add ismapper module (#773)

* add ismapper module

* Update main.nf

* Update main.nf

* Update meta.yml

* Apply suggestions from code review

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

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+    // Do not publish versions.yml unless running from pytest workflow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

@@ -0,0 +1,44 @@
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName; getProcessName } from './functions'
+
+params.options = [:]
+options        = initOptions(params.options)
+
+process ISMAPPER {
+    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::ismapper=2.0.2" : null)
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container "https://depot.galaxyproject.org/singularity/ismapper:2.0.2--pyhdfd78af_1"
+    } else {
+        container "quay.io/biocontainers/ismapper:2.0.2--pyhdfd78af_1"
+    }
+
+    input:
+    tuple val(meta), path(reads), path(reference), path(query)
+
+    output:
+    tuple val(meta), path("results/*"), emit: results
+    path "versions.yml"               , emit: versions
+
+    script:
+    def prefix = options.suffix ? "${meta.id}${options.suffix}" : "${meta.id}"
+    """
+    ismap \\
+        $options.args \\
+        --t $task.cpus \\
+        --output_dir results \\
+        --queries $query \\
+        --reference $reference \\
+        --reads $reads
+
+    cat <<-END_VERSIONS > versions.yml
+    ${getProcessName(task.process)}:
+        ${getSoftwareName(task.process)}: \$( echo \$( ismap --version 2>&1 ) | sed 's/^.*ismap //' )
+    END_VERSIONS
+    """
+}

modules/ismapper/meta.yml

@@ -0,0 +1,50 @@
+name: ismapper
+description: Identify insertion sites positions in bacterial genomes
+keywords:
+  - fastq
+  - insertion sequences
+tools:
+  - ismapper:
+      description: A mapping-based tool for identification of the site and orientation of IS insertions in bacterial genomes.
+      homepage: https://github.com/jhawkey/IS_mapper
+      documentation: https://github.com/jhawkey/IS_mapper
+      tool_dev_url: https://github.com/jhawkey/IS_mapper
+      doi: "https://doi.org/10.1186/s12864-015-1860-2"
+      licence: ['BSD']
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - reads:
+      type: file
+      description: A set of paired-end FASTQ files
+      pattern: "*.{fastq.gz,fq.gz}"
+  - reference:
+      type: file
+      description: Reference genome in GenBank format
+      pattern: "*.{gbk}"
+  - query:
+      type: file
+      description: Insertion sequences to query in FASTA format
+      pattern: "*.{fasta,fa}"
+
+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"
+  - results:
+      type: directory
+      description: Directory containing ISMapper result files
+      pattern: "*/*"
+
+authors:
+  - "@rpetit3"

tests/config/pytest_modules.yml

@@ -521,6 +521,10 @@ iqtree:
   - modules/iqtree/**
   - tests/modules/iqtree/**
 
+ismapper:
+  - modules/ismapper/**
+  - tests/modules/ismapper/**
+
 ivar/consensus:
   - modules/ivar/consensus/**
   - tests/modules/ivar/consensus/**

tests/modules/ismapper/main.nf

@@ -0,0 +1,18 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { ISMAPPER } from '../../../modules/ismapper/main.nf' addParams( options: [:] )
+
+workflow test_ismapper {
+    
+    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) ],
+        file("https://github.com/jhawkey/IS_mapper/raw/master/test/inputs/S_suis_P17.gbk", checkIfExists: true),
+        file("https://github.com/jhawkey/IS_mapper/raw/master/test/inputs/ISSsu3.fasta", checkIfExists: true)
+    ]
+
+    ISMAPPER ( input )
+}

tests/modules/ismapper/test.yml

@@ -0,0 +1,27 @@
+- name: ismapper test_ismapper
+  command: nextflow run tests/modules/ismapper -entry test_ismapper -c tests/config/nextflow.config
+  tags:
+    - ismapper
+  files:
+    - path: output/ismapper/results/test/ISSsu3/test_ISSsu3_left_final.fastq
+      md5sum: d41d8cd98f00b204e9800998ecf8427e
+    - path: output/ismapper/results/test/ISSsu3/test_ISSsu3_right_final.fastq
+      md5sum: d41d8cd98f00b204e9800998ecf8427e
+    - path: output/ismapper/results/test/ISSsu3/test__AM946016.1_closest.bed
+      md5sum: d41d8cd98f00b204e9800998ecf8427e
+    - path: output/ismapper/results/test/ISSsu3/test__AM946016.1_intersect.bed
+      md5sum: d41d8cd98f00b204e9800998ecf8427e
+    - path: output/ismapper/results/test/ISSsu3/test__AM946016.1_table.txt
+      md5sum: 9e05cda3990cb841db2bfb6e6e04a1f5
+    - path: output/ismapper/results/test/ISSsu3/test_left_AM946016.1_finalcov.bed
+      md5sum: d41d8cd98f00b204e9800998ecf8427e
+    - path: output/ismapper/results/test/ISSsu3/test_left_AM946016.1_merged.sorted.bed
+      md5sum: d41d8cd98f00b204e9800998ecf8427e
+    - path: output/ismapper/results/test/ISSsu3/test_left_AM946016.1_unpaired.bed
+      md5sum: d41d8cd98f00b204e9800998ecf8427e
+    - path: output/ismapper/results/test/ISSsu3/test_right_AM946016.1_finalcov.bed
+      md5sum: d41d8cd98f00b204e9800998ecf8427e
+    - path: output/ismapper/results/test/ISSsu3/test_right_AM946016.1_merged.sorted.bed
+      md5sum: d41d8cd98f00b204e9800998ecf8427e
+    - path: output/ismapper/results/test/ISSsu3/test_right_AM946016.1_unpaired.bed
+      md5sum: d41d8cd98f00b204e9800998ecf8427e