add spatyper module (#784)

* add spatyper module

* lint fix

* Apply suggestions from code review

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

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+    // Do not publish versions.yml unless running from pytest workflow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

@@ -0,0 +1,46 @@
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName; getProcessName } from './functions'
+
+params.options = [:]
+options        = initOptions(params.options)
+
+process SPATYPER {
+    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::spatyper=0.3.3" : null)
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container "https://depot.galaxyproject.org/singularity/spatyper%3A0.3.3--pyhdfd78af_3"
+    } else {
+        container "quay.io/biocontainers/spatyper:0.3.3--pyhdfd78af_3"
+    }
+
+    input:
+    tuple val(meta), path(fasta)
+    path repeats
+    path repeat_order
+
+    output:
+    tuple val(meta), path("*.tsv"), emit: tsv
+    path "versions.yml"           , emit: versions
+
+    script:
+    def prefix = options.suffix ? "${meta.id}${options.suffix}" : "${meta.id}"
+    def input_args = repeats && repeat_order ? "-r ${repeats} -o ${repeat_order}" : ""
+    """
+    env
+    spaTyper \\
+        $options.args \\
+        $input_args \\
+        --fasta $fasta \\
+        --output ${prefix}.tsv
+
+    cat <<-END_VERSIONS > versions.yml
+    ${getProcessName(task.process)}:
+        ${getSoftwareName(task.process)}: \$( echo \$(spaTyper --version 2>&1) | sed 's/^.*spaTyper //' )
+    END_VERSIONS
+    """
+}

modules/spatyper/meta.yml

@@ -0,0 +1,50 @@
+name: spatyper
+description: Computational method for finding spa types.
+keywords:
+  - fasta
+  - spatype
+tools:
+  - spatyper:
+      description: Computational method for finding spa types.
+      homepage: https://github.com/HCGB-IGTP/spaTyper
+      documentation: https://github.com/HCGB-IGTP/spaTyper
+      tool_dev_url: https://github.com/HCGB-IGTP/spaTyper
+      doi: https://doi.org/10.5281/zenodo.4063625
+      licence: ['LGPL 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}"
+  - repeats:
+      type: file
+      description: spa repeat sequences in FASTA format (Optional)
+      pattern: "*.{fasta}"
+  - repeat_order:
+      type: file
+      description: spa types and order of repeats (Optional)
+      pattern: "*.{txt}"
+
+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 results
+      pattern: "*.{tsv}"
+
+authors:
+  - "@rpetit3"

tests/config/pytest_modules.yml

@@ -951,6 +951,10 @@ spades:
   - modules/spades/**
   - tests/modules/spades/**
 
+spatyper:
+  - modules/spatyper/**
+  - tests/modules/spatyper/**
+
 staphopiasccmec:
   - modules/staphopiasccmec/**
   - tests/modules/staphopiasccmec/**

tests/modules/spatyper/main.nf

@@ -0,0 +1,26 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { SPATYPER } from '../../../modules/spatyper/main.nf' addParams( options: [:] )
+include { SPATYPER as SPATYPER_ENRICH } from '../../../modules/spatyper/main.nf' addParams( options: [args: '--do_enrich'] )
+
+workflow test_spatyper {
+    input = [ [ id:'test' ],
+              file(params.test_data['sarscov2']['genome']['genome_fasta'], checkIfExists: true) ]
+
+    repeats = []
+    repeat_order = []
+
+    SPATYPER ( input, repeats, repeat_order )
+}
+
+workflow test_spatyper_enrich {
+    input = [ [ id:'test' ],
+              file(params.test_data['sarscov2']['genome']['genome_fasta'], checkIfExists: true) ]
+
+    repeats = []
+    repeat_order = []
+
+    SPATYPER_ENRICH ( input, repeats, repeat_order )
+}

tests/modules/spatyper/test.yml

@@ -0,0 +1,15 @@
+- name: spatyper test_spatyper
+  command: nextflow run tests/modules/spatyper -entry test_spatyper -c tests/config/nextflow.config
+  tags:
+    - spatyper
+  files:
+    - path: output/spatyper/test.tsv
+      md5sum: a698352823875171696e5e7ed7015c13
+
+- name: spatyper test_spatyper_enrich
+  command: nextflow run tests/modules/spatyper -entry test_spatyper_enrich -c tests/config/nextflow.config
+  tags:
+    - spatyper
+  files:
+    - path: output/spatyper/test.tsv
+      md5sum: a698352823875171696e5e7ed7015c13