Merge pull request #627 from daichengxin/prbranch

move thermorawfileparser from local to bigbio nf-core

Author: ypriverol

modules.json

@@ -36,6 +36,17 @@
                     }
                 }
             }
+        },
+        "https://github.com/bigbio/nf-modules.git": {
+            "modules": {
+                "bigbio": {
+                    "thermorawfileparser": {
+                        "branch": "main",
+                        "git_sha": "f85bcb529b16e03e8f46374d0fde5bfaf604b676",
+                        "installed_by": ["modules"]
+                    }
+                }
+            }
         }
     }
 }

modules/bigbio/thermorawfileparser/environment.yml

@@ -0,0 +1,7 @@
+name: thermorawfileparser
+channels:
+  - conda-forge
+  - bioconda
+  - defaults
+dependencies:
+  - bioconda::thermorawfileparser=1.4.5

modules/bigbio/thermorawfileparser/main.nf

@@ -0,0 +1,67 @@
+process THERMORAWFILEPARSER {
+    tag "$meta.mzml_id"
+    label 'process_low'
+    label 'process_single'
+    label 'error_retry'
+
+    conda "${moduleDir}/environment.yml"
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/thermorawfileparser:1.4.5--h05cac1d_1' :
+        'biocontainers/thermorawfileparser:1.4.5--h05cac1d_1' }"
+
+    stageInMode {
+        if (task.attempt == 1) {
+            if (task.executor == "awsbatch") {
+                'symlink'
+            } else {
+                'link'
+            }
+        } else if (task.attempt == 2) {
+            if (task.executor == "awsbatch") {
+                'copy'
+            } else {
+                'symlink'
+            }
+        } else {
+            'copy'
+        }
+    }
+    input:
+    tuple val(meta), path(rawfile)
+
+    output:
+    tuple val(meta), path("*.{mzML,mgf,parquet}"), emit: convert_files
+    path "versions.yml",   emit: versions
+    path "*.log",   emit: log
+
+    script:
+    def args = task.ext.args ?: ''
+    // Default to indexed mzML format (-f=2) if not specified in args
+    def formatArg = args.contains('-f=') ? '' : '-f=2'
+
+    """
+    ThermoRawFileParser.sh -i='${rawfile}' ${formatArg} ${args} -o=./ 2>&1 | tee '${rawfile.baseName}_conversion.log'
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        ThermoRawFileParser: \$(ThermoRawFileParser.sh --version)
+    END_VERSIONS
+    """
+
+    stub:
+    def prefix = task.ext.prefix ?: "${meta.mzml_id}"
+    def args = task.ext.args ?: ''
+    // Determine output format from args, default to mzML
+    // Format 0 = MGF, formats 1-2 = mzML, format 3 = Parquet, format 4 = None
+    def outputExt = (args =~ /-f=0\b/).find() ? 'mgf' : 'mzML'
+
+    """
+    touch '${prefix}.${outputExt}'
+    touch '${prefix}_conversion.log'
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        ThermoRawFileParser: \$(ThermoRawFileParser.sh --version)
+    END_VERSIONS
+    """
+}

modules/bigbio/thermorawfileparser/meta.yml

@@ -0,0 +1,49 @@
+name: thermorawfileparser
+description: Convert RAW file to mzML or MGF files
+keywords:
+  - raw
+  - mzML
+  - MGF
+  - OpenMS
+tools:
+  - thermorawfileparser:
+      description: |
+        ThermoRawFileParser converts Thermo RAW files to open standard formats like mzML, producing indexed output files.
+        Use `task.ext.args` to pass additional arguments, e.g.:
+        - `-f=0` for MGF output, `-f=1` for mzML, `-f=2` for indexed mzML (default), `-f=3` for Parquet, `-f=4` for None
+        - `-L` or `--msLevel=VALUE` to select MS levels (e.g., `-L=1,2` or `--msLevel=1-3`)
+      homepage: https://github.com/compomics/ThermoRawFileParser
+      documentation: https://github.com/compomics/ThermoRawFileParser
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+  - rawfile:
+      type: file
+      description: |
+        Thermo RAW file
+      pattern: "*.{raw,RAW}"
+output:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'sample1', mzml_id:'UPS1_50amol_R3' ]
+  - convert_files:
+      type: file
+      description: |
+        Converted files in mzML or MGF format depending on the format parameter (-f).
+        Format options: 0 for MGF, 1 for mzML, 2 for indexed mzML (default), 3 for Parquet, 4 for None.
+      pattern: "*.{mzML,mgf}"
+  - log:
+      type: file
+      description: log file
+      pattern: "*.log"
+  - versions:
+      type: file
+      description: File containing software version
+      pattern: "versions.yml"
+authors:
+  - "@daichengxin"
+  - "@ypriverol"

modules/bigbio/thermorawfileparser/tests/main.nf.test

@@ -0,0 +1,53 @@
+nextflow_process {
+
+    name "Test Process THERMORAWFILEPARSER"
+    script "../main.nf"
+    process "THERMORAWFILEPARSER"
+    tag "modules"
+    tag "modules_bigbio"
+    tag "thermorawfileparser"
+
+    test("Should convert RAW to mzML") {
+
+        when {
+            process {
+                """
+                input[0] = [
+                    [ id: 'test', mzml_id: 'UPS1_50amol_R3' ],
+                    file(params.test_data['proteomics']['msspectra']['ups1_50amol_r3'], checkIfExists: false)
+                ]
+                """
+            }
+        }
+
+        then {
+            assert process.success
+            assert snapshot(process.out.versions).match("versions")
+            assert new File(process.out.convert_files[0][1]).name == 'TMT_Erwinia_1uLSike_Top10HCD_isol2_45stepped_60min_01.mzML'
+            assert process.out.log.size() == 1
+        }
+    }
+
+    test("Should run stub mode") {
+
+        options "-stub"
+
+        when {
+            process {
+                """
+                input[0] = [
+                    [ id: 'test', mzml_id: 'test_sample' ],
+                    file(params.test_data['proteomics']['msspectra']['ups1_50amol_r3'], checkIfExists: false)
+                ]
+                """
+            }
+        }
+
+        then {
+            assert process.success
+            assert snapshot(process.out.versions).match("versions_stub")
+            assert new File(process.out.convert_files[0][1]).name == 'test_sample.mzML'
+            assert process.out.log.size() == 1
+        }
+    }
+}

modules/bigbio/thermorawfileparser/tests/main.nf.test.snap

@@ -0,0 +1,26 @@
+{
+  "versions": {
+    "content": [
+      [
+        "versions.yml:md5,dc9625538c025d615109ef8cac3a86ab"
+      ]
+    ],
+    "meta": {
+      "nf-test": "0.9.3",
+      "nextflow": "25.04.8"
+    },
+    "timestamp": "2025-12-11T06:27:00.000000"
+  },
+  "versions_stub": {
+    "content": [
+      [
+        "versions.yml:md5,dc9625538c025d615109ef8cac3a86ab"
+      ]
+    ],
+    "meta": {
+      "nf-test": "0.9.3",
+      "nextflow": "25.04.8"
+    },
+    "timestamp": "2025-12-11T06:27:00.000000"
+  }
+}

modules/bigbio/thermorawfileparser/tests/nextflow.config

@@ -0,0 +1,3 @@
+process {
+    publishDir = { "${params.outdir}/${task.process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()}" }
+}