Add ucsc/bedclip module (#450)

* Add ucsc/bedclip module

* Fix tests

* Fix nf-core lint

software/ucsc/bedclip/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/ucsc/bedclip/main.nf

@@ -0,0 +1,42 @@
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName } from './functions'
+
+params.options = [:]
+options        = initOptions(params.options)
+
+def VERSION = '377'
+
+process UCSC_BEDCLIP {
+    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::ucsc-bedclip=377" : null)
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container "https://depot.galaxyproject.org/singularity/ucsc-bedclip:377--h0b8a92a_2"
+    } else {
+        container "quay.io/biocontainers/ucsc-bedclip:377--h0b8a92a_2"
+    }
+
+    input:
+    tuple val(meta), path(bedgraph)
+    path  sizes
+
+    output:
+    tuple val(meta), path("*.bedGraph"), emit: bedgraph
+    path "*.version.txt"               , emit: version
+
+    script:
+    def software = getSoftwareName(task.process)
+    def prefix   = options.suffix ? "${meta.id}${options.suffix}" : "${meta.id}"
+    """
+    bedClip \\
+        $bedgraph \\
+        $sizes \\
+        ${prefix}.bedGraph
+
+    echo $VERSION > ${software}.version.txt
+    """
+}

software/ucsc/bedclip/meta.yml

@@ -0,0 +1,41 @@
+name: ucsc_bedclip
+description: See http://hgdownload.cse.ucsc.edu/admin/exe/
+keywords:
+  - sort
+tools:
+  - ucsc:
+      description: Remove lines from bed file that refer to off-chromosome locations.
+      homepage: None
+      documentation: None
+      tool_dev_url: None
+      doi: ""
+      licence: ['varies; see http://genome.ucsc.edu/license']
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - bedgraph:
+      type: file
+      description: bedGraph file
+      pattern: "*.{bedgraph}"
+
+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}"
+  - bedgraph:
+      type: file
+      description: bedGraph file
+      pattern: "*.{bedgraph}"
+
+authors:
+  - "@drpatelh"

tests/config/pytest_software.yml

@@ -571,6 +571,10 @@ ucsc/bed12tobigbed:
   - software/ucsc/bed12tobigbed/**
   - tests/software/ucsc/bed12tobigbed/**
 
+ucsc/bedclip:
+  - software/ucsc/bedclip/**
+  - tests/software/ucsc/bedclip/**
+
 ucsc/bedgraphtobigwig:
   - software/ucsc/bedgraphtobigwig/**
   - tests/software/ucsc/bedgraphtobigwig/**

tests/software/ucsc/bedclip/main.nf

@@ -0,0 +1,14 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { UCSC_BEDCLIP } from '../../../../software/ucsc/bedclip/main.nf' addParams( options: [suffix:'.clip'] )
+
+workflow test_ucsc_bedclip {
+    input = [ [ id:'test', single_end:false ], // meta map
+              file(params.test_data['sarscov2']['illumina']['test_bedgraph'], checkIfExists: true)
+            ]
+    sizes = file(params.test_data['sarscov2']['genome']['genome_sizes'], checkIfExists: true)
+
+    UCSC_BEDCLIP ( input, sizes )
+}

tests/software/ucsc/bedclip/test.yml

@@ -0,0 +1,8 @@
+- name: ucsc bedclip
+  command: nextflow run ./tests/software/ucsc/bedclip -entry test_ucsc_bedclip -c tests/config/nextflow.config
+  tags:
+    - ucsc
+    - ucsc/bedclip
+  files:
+    - path: output/ucsc/test.clip.bedGraph
+      md5sum: e02395e1f7c593b3f79563067159ebc2