Request to review code for seqtk/sample module (#488)

* added files

* removed file

* added file

* changed files

* changed files

* edited file

* edited file

* edited files

* edited files

* edited files

* edited tags

* edited tags

* edited tags

* edited tags

Co-authored-by: kaurravneet4123 <kaurravneet4123@yahoo.com@users.noreply.github.com>

software/seqtk/sample/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/seqtk/sample/main.nf

@@ -0,0 +1,62 @@
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName } from './functions'
+
+params.options = [:]
+options        = initOptions(params.options)
+
+process SEQTK_SAMPLE {
+    tag "$meta.id"
+    label 'process_low'
+    publishDir "${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::seqtk=1.3" : null)
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container "https://depot.galaxyproject.org/singularity/seqtk:1.3--h5bf99c6_3"
+    } else {
+        container "quay.io/biocontainers/seqtk:1.3--h5bf99c6_3"
+    }
+
+    input:
+    tuple val(meta), path(reads)
+    val sample_size
+
+    output:
+    tuple val(meta), path("*.fastq.gz"), emit: reads
+    path "*.version.txt"               , emit: version
+
+    script:
+    def software = getSoftwareName(task.process)
+    def prefix   = options.suffix ? "${meta.id}${options.suffix}" : "${meta.id}"
+    if (meta.single_end) {
+        """
+        seqtk \\
+            sample \\
+            $options.args \\
+            $reads \\
+            $sample_size \\
+            | gzip > ${prefix}.fastq.gz \\
+
+        echo \$(seqtk 2>&1) | sed 's/^.*Version: //; s/ .*\$//' > ${software}.version.txt
+        """
+    } else {
+        """
+        seqtk \\
+            sample \\
+            $options.args \\
+            $reads[0] \\
+            $sample_size \\
+            | gzip > ${prefix}_1.fastq.gz \\
+
+        seqtk \\
+            sample \\
+            $options.args \\
+            $reads[1] \\
+            $sample_size \\
+            | gzip > ${prefix}_2.fastq.gz \\
+
+        echo \$(seqtk 2>&1) | sed 's/^.*Version: //; s/ .*\$//' > ${software}.version.txt
+        """
+    }
+}

software/seqtk/sample/meta.yml

@@ -0,0 +1,39 @@
+name: seqtk_sample
+description: Subsample reads from FASTQ files
+keywords:
+  - sample
+tools:
+  - seqtk:
+      description: Seqtk is a fast and lightweight tool for processing sequences in the FASTA or FASTQ format. Seqtk sample command subsamples sequences.
+      homepage: https://github.com/lh3/seqtk
+      documentation: https://docs.csc.fi/apps/seqtk/
+      licence: ['MIT']
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - reads:
+      type: file
+      description: List of input FastQ files of size 1 and 2 for single-end and paired-end data,respectively.
+      pattern: "*.{fastq.gz}"
+
+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}"
+  - reads:
+      type: file
+      description: Subsampled FastQ files of size 1 and 2 for single-end and paired-end data,respectively.
+      pattern: "*.{fastq.gz}"
+
+authors:
+  - "@kaurravneet4123"

tests/config/pytest_software.yml

@@ -563,6 +563,10 @@ seqkit/split2:
   - software/seqkit/split2/**
   - tests/software/seqkit/split2/**
 
+seqtk/sample:
+  - software/seqtk/sample/**
+  - tests/software/seqtk/sample/**
+
 sequenzautils/bam2seqz:
   - software/sequenzautils/bam2seqz/**
   - tests/software/sequenzautils/bam2seqz/**

tests/software/seqtk/sample/main.nf

@@ -0,0 +1,29 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { SEQTK_SAMPLE } from '../../../../software/seqtk/sample/main.nf' addParams( options: [ 'args': '-s100', 'suffix':'.sampled' ] )
+
+/*
+ * Test with single-end data
+ */
+workflow test_seqtk_sample_single_end {
+    
+    input = [ [ id:'test', single_end:true ], // meta map
+              file(params.test_data['sarscov2']['illumina']['test_1_fastq_gz'], checkIfExists: true) ]
+
+    SEQTK_SAMPLE ( input, 250000 )
+}
+
+/*
+ * Test with paired-end data
+ */
+workflow test_seqtk_sample_paired_end {
+
+    input = [ [ id:'test', single_end:false ], // meta map
+              [ file(params.test_data['sarscov2']['illumina']['test_1_fastq_gz'], checkIfExists: true),
+                file(params.test_data['sarscov2']['illumina']['test_2_fastq_gz'], checkIfExists: true) ]
+            ]
+
+    SEQTK_SAMPLE ( input, 250000 )
+}

tests/software/seqtk/sample/test.yml

@@ -0,0 +1,19 @@
+- name: seqtk sample single-end
+  command: nextflow run ./tests/software/seqtk/sample -entry test_seqtk_sample_single_end -c tests/config/nextflow.config
+  tags:
+    - seqtk
+    - seqtk/sample
+  files:
+    - path: output/seqtk/test.sampled.fastq.gz
+      md5sum: 875863b402f67403dac63ef59b9c9a8a
+
+- name: seqtk sample paired-end
+  command: nextflow run ./tests/software/seqtk/sample -entry test_seqtk_sample_paired_end -c tests/config/nextflow.config
+  tags:
+    - seqtk
+    - seqtk/sample
+  files:
+    - path: output/seqtk/test.sampled_1.fastq.gz
+      md5sum: 7029066c27ac6f5ef18d660d5741979a
+    - path: output/seqtk/test.sampled_2.fastq.gz
+      md5sum: 7029066c27ac6f5ef18d660d5741979a