conda.anaconda.org/bioconda/noarch/snakemake-wrapper-utils-0.6.2-pyhdfd78af_0.tar.bz2#fd8759bbd04116eace828c4fab906096
 https://conda.anaconda.org/conda-forge/noarch/tenacity-8.2.3-pyhd8ed1ab_0.conda#1482e77f87c6a702a7e05ef22c9b197b
 https://conda.anaconda.org/conda-forge/noarch/typing_extensions-4.9.0-pyha770c72_0.conda#a92a6440c3fe7052d63244f3aba2a4a7
 https://conda.anaconda.org/conda-forge/noarch/wheel-0.42.0-pyhd8ed1ab_0.conda#1cdea58981c5cbc17b51973bcaddcea7

bio/multiqc/environment.yaml

@@ -4,3 +4,4 @@ channels:
   - nodefaults
 dependencies:
   - multiqc =1.20
+  - snakemake-wrapper-utils =0.6.2

bio/multiqc/meta.yaml

@@ -1,13 +1,17 @@
 name: multiqc
 description: |
-  Generate qc report using multiqc.
+  Generate qc report using MultiQC.
 url: https://multiqc.info/
 authors:
   - Julian de Ruiter
+  - Filipe G. Vieira
 input:
   - input directory containing qc files, default behaviour is to extract folder path from the provided files or parent folder if a folder is provided.
 output:
   - qc report (html)
   - multiqc data folder or zip (optional)
 params:
+  - extra: additional program arguments.
   - use_input_files_only: if this variable is set to True input will be used as it is, i.e no folder will be extract from provided file names
+notes: |
+  * options `--data-dir`, `--no-data-dir`, `--zip-data-dir`, and `--no-report` are automaticall inferred.

bio/multiqc/test/Snakefile

@@ -3,9 +3,9 @@ rule multiqc_dir:
         expand("samtools_stats/{sample}.txt", sample=["a", "b"]),
     output:
         "qc/multiqc.html",
-        directory("qc/multiqc_data"),
+        directory("qc_data/multiqc_data"),
     params:
-        extra="--data-dir",  # Optional: extra parameters for multiqc.
+        extra="--verbose",  # Optional: extra parameters for multiqc.
     log:
         "logs/multiqc.log",
     wrapper:
@@ -17,9 +17,9 @@ rule multiqc_file:
         expand("samtools_stats/{sample}.txt", sample=["a"]),
     output:
         "qc/multiqc.a.html",
-        "qc/multiqc.a_data.zip",
+        "qc_data/multiqc.a_data.zip",
     params:
-        extra="--zip-data-dir",  # Optional: extra parameters for multiqc.
+        extra="--verbose",  # Optional: extra parameters for multiqc.
         use_input_files_only=True,  # Optional, use only a.txt and don't search folder samtools_stats for files
     log:
         "logs/multiqc.log",

bio/multiqc/wrapper.py

@@ -6,9 +6,11 @@
 __license__ = "MIT"
 
 
-import tempfile
+# No need for explicit temp folder, since MultiQC already uses TMPDIR (https://multiqc.info/docs/usage/troubleshooting/#no-space-left-on-device)
+
 from pathlib import Path
 from snakemake.shell import shell
+from snakemake_wrapper_utils.snakemake import is_arg
 
 
 extra = snakemake.params.get("extra", "")
@@ -24,21 +26,48 @@
     input_data = set(snakemake.input)
 
 
-with tempfile.TemporaryDirectory() as tmpdir:
-    shell(
-        "multiqc"
-        " {extra}"
-        " --outdir {tmpdir}"
-        " --filename out"
-        " {input_data}"
-        " {log}"
-    )
-
-    for output in snakemake.output:
-        if output.endswith("_data"):
-            ext = "_data"
-        elif output.endswith(".zip"):
-            ext = "_data.zip"
-        else:
-            ext = Path(output).suffix
-        shell("mv {tmpdir}/out{ext} {output}")
+# Add extra options depending on output files
+no_report = True
+for output in snakemake.output:
+    if output.endswith(".html"):
+        no_report = False
+    if output.endswith("_data"):
+        extra += " --data-dir"
+    if output.endswith(".zip"):
+        extra += " --zip-data-dir"
+if no_report:
+    extra += " --no-report"
+if (
+    not is_arg("--data-dir", extra)
+    and not is_arg("-z", extra)
+    and not is_arg("--zip-data-dir", extra)
+):
+    extra += " --no-data-dir"
+
+# Specify output dir and file name, since they are stored in the JSON file
+out_dir = Path(snakemake.output[0]).parent
+file_name = Path(snakemake.output[0]).with_suffix("").name
+
+
+shell(
+    "multiqc"
+    " {extra}"
+    " --outdir {out_dir}"
+    " --filename {file_name}"
+    " {input_data}"
+    " {log}"
+)
+
+
+# Move files to another destination (if needed)
+for output in snakemake.output:
+    if output.endswith("_data"):
+        ext = "_data"
+    elif output.endswith(".zip"):
+        ext = "_data.zip"
+    else:
+        ext = Path(output).suffix
+
+    default_dest = f"{out_dir}/{file_name}{ext}"
+    if default_dest != output:
+        shell("mv {default_dest} {output}")