add bedtools bamtobed module (from nanoseq modules) (#466)

* add bedtools bamtobed module

* fix erros

* fix linting problem

software/bedtools/bamtobed/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/bedtools/bamtobed/main.nf

@@ -0,0 +1,34 @@
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName } from './functions'
+
+params.options = [:]
+def options    = initOptions(params.options)
+
+process BEDTOOLS_BAMTOBED {
+    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::bedtools=2.29.2" : null)
+    container "quay.io/biocontainers/bedtools:2.29.2--hc088bd4_0"
+
+    input:
+    tuple val(meta), path(sizes), path(bam), path(bai)
+
+    output:
+    tuple val(meta), path(sizes), path("*.bed12"), emit: bed12
+    path "*.version.txt"                         , emit: version
+
+    script:
+    """
+    bedtools \\
+        bamtobed \\
+        -bed12 \\
+        -cigar \\
+        -i ${bam[0]} \\
+        | bedtools sort > ${meta.id}.bed12
+    bedtools --version | sed -e "s/bedtools v//g" > bedtools.version.txt
+    """
+}

software/bedtools/bamtobed/meta.yml

@@ -0,0 +1,40 @@
+name: bedtools_bamtobed
+description: Converts a bam file to a bed12 file.
+keywords:
+    - bam
+    - bed
+tools:
+    - bedtools:
+        description: |
+            A set of tools for genomic analysis tasks, specifically enabling genome arithmetic (merge, count, complement) on various file types.
+        documentation: https://bedtools.readthedocs.io/en/latest/content/tools/complement.html
+input:
+    - meta:
+        type: map
+        description: |
+            Groovy Map containing sample information
+            e.g. [ id:'test', single_end:false ]
+    - bam:
+        type: file
+        description: Input BAM file
+        pattern: "*.{bam}"
+    - sizes:
+        type: file
+        description: File which defines the chromosome lengths for a given genome
+        pattern: "*.{sizes}"
+output:
+    - meta:
+        type: map
+        description: |
+            Groovy Map containing sample information
+            e.g. [ id:'test', single_end:false ]
+    - bed12:
+        type: file
+        description: Bed file with 12 columns containing all genomic intervals.
+        pattern: "*.{bed}"
+    - version:
+        type: file
+        description: File containing software version
+        pattern: "*.{version.txt}"
+authors:
+    - "@yuukiiwa"

tests/config/pytest_software.yml

@@ -42,6 +42,10 @@ bcftools/stats:
   - software/bcftools/stats/**
   - tests/software/bcftools/stats/**
 
+bedtools/bamtobed:
+  - software/bedtools/bamtobed/**
+  - tests/software/bedtools/bamtobed/**
+
 bedtools/complement:
   - software/bedtools/complement/**
   - tests/software/bedtools/complement/**

tests/software/bedtools/bamtobed/main.nf

@@ -0,0 +1,15 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { BEDTOOLS_BAMTOBED } from '../../../../software/bedtools/bamtobed/main.nf' addParams( options: [:] )
+
+workflow test_bedtools_bamtobed {
+    input = [ [ id:'test'], //meta map
+              file(params.test_data['sarscov2']['genome']['genome_sizes'], checkIfExists: true),
+              file(params.test_data['sarscov2']['nanopore']['test_sorted_bam'], checkIfExists: true),
+              file(params.test_data['sarscov2']['nanopore']['test_sorted_bam_bai'], checkIfExists: true)
+            ]
+
+    BEDTOOLS_BAMTOBED ( input )
+}

tests/software/bedtools/bamtobed/test.yml

@@ -0,0 +1,8 @@
+- name: bedtools bamtobed
+  command: nextflow run ./tests/software/bedtools/bamtobed -entry test_bedtools_bamtobed -c tests/config/nextflow.config
+  tags:
+    - bedtools
+    - bedtools/bamtobed
+  files:
+    - path: ./output/bedtools/test.bed12
+      md5sum: d41d8cd98f00b204e9800998ecf8427e