Add modules bgziptabix (#544)

* feat: add modules bgziptabix

* fix: add module to config

* fix: lint

* fix: tests

software/tabix/bgziptabix/functions.nf

@@ -0,0 +1,68 @@
+//
+//  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/tabix/bgziptabix/main.nf

@@ -0,0 +1,37 @@
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName } from './functions'
+
+params.options = [:]
+options        = initOptions(params.options)
+
+process TABIX_BGZIPTABIX {
+    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::tabix=0.2.6" : null)
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container "https://depot.galaxyproject.org/singularity/tabix:0.2.6--ha92aebf_0"
+    } else {
+        container "quay.io/biocontainers/tabix:0.2.6--ha92aebf_0"
+    }
+
+    input:
+    tuple val(meta), path(input)
+
+    output:
+    tuple val(meta), path("*.gz"), path("*.tbi"), emit: tbi
+    path  "*.version.txt",                        emit: version
+
+    script:
+    def software = getSoftwareName(task.process)
+    def prefix   = options.suffix ? "${meta.id}${options.suffix}" : "${meta.id}"
+    """
+    bgzip -c $options.args $input > ${prefix}.gz
+    tabix $options.args2 ${prefix}.gz
+
+    echo \$(tabix -h 2>&1) | sed 's/^.*Version: //; s/(.*\$//' > ${software}.version.txt
+    """
+}

software/tabix/bgziptabix/meta.yml

@@ -0,0 +1,44 @@
+name: tabix_bgziptabix
+description: bgzip a sorted tab-delimited genome file and then create tabix index
+keywords:
+    - bgzip
+    - compress
+    - index
+    - tabix
+    - vcf
+tools:
+    - tabix:
+        description: Generic indexer for TAB-delimited genome position files.
+        homepage: https://www.htslib.org/doc/tabix.html
+        documentation: https://www.htslib.org/doc/tabix.1.html
+        doi: 10.1093/bioinformatics/btq671
+input:
+    - meta:
+        type: map
+        description: |
+          Groovy Map containing sample information
+          e.g. [ id:'test', single_end:false ]
+    - tab:
+        type: file
+        description: TAB-delimited genome position file
+        pattern: "*.{bed,gff,sam,vcf}"
+output:
+    - meta:
+        type: map
+        description: |
+          Groovy Map containing sample information
+          e.g. [ id:'test', single_end:false ]
+    - gz:
+        type: file
+        description: Output compressed file
+        pattern: "*.{gz}"
+    - tbi:
+        type: file
+        description: tabix index file
+        pattern: "*.{gz.tbi}"
+    - version:
+        type: file
+        description: File containing software version
+        pattern: "*.{version.txt}"
+authors:
+    - "@maxulysse"

tests/config/pytest_software.yml

@@ -699,6 +699,10 @@ tabix/bgzip:
   - software/tabix/bgzip/**
   - tests/software/tabix/bgzip/**
 
+tabix/bgziptabix:
+  - software/tabix/bgziptabix/**
+  - tests/software/tabix/bgziptabix/**
+
 tabix/tabix:
   - software/tabix/tabix/**
   - tests/software/tabix/tabix/**

tests/software/tabix/bgziptabix/main.nf

@@ -0,0 +1,13 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { TABIX_BGZIPTABIX } from '../../../../software/tabix/bgziptabix/main.nf' addParams( options: ['args2': '-p vcf'] )
+
+workflow test_tabix_bgziptabix {
+    input = [ [ id:'test' ], // meta map
+              [ file(params.test_data['sarscov2']['illumina']['test_vcf'], checkIfExists: true) ]
+            ]
+
+    TABIX_BGZIPTABIX ( input )
+}

tests/software/tabix/bgziptabix/test.yml

@@ -0,0 +1,10 @@
+- name: tabix bgziptabix
+  command: nextflow run ./tests/software/tabix/bgziptabix -entry test_tabix_bgziptabix -c tests/config/nextflow.config
+  tags:
+    - tabix
+    - tabix/bgziptabix
+  files:
+    - path: ./output/tabix/test.gz
+      md5sum: 0f1c94af3aa3e7e203d9e034ef6f8f4d
+    - path: ./output/tabix/test.gz.tbi
+      md5sum: bbec39fd53cf2834909d52094980d094