Merge pull request #158 from JoseEspinosa/untar

Add untar module

.github/filters.yml

@@ -201,6 +201,10 @@ trimgalore:
   - software/trimgalore/**
   - tests/software/trimgalore/**
 
+untar:
+  - software/untar/**
+  - tests/software/untar/**
+
 ucsc_bedgraphtobigwig:
   - software/ucsc/bedgraphtobigwig/**
   - tests/software/ucsc/bedgraphtobigwig/**

software/untar/functions.nf

@@ -0,0 +1,59 @@
+/*
+ * -----------------------------------------------------
+ *  Utility functions used in nf-core DSL2 module files
+ * -----------------------------------------------------
+ */
+
+/*
+ * Extract name of software tool from process name using $task.process
+ */
+def getSoftwareName(task_process) {
+    return task_process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()
+}
+
+/*
+ * Function to initialise default values and to generate a Groovy Map of available options for nf-core modules
+ */
+def initOptions(Map args) {
+    def Map options = [:]
+    options.args          = args.args ?: ''
+    options.args2         = args.args2 ?: ''
+    options.publish_by_id = args.publish_by_id ?: false
+    options.publish_dir   = args.publish_dir ?: ''
+    options.publish_files = args.publish_files
+    options.suffix        = args.suffix ?: ''
+    return options
+}
+
+/*
+ * Tidy up and join elements of a list to return a path string
+ */
+def getPathFromList(path_list) {
+    def paths = path_list.findAll { item -> !item?.trim().isEmpty() }  // Remove empty entries
+    paths = paths.collect { it.trim().replaceAll("^[/]+|[/]+\$", "") } // Trim whitespace and trailing slashes
+    return paths.join('/')
+}
+
+/*
+ * Function to save/publish module results
+ */
+def saveFiles(Map args) {
+    if (!args.filename.endsWith('.version.txt')) {
+        def ioptions = initOptions(args.options)
+        def path_list = [ ioptions.publish_dir ?: args.publish_dir ]
+        if (ioptions.publish_by_id) {
+            path_list.add(args.publish_id)
+        }
+        if (ioptions.publish_files instanceof Map) {
+            for (ext in ioptions.publish_files) {
+                if (args.filename.endsWith(ext.key)) {
+                    def ext_list = path_list.collect()
+                    ext_list.add(ext.value)
+                    return "${getPathFromList(ext_list)}/$args.filename"
+                }
+            }
+        } else if (ioptions.publish_files == null) {
+            return "${getPathFromList(path_list)}/$args.filename"
+        }
+    }
+}

software/untar/main.nf

@@ -0,0 +1,34 @@
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName } from './functions'
+
+params.options = [:]
+def options    = initOptions(params.options)
+
+process UNTAR {
+    tag "$archive"
+    publishDir "${params.outdir}",
+        mode: params.publish_dir_mode,
+        saveAs: { filename -> saveFiles(filename:filename, options:params.options, publish_dir:getSoftwareName(task.process), publish_id:'') }
+
+    conda (params.enable_conda ? "conda-forge::sed=4.7" : null)
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container "https://containers.biocontainers.pro/s3/SingImgsRepo/biocontainers/v1.2.0_cv1/biocontainers_v1.2.0_cv1.img"
+    } else {
+        container "biocontainers/biocontainers:v1.2.0_cv1"
+    }
+
+    input:
+    path archive
+
+    output:
+    path "$untar"       , emit: untar
+    path "*.version.txt", emit: version
+
+    script:
+    def software = getSoftwareName(task.process)
+    untar        = archive.toString() - '.tar.gz'
+    """
+    tar -xzvf $options.args $archive
+    echo \$(tar --version 2>&1) | sed 's/^.*(GNU tar) //; s/ Copyright.*\$//' > ${software}.version.txt
+    """
+}

software/untar/meta.yml

@@ -0,0 +1,48 @@
+name: untar
+description: Extract files.
+keywords:
+    - untar
+    - uncompress
+tools:
+    - untar:
+        description: |
+            Extract tar.gz files.
+        documentation: https://www.gnu.org/software/tar/manual/
+params:
+    - outdir:
+        type: string
+        description: |
+            The pipeline's output directory. By default, the module will
+            output files into `$params.outdir/<SOFTWARE>`
+    - publish_dir_mode:
+        type: string
+        description: |
+            Value for the Nextflow `publishDir` mode parameter.
+            Available: symlink, rellink, link, copy, copyNoFollow, move.
+    - enable_conda:
+        type: boolean
+        description: |
+            Run the module with Conda using the software specified
+            via the `conda` directive
+    - singularity_pull_docker_container:
+        type: boolean
+        description: |
+            Instead of directly downloading Singularity images for use with Singularity,
+            force the workflow to pull and convert Docker containers instead.
+input:
+    - archive:
+        type: file
+        description: File to be untar
+        pattern: "*.{tar}.{gz}"
+output:
+    - untar:
+        type: file
+        description:
+        pattern: "*.*"
+    - version:
+        type: file
+        description: File containing software version
+        pattern: "*.{version.txt}"
+authors:
+    - "@joseespinosa"
+    - "@drpatelh"

tests/software/untar/main.nf

@@ -0,0 +1,13 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { UNTAR } from '../../../software/untar/main.nf' addParams( options: [:] )
+
+workflow test_untar {
+
+    def input = []
+    input = [ file("${launchDir}/tests/data/tar_gz/test.txt.tar.gz", checkIfExists: true) ]
+
+    UNTAR ( input )
+}

tests/software/untar/test.yml

@@ -0,0 +1,7 @@
+- name: untar
+  command: nextflow run ./tests/software/untar -entry test_untar -c tests/config/nextflow.config
+  tags:
+    - untar
+  files:
+    - path: output/untar/test.txt
+      md5sum: 7171d4ec2ad0882fab4b0142536d93a1