add porechop module (#914)

* add module main porechop

* update porechop main

* add porechop functions

* update meta porechop

* add test main porechop

* add porechop pytest yml

* add porechop test.yml

* Update modules/porechop/meta.yml

Co-authored-by: Robert A. Petit III <robbie.petit@gmail.com>

* re-add porechop avoid conflict

* Update modules/porechop/meta.yml

Co-authored-by: Robert A. Petit III <robbie.petit@gmail.com>

* fix prefix suffix

Co-authored-by: Robert A. Petit III <robbie.petit@gmail.com>

modules/porechop/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+    // Do not publish versions.yml unless running from pytest workflow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/porechop/main.nf

@@ -0,0 +1,42 @@
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName; getProcessName } from './functions'
+
+params.options = [:]
+options        = initOptions(params.options)
+
+process PORECHOP {
+    tag "$meta.id"
+    label 'process_medium'
+    publishDir "${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::porechop=0.2.4" : null)
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container "https://depot.galaxyproject.org/singularity/porechop:0.2.4--py39h7cff6ad_2"
+    } else {
+        container "quay.io/biocontainers/porechop:0.2.4--py38h8c62d01_2"
+    }
+
+    input:
+    tuple val(meta), path(reads)
+
+    output:
+    tuple val(meta), path("*.fastq.gz")  , emit: reads
+    path "versions.yml"                          , emit: versions
+
+    script:
+    def prefix = options.suffix ? "${meta.id}${options.suffix}" : "${meta.id}"
+    """
+    porechop \\
+        -i ${reads} \\
+        -t ${task.cpus} \\
+        ${options.args} \\
+        -o ${prefix}.fastq.gz
+
+    cat <<-END_VERSIONS > versions.yml
+    ${getProcessName(task.process)}:
+        ${getSoftwareName(task.process)}: \$( porechop --version )
+    END_VERSIONS
+    """
+}

modules/porechop/meta.yml

@@ -0,0 +1,50 @@
+name: porechop
+description: Adapter removal and demultiplexing of Oxford Nanopore reads
+keywords:
+  - adapter
+  - nanopore
+  - demultiplexing
+tools:
+  - porechop:
+      description: Adapter removal and demultiplexing of Oxford Nanopore reads
+      homepage: "https://github.com/rrwick/Porechop"
+      documentation: "https://github.com/rrwick/Porechop"
+      tool_dev_url: "https://github.com/rrwick/Porechop"
+      doi: "10.1099/mgen.0.000132"
+      licence: ["GPL v3"]
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - reads:
+      type: file
+      description: fastq/fastq.gz file
+      pattern: "*.{fastq,fastq.gz,fq,fq.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"
+  - reads:
+      type: file
+      description: Demultiplexed and/or adapter-trimmed fastq.gz file
+      pattern: "*.{fastq.gz}"
+
+authors:
+  - "@ggabernet"
+  - "@jasmezz"
+  - "@d4straub"
+  - "@LaurenceKuhl"
+  - "@SusiJo"
+  - "@jonasscheid"
+  - "@jonoave"
+  - "@GokceOGUZ"

tests/config/pytest_modules.yml

@@ -863,6 +863,10 @@ plink/vcf:
   - modules/plink/vcf/**
   - tests/modules/plink/vcf/**
 
+porechop:
+  - modules/porechop/**
+  - tests/modules/porechop/**
+
 preseq/lcextrap:
   - modules/preseq/lcextrap/**
   - tests/modules/preseq/lcextrap/**

tests/modules/porechop/main.nf

@@ -0,0 +1,13 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { PORECHOP } from '../../../modules/porechop/main.nf' addParams( options: [args: '', suffix: '_porechop'] )
+
+workflow test_porechop {
+
+    input = [ [ id:'test', single_end:true ], // meta map
+                file(params.test_data['sarscov2']['nanopore']['test_fastq_gz'], checkIfExists: true) ]
+
+    PORECHOP ( input )
+}

tests/modules/porechop/test.yml

@@ -0,0 +1,7 @@
+- name: porechop test_porechop
+  command: nextflow run tests/modules/porechop -entry test_porechop -c tests/config/nextflow.config
+  tags:
+    - porechop
+  files:
+    - path: output/porechop/test_porechop.fastq.gz
+      md5sum: 08f314ae9f162c8dcc27e5b513d2064d