New module: checkqc

.github/workflows/test.yml

@@ -544,6 +544,8 @@ jobs:
             path: modules/nf-core/cellranger/multi
           - profile: conda
             path: modules/nf-core/cellranger/vdj
+          - profile: conda
+            path: modules/nf-core/checkqc
           - profile: conda
             path: modules/nf-core/custom/dumpsoftwareversions
           - profile: conda

modules/nf-core/checkqc/environment.yml

@@ -0,0 +1,11 @@
+name: checkqc
+channels:
+  - conda-forge
+  - bioconda
+  - defaults
+dependencies:
+  - python=3.8
+  - pip
+  - pip: # FIXME https://github.com/nf-core/modules/issues/5814
+      - checkqc==3.8.0
+      - interop==1.2.4

modules/nf-core/checkqc/main.nf

@@ -0,0 +1,50 @@
+process CHECKQC {
+    tag "$meta.id"
+    label 'process_single'
+
+    container "community.wave.seqera.io/library/python_pip_interop_checkqc:d76c912c8fadc561"
+
+    input:
+    tuple val(meta), path(run_dir)
+    path(checkqc_config)
+
+    output:
+    tuple val(meta), path("*checkqc_report.json"), emit: report
+    path "versions.yml"                          , emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    // Exit if running this module with -profile conda / -profile mamba
+    if (workflow.profile.tokenize(',').intersect(['conda', 'mamba']).size() >= 1) {
+        error "CheckQC module does not support Conda yet. Please use Docker / Singularity / Podman instead."
+    }
+
+    def args = task.ext.args ?: ''
+    def config = checkqc_config ? "--config $checkqc_config" : ''
+
+    """
+    checkqc \
+        $args \
+        $config \
+        --json \
+        $run_dir > checkqc_report.json
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        checkqc: \$( checkqc --version | sed -e "s/checkqc, version //g" )
+    END_VERSIONS
+    """
+
+    stub:
+    def args = task.ext.args ?: ''
+    """
+    touch checkqc_report.json
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        checkqc: \$( checkqc --version | sed -e "s/checkqc, version //g" )
+    END_VERSIONS
+    """
+}

modules/nf-core/checkqc/meta.yml

@@ -0,0 +1,39 @@
+---
+# yaml-language-server: $schema=https://raw.githubusercontent.com/nf-core/modules/master/modules/yaml-schema.json
+name: "checkqc"
+description: A simple program to parse Illumina NGS data and check it for quality criteria
+keywords:
+  - QC
+  - Illumina
+  - genomics
+tools:
+  - "checkqc":
+      description: "A simple program to parse Illumina NGS data and check it for quality criteria."
+      homepage: "https://github.com/Molmed/checkQC"
+      documentation: "http://checkqc.readthedocs.io/en/latest/"
+      doi: "10.21105/joss.00556"
+      licence: ["GPL v3"]
+
+input:
+  - run_dir:
+      type: file
+      description: |
+        Illumina sequencing run directory
+        Can be directory or a compressed tar (tar.gz) of the directory
+  - checkqc_config:
+      type: file
+      description: CheckQC configuration file
+      pattern: "*.{yml,yaml}"
+
+output:
+  - versions:
+      type: file
+      description: File containing software versions
+      pattern: "versions.yml"
+  - report:
+      type: file
+      description: CheckQC report in json format
+      pattern: "*checkqc_report.json"
+
+authors:
+  - "@matrulda"

modules/nf-core/checkqc/tests/main.nf.test

@@ -0,0 +1,44 @@
+nextflow_process {
+
+    name "Test Process CHECKQC"
+    script "../main.nf"
+    process "CHECKQC"
+    tag "modules"
+    tag "modules_nfcore"
+    tag "untar"
+    tag "checkqc"
+
+    test("homo_sapiens illumina test_interop_bcl2fastqstats_flowcell - QC OK") {
+        config "./nextflow.config"
+
+        setup {
+            run("UNTAR") {
+                script "../../untar/main.nf"
+                process {
+                    """
+                    input[0] = Channel.of([[id: 'test'], file(params.modules_testdata_base_path + 'genomics/homo_sapiens/illumina/interop_bcl2fastqstats/test_flowcell_stats.tar.gz', checkIfExists: true)])
+                    """
+                }
+            }
+        }
+
+        when {
+            process {
+                """
+                    input[0] = UNTAR.out.untar
+                    input[1] = []
+                """
+            }
+        }
+
+        then {
+            assertAll(
+            { assert process.success },
+            { assert snapshot(process.out).match() }
+          )
+        }
+
+    }
+    }
+
+

modules/nf-core/checkqc/tests/main.nf.test.snap

@@ -0,0 +1,35 @@
+{
+    "homo_sapiens illumina test_interop_bcl2fastqstats_flowcell - QC OK": {
+        "content": [
+            {
+                "0": [
+                    [
+                        {
+                            "id": "test"
+                        },
+                        "checkqc_report.json:md5,c69cb912a8c5bf53544764bd5bd73ae5"
+                    ]
+                ],
+                "1": [
+                    "versions.yml:md5,c37381fc450f4ed00da2987430db8177"
+                ],
+                "report": [
+                    [
+                        {
+                            "id": "test"
+                        },
+                        "checkqc_report.json:md5,c69cb912a8c5bf53544764bd5bd73ae5"
+                    ]
+                ],
+                "versions": [
+                    "versions.yml:md5,c37381fc450f4ed00da2987430db8177"
+                ]
+            }
+        ],
+        "meta": {
+            "nf-test": "0.8.4",
+            "nextflow": "24.04.3"
+        },
+        "timestamp": "2024-07-17T18:30:52.712542733"
+    }
+}

modules/nf-core/checkqc/tests/nextflow.config

@@ -0,0 +1,5 @@
+process {
+    withName: CHECKQC {
+        ext.args = '--downgrade-errors ReadsPerSampleHandler'
+    }
+}

modules/nf-core/checkqc/tests/tags.yml

@@ -0,0 +1,2 @@
+checkqc:
+  - modules/nf-core/checkqc/**

tests/config/test_data.config

@@ -536,6 +536,8 @@ params {
 
                 varlociraptor_scenario                                  = "${params.test_data_base}/data/genomics/homo_sapiens/illumina/varlociraptor/scenario.yml"
 
+                test_interop_bcl2fastqstats_flowcell                    = "${params.test_data_base}/data/genomics/homo_sapiens/illumina/interop_bcl2fastqstats/test_flowcell_stats.tar.gz"
+
                 contig_ploidy_priors_table                              = "${params.test_data_base}/data/genomics/homo_sapiens/illumina/gatk/contig_ploidy_priors_table.tsv"
 
                 purecn_ex1_bam                                          = "${params.test_data_base}/data/genomics/homo_sapiens/illumina/purecn/purecn_ex1.bam"