add emmtyper module (#1028)

Co-authored-by: Gregor Sturm <mail@gregor-sturm.de>

modules/emmtyper/functions.nf

@@ -0,0 +1,78 @@
+//
+//  Utility functions used in nf-core DSL2 module files
+//
+
+//
+// Extract name of software tool from process name using $task.process
+//
+def getSoftwareName(task_process) {
+    return task_process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()
+}
+
+//
+// Extract name of module from process name using $task.process
+//
+def getProcessName(task_process) {
+    return task_process.tokenize(':')[-1]
+}
+
+//
+// Function to initialise default values and to generate a Groovy Map of available options for nf-core modules
+//
+def initOptions(Map args) {
+    def Map options = [:]
+    options.args            = args.args ?: ''
+    options.args2           = args.args2 ?: ''
+    options.args3           = args.args3 ?: ''
+    options.publish_by_meta = args.publish_by_meta ?: []
+    options.publish_dir     = args.publish_dir ?: ''
+    options.publish_files   = args.publish_files
+    options.suffix          = args.suffix ?: ''
+    return options
+}
+
+//
+// Tidy up and join elements of a list to return a path string
+//
+def getPathFromList(path_list) {
+    def paths = path_list.findAll { item -> !item?.trim().isEmpty() }      // Remove empty entries
+    paths     = paths.collect { it.trim().replaceAll("^[/]+|[/]+\$", "") } // Trim whitespace and trailing slashes
+    return paths.join('/')
+}
+
+//
+// Function to save/publish module results
+//
+def saveFiles(Map args) {
+    def ioptions  = initOptions(args.options)
+    def path_list = [ ioptions.publish_dir ?: args.publish_dir ]
+
+    // Do not publish versions.yml unless running from pytest workflow
+    if (args.filename.equals('versions.yml') && !System.getenv("NF_CORE_MODULES_TEST")) {
+        return null
+    }
+    if (ioptions.publish_by_meta) {
+        def key_list = ioptions.publish_by_meta instanceof List ? ioptions.publish_by_meta : args.publish_by_meta
+        for (key in key_list) {
+            if (args.meta && key instanceof String) {
+                def path = key
+                if (args.meta.containsKey(key)) {
+                    path = args.meta[key] instanceof Boolean ? "${key}_${args.meta[key]}".toString() : args.meta[key]
+                }
+                path = path instanceof String ? path : ''
+                path_list.add(path)
+            }
+        }
+    }
+    if (ioptions.publish_files instanceof Map) {
+        for (ext in ioptions.publish_files) {
+            if (args.filename.endsWith(ext.key)) {
+                def ext_list = path_list.collect()
+                ext_list.add(ext.value)
+                return "${getPathFromList(ext_list)}/$args.filename"
+            }
+        }
+    } else if (ioptions.publish_files == null) {
+        return "${getPathFromList(path_list)}/$args.filename"
+    }
+}

modules/emmtyper/main.nf

@@ -0,0 +1,41 @@
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName; getProcessName } from './functions'
+
+params.options = [:]
+options        = initOptions(params.options)
+
+process EMMTYPER {
+    tag "$meta.id"
+    label 'process_low'
+    publishDir "${params.outdir}",
+        mode: params.publish_dir_mode,
+        saveAs: { filename -> saveFiles(filename:filename, options:params.options, publish_dir:getSoftwareName(task.process), meta:meta, publish_by_meta:['id']) }
+
+    conda (params.enable_conda ? "bioconda::emmtyper=0.2.0" : null)
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container "https://depot.galaxyproject.org/singularity/emmtyper:0.2.0--py_0"
+    } else {
+        container "quay.io/biocontainers/emmtyper:0.2.0--py_0"
+    }
+
+    input:
+    tuple val(meta), path(fasta)
+
+    output:
+    tuple val(meta), path("*.tsv"), emit: tsv
+    path "versions.yml"           , emit: versions
+
+    script:
+    def prefix = options.suffix ? "${meta.id}${options.suffix}" : "${meta.id}"
+    """
+    emmtyper \\
+        $options.args \\
+        $fasta \\
+        > ${prefix}.tsv
+
+    cat <<-END_VERSIONS > versions.yml
+    ${getProcessName(task.process)}:
+        ${getSoftwareName(task.process)}: \$( echo \$(emmtyper --version 2>&1) | sed 's/^.*emmtyper v//' )
+    END_VERSIONS
+    """
+}

modules/emmtyper/meta.yml

@@ -0,0 +1,43 @@
+name: emmtyper
+description: EMM typing of Streptococcus pyogenes assemblies
+keywords:
+  - fasta
+  - Streptococcus pyogenes
+  - typing
+tools:
+  - emmtyper:
+      description: Streptococcus pyogenes in silico EMM typer
+      homepage: https://github.com/MDU-PHL/emmtyper
+      documentation: https://github.com/MDU-PHL/emmtyper
+      tool_dev_url: https://github.com/MDU-PHL/emmtyper
+      doi: ""
+      licence: ['GNU General Public v3 (GPL v3)']
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - fasta:
+      type: file
+      description: FASTA assembly file
+      pattern: "*.{fasta,fasta.gz,fa,fa.gz,fna,fna.gz}"
+
+output:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - versions:
+      type: file
+      description: File containing software versions
+      pattern: "versions.yml"
+  - tsv:
+      type: file
+      description: Tab-delimited result file
+      pattern: "*.tsv"
+
+authors:
+  - "@rpetit3"

tests/config/pytest_modules.yml

@@ -362,6 +362,10 @@ dshbio/splitgff3:
   - modules/dshbio/splitgff3/**
   - tests/modules/dshbio/splitgff3/**
 
+emmtyper:
+  - modules/emmtyper/**
+  - tests/modules/emmtyper/**
+
 ensemblvep:
   - modules/ensemblvep/**
   - tests/modules/ensemblvep/**

tests/modules/emmtyper/main.nf

@@ -0,0 +1,13 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { EMMTYPER } from '../../../modules/emmtyper/main.nf' addParams( options: [:] )
+
+workflow test_emmtyper {
+    
+    input = [ [ id:'test', single_end:false ], // meta map
+              file(params.test_data['sarscov2']['genome']['genome_fasta'], checkIfExists: true) ]
+
+    EMMTYPER ( input )
+}

tests/modules/emmtyper/test.yml

@@ -0,0 +1,7 @@
+- name: emmtyper test_emmtyper
+  command: nextflow run tests/modules/emmtyper -entry test_emmtyper -c tests/config/nextflow.config
+  tags:
+    - emmtyper
+  files:
+    - path: output/emmtyper/test.tsv
+      md5sum: c727ba859adec9ca8ff0e091ecf79c62