🏋️‍♀️ new submodule mash/sketch🕺 (#426)

* new submodule mash/sketch

* fixed submodule naming

* OK, tag is diff to keyword

* OK  another round 🤣

* removed TODO comments

* updated as per review comments 🙆‍♂️

* updated functions.nf 😁

* Update software/mash/sketch/main.nf

* Update main.nf

Removed blank line at the 12th

Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com>

software/mash/sketch/functions.nf

@@ -0,0 +1,70 @@
+/*
+ * -----------------------------------------------------
+ *  Utility functions used in nf-core DSL2 module files
+ * -----------------------------------------------------
+ */
+
+/*
+ * Extract name of software tool from process name using $task.process
+ */
+def getSoftwareName(task_process) {
+    return task_process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()
+}
+
+/*
+ * Function to initialise default values and to generate a Groovy Map of available options for nf-core modules
+ */
+def initOptions(Map args) {
+    def Map options = [:]
+    options.args            = args.args ?: ''
+    options.args2           = args.args2 ?: ''
+    options.args3           = args.args3 ?: ''
+    options.publish_by_meta = args.publish_by_meta ?: []
+    options.publish_dir     = args.publish_dir ?: ''
+    options.publish_files   = args.publish_files
+    options.suffix          = args.suffix ?: ''
+    return options
+}
+
+/*
+ * Tidy up and join elements of a list to return a path string
+ */
+def getPathFromList(path_list) {
+    def paths = path_list.findAll { item -> !item?.trim().isEmpty() }      // Remove empty entries
+    paths     = paths.collect { it.trim().replaceAll("^[/]+|[/]+\$", "") } // Trim whitespace and trailing slashes
+    return paths.join('/')
+}
+
+/*
+ * Function to save/publish module results
+ */
+def saveFiles(Map args) {
+    if (!args.filename.endsWith('.version.txt')) {
+        def ioptions  = initOptions(args.options)
+        def path_list = [ ioptions.publish_dir ?: args.publish_dir ]
+        if (ioptions.publish_by_meta) {
+            def key_list = ioptions.publish_by_meta instanceof List ? ioptions.publish_by_meta : args.publish_by_meta
+            for (key in key_list) {
+                if (args.meta && key instanceof String) {
+                    def path = key
+                    if (args.meta.containsKey(key)) {
+                        path = args.meta[key] instanceof Boolean ? "${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"
+        }
+    }
+}

software/mash/sketch/main.nf

@@ -0,0 +1,40 @@
+include { initOptions; saveFiles; getSoftwareName } from './functions'
+
+params.options = [:]
+options        = initOptions(params.options)
+
+process MASH_SKETCH {
+    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::mash=2.3" : null)
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container "https://depot.galaxyproject.org/singularity/mash:2.3--he348c14_1"
+    } else {
+        container "quay.io/biocontainers/mash:2.3--he348c14_1"
+    }
+
+    input:
+    tuple val(meta), path(reads)
+
+    output:
+    tuple val(meta), path("*.msh")        , emit: mash
+    tuple val(meta), path("*.mash_stats") , emit: stats
+    path "*.version.txt"                  , emit: version
+
+    script:
+    def software = getSoftwareName(task.process)
+    def prefix   = options.suffix ? "${meta.id}${options.suffix}" : "${meta.id}"
+    """
+    mash \\
+        sketch \\
+        $options.args \\
+        -p $task.cpus \\
+        -o ${prefix} \\
+        -r $reads \\
+        2> ${prefix}.mash_stats
+    echo \$(mash --version 2>&1) > ${software}.version.txt
+    """
+}

software/mash/sketch/meta.yml

@@ -0,0 +1,40 @@
+name: mash_sketch
+description: Creates vastly reduced representations of sequences using MinHash
+keywords:
+  - mash/sketch
+tools:
+  - mash:
+      description: Fast sequence distance estimator that uses MinHash
+      homepage: https://github.com/marbl/Mash
+      documentation: https://mash.readthedocs.io/en/latest/sketches.html
+      tool_dev_url: https://github.com/marbl/Mash
+      doi: "10.1186/s13059-016-0997-x"
+      licence: ['https://github.com/marbl/Mash/blob/master/LICENSE.txt']
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - reads:
+      type: file
+      description: List of input paired-end FastQ files
+
+output:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - version:
+      type: file
+      description: Sketch output
+      pattern: "*.{mash}"
+  - stats:
+      type: file
+      description: Sketch statistics
+      pattern: "*.{mash_stats}"
+
+authors:
+  - "@thanhleviet"

tests/config/pytest_software.yml

@@ -301,6 +301,10 @@ kraken2/run:
   - software/kraken2/run/**
   - tests/software/kraken2/run/**
 
+mash/sketch:
+  - software/mash/sketch/**
+  - tests/software/mash/sketch/**
+
 methyldackel/extract:
   - software/methyldackel/extract/**
   - tests/software/methyldackel/extract/**
@@ -386,14 +390,14 @@ quast:
   - software/quast/**
   - tests/software/quast/**
 
-rasusa:
-  - software/rasusa/**
-  - tests/software/rasusa/**
-
 rapidnj:
   - software/rapidnj/**
   - tests/software/rapidnj/**
 
+rasusa:
+  - software/rasusa/**
+  - tests/software/rasusa/**
+
 salmon/index:
   - software/salmon/index/**
   - tests/software/salmon/index/**

tests/software/mash/sketch/main.nf

@@ -0,0 +1,16 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { MASH_SKETCH } from '../../../../software/mash/sketch/main.nf' addParams( options: [:] )
+
+workflow test_mash_sketch {
+    
+    input = [ [ id:'test', single_end:false], // meta map
+              [ file(params.test_data['sarscov2']['illumina']['test_1_fastq_gz'], checkIfExists: true),
+                file(params.test_data['sarscov2']['illumina']['test_2_fastq_gz'], checkIfExists: true)
+              ]
+            ]
+
+    MASH_SKETCH ( input )
+}

tests/software/mash/sketch/test.yml

@@ -0,0 +1,9 @@
+- name: mash sketch
+  command: nextflow run ./tests/software/mash/sketch -entry test_mash_sketch -c tests/config/nextflow.config
+  tags:
+    - mash/sketch
+  files:
+    - path: output/mash/test.msh
+      md5sum: d747145a43dad5f82342036f8f5d9133
+    - path: output/mash/test.mash_stats
+      md5sum: 2a6f297d8e69a5e4160243bc6c89129c