PR for sequenzautils/gcwiggle (#345)

* added files

* edited files

* edited files

* edited file

* test(sequenzautils): Remove md5sum

Co-authored-by: kaurravneet4123 <kaurravneet4123@yahoo.com@users.noreply.github.com>
Co-authored-by: Edmund Miller <edmund.a.miller@protonmail.com>

software/sequenzautils/gcwiggle/functions.nf

@@ -0,0 +1,60 @@
+/*
+ * -----------------------------------------------------
+ *  Utility functions used in nf-core DSL2 module files
+ * -----------------------------------------------------
+ */
+
+/*
+ * Extract name of software tool from process name using $task.process
+ */
+def getSoftwareName(task_process) {
+    return task_process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()
+}
+
+/*
+ * Function to initialise default values and to generate a Groovy Map of available options for nf-core modules
+ */
+def initOptions(Map args) {
+    def Map options = [:]
+    options.args          = args.args ?: ''
+    options.args2         = args.args2 ?: ''
+    options.args3         = args.args3 ?: ''
+    options.publish_by_id = args.publish_by_id ?: false
+    options.publish_dir   = args.publish_dir ?: ''
+    options.publish_files = args.publish_files
+    options.suffix        = args.suffix ?: ''
+    return options
+}
+
+/*
+ * Tidy up and join elements of a list to return a path string
+ */
+def getPathFromList(path_list) {
+    def paths = path_list.findAll { item -> !item?.trim().isEmpty() }  // Remove empty entries
+    paths = paths.collect { it.trim().replaceAll("^[/]+|[/]+\$", "") } // Trim whitespace and trailing slashes
+    return paths.join('/')
+}
+
+/*
+ * Function to save/publish module results
+ */
+def saveFiles(Map args) {
+    if (!args.filename.endsWith('.version.txt')) {
+        def ioptions = initOptions(args.options)
+        def path_list = [ ioptions.publish_dir ?: args.publish_dir ]
+        if (ioptions.publish_by_id) {
+            path_list.add(args.publish_id)
+        }
+        if (ioptions.publish_files instanceof Map) {
+            for (ext in ioptions.publish_files) {
+                if (args.filename.endsWith(ext.key)) {
+                    def ext_list = path_list.collect()
+                    ext_list.add(ext.value)
+                    return "${getPathFromList(ext_list)}/$args.filename"
+                }
+            }
+        } else if (ioptions.publish_files == null) {
+            return "${getPathFromList(path_list)}/$args.filename"
+        }
+    }
+}

software/sequenzautils/gcwiggle/main.nf

@@ -0,0 +1,40 @@
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName } from './functions'
+
+params.options = [:]
+options        = initOptions(params.options)
+
+process SEQUENZAUTILS_GCWIGGLE {
+    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), publish_id:meta.id) }
+
+    conda (params.enable_conda ? "bioconda::sequenza-utils=3.0.0" : null)
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container "https://depot.galaxyproject.org/singularity/sequenza-utils:3.0.0--py38h6ed170a_2"
+    } else {
+        container "quay.io/biocontainers/sequenza-utils:3.0.0--py38h6ed170a_2"
+    }
+
+    input:
+    tuple val(meta), path(fasta)
+
+    output:
+    tuple val(meta), path("*.wig.gz"), emit: wig
+    path "*.version.txt"          , emit: version
+
+    script:
+    def software = getSoftwareName(task.process)
+    def prefix   = options.suffix ? "${meta.id}${options.suffix}" : "${meta.id}"
+    """
+    sequenza-utils \\
+        gc_wiggle \\
+        $options.args \\
+        --fasta $fasta \\
+        -o ${prefix}.wig.gz
+
+    echo \$(sequenzautils --version 2>&1) | sed 's/^.*sequenzautils //; s/Using.*\$//' > ${software}.version.txt
+    """
+}

software/sequenzautils/gcwiggle/meta.yml

@@ -0,0 +1,38 @@
+name: sequenzautils_gcwiggle
+description: Sequenza-utils gc_wiggle computes the GC percentage across the sequences, and returns a file in the UCSC wiggle format, given a fasta file and a window size.
+keywords:
+  - gc_wiggle
+tools:
+  - sequenzautils:
+      description: Sequenza-utils provides 3 main command line programs to transform common NGS file format - such as FASTA, BAM - to input files for the Sequenza R package. The program -gc_wiggle- takes fasta file as an input, computes GC percentage across the sequences and returns a file in the UCSC wiggle format.
+      homepage: https://sequenza-utils.readthedocs.io/en/latest/index.html
+      documentation: https://sequenza-utils.readthedocs.io/en/latest/index.html
+      doi: 10.1093/annonc/mdu479
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - fasta:
+      type: file
+      description: FASTA file
+      pattern: "*.{fasta}"
+
+output:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - version:
+      type: file
+      description: File containing software version
+      pattern: "*.{version.txt}"
+  - wig:
+      type: file
+      description: GC Wiggle track file
+      pattern: "*.{wig.gz}"
+
+authors:
+  - "@kaurravneet4123"

tests/config/pytest_software.yml

@@ -375,10 +375,18 @@ seqkit_split2:
   - software/seqkit/split2/**
   - tests/software/seqkit/split2/**
 
+sequenza_bam2seqz:
+  - software/sequenza/bam2seqz/**
+  - tests/software/sequenza/bam2seqz/**
+
 sequenza_wiggle:
   - software/sequenza/wiggle/**
   - tests/software/sequenza/wiggle/**
 
+sequenzautils_gcwiggle:
+  - software/sequenzautils/gcwiggle/**
+  - tests/software/sequenzautils/gcwiggle/**
+
 seqwish_induce:
   - software/seqwish/induce/**
   - tests/software/seqwish/induce/**

tests/software/sequenzautils/gcwiggle/main.nf

@@ -0,0 +1,14 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { SEQUENZAUTILS_GCWIGGLE } from '../../../../software/sequenzautils/gcwiggle/main.nf' addParams( options: [ 'args': '-w 50' ] )
+
+workflow test_sequenzautils_gcwiggle {
+    
+    def input = []
+    input = [ [ id:'test' ], // meta map
+              file("${launchDir}/tests/data/genomics/sarscov2/fasta/test_genome.fasta", checkIfExists: true) ]
+
+    SEQUENZAUTILS_GCWIGGLE ( input )
+}

tests/software/sequenzautils/gcwiggle/test.yml

@@ -0,0 +1,9 @@
+## TODO nf-core: Please run the following command to build this file:
+#                nf-core modules create-test-yml sequenzautils/gcwiggle
+- name: sequenzautils gcwiggle
+  command: nextflow run ./tests/software/sequenzautils/gcwiggle -entry test_sequenzautils_gcwiggle -c tests/config/nextflow.config
+  tags:
+    - sequenzautils
+    - sequenzautils_gcwiggle
+  files:
+    - path: output/sequenzautils/test.wig.gz