Add a module for sra-tools prefetch (#714)

* chore: apply module template

* refactor: add NCBI settings to options

* docs: complete meta information

* feat: add prefetch process

* fix: correct bash commands

* tests: define the right tests

* style: move option definition to satisfy linting

* fix: extract version correctly

* fix: correct newline issues

* refactor: address review comments

* Apply suggestions from code review

* chore: add retrying via nf-core label

* refactor: validate download thoroughly

* refactor: remove vdb-config input

Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com>

modules/sratools/prefetch/functions.nf

@@ -0,0 +1,78 @@
+//
+//  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()
+}
+
+//
+// Extract name of module from process name using $task.process
+//
+def getProcessName(task_process) {
+    return task_process.tokenize(':')[-1]
+}
+
+//
+// Function to initialise default values and to generate a Groovy Map of available options for nf-core modules
+//
+def initOptions(Map args) {
+    def Map options = [:]
+    options.args            = args.args ?: ''
+    options.args2           = args.args2 ?: ''
+    options.args3           = args.args3 ?: ''
+    options.publish_by_meta = args.publish_by_meta ?: []
+    options.publish_dir     = args.publish_dir ?: ''
+    options.publish_files   = args.publish_files
+    options.suffix          = args.suffix ?: ''
+    return options
+}
+
+//
+// Tidy up and join elements of a list to return a path string
+//
+def getPathFromList(path_list) {
+    def paths = path_list.findAll { item -> !item?.trim().isEmpty() }      // Remove empty entries
+    paths     = paths.collect { it.trim().replaceAll("^[/]+|[/]+\$", "") } // Trim whitespace and trailing slashes
+    return paths.join('/')
+}
+
+//
+// Function to save/publish module results
+//
+def saveFiles(Map args) {
+    def ioptions  = initOptions(args.options)
+    def path_list = [ ioptions.publish_dir ?: args.publish_dir ]
+
+    // Do not publish versions.yml unless running from pytest workflow
+    if (args.filename.equals('versions.yml') && !System.getenv("NF_CORE_MODULES_TEST")) {
+        return null
+    }
+    if (ioptions.publish_by_meta) {
+        def key_list = ioptions.publish_by_meta instanceof List ? ioptions.publish_by_meta : args.publish_by_meta
+        for (key in key_list) {
+            if (args.meta && key instanceof String) {
+                def path = key
+                if (args.meta.containsKey(key)) {
+                    path = args.meta[key] instanceof Boolean ? "${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"
+    }
+}

modules/sratools/prefetch/main.nf

@@ -0,0 +1,50 @@
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName; getProcessName } from './functions'
+
+params.options = [:]
+options        = initOptions(params.options)
+
+process SRATOOLS_PREFETCH {
+    tag "$id"
+    label 'process_low'
+    label 'error_retry'
+    publishDir "${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::sra-tools=2.11.0' : null)
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container 'https://depot.galaxyproject.org/singularity/sra-tools:2.11.0--pl5262h314213e_0'
+    } else {
+        container 'quay.io/biocontainers/sra-tools:2.11.0--pl5262h314213e_0'
+    }
+
+    input:
+    tuple val(meta), val(id)
+
+    output:
+    tuple val(meta), path("$id"), emit: sra
+    path "versions.yml"         , emit: versions
+
+    script:
+    def config = "/LIBS/GUID = \"${UUID.randomUUID().toString()}\"\\n/libs/cloud/report_instance_identity = \"true\"\\n"
+    """
+    eval "\$(vdb-config -o n NCBI_SETTINGS | sed 's/[" ]//g')"
+    if [[ ! -f "\${NCBI_SETTINGS}" ]]; then
+        mkdir -p "\$(dirname "\${NCBI_SETTINGS}")"
+        printf '${config}' > "\${NCBI_SETTINGS}"
+    fi
+
+    prefetch \\
+        $options.args \\
+        --progress \\
+        $id
+
+    vdb-validate $id
+
+    cat <<-END_VERSIONS > versions.yml
+    ${getProcessName(task.process)}:
+        ${getSoftwareName(task.process)}: \$(prefetch --version 2>&1 | grep -Eo '[0-9.]+')
+    END_VERSIONS
+    """
+}

modules/sratools/prefetch/meta.yml

@@ -0,0 +1,43 @@
+name: sratools_prefetch
+description: Download sequencing data from the NCBI Sequence Read Archive (SRA).
+keywords:
+  - sequencing
+  - fastq
+  - prefetch
+tools:
+  - sratools:
+      description: SRA Toolkit and SDK from NCBI
+      homepage: https://github.com/ncbi/sra-tools
+      documentation: https://github.com/ncbi/sra-tools/wiki
+      tool_dev_url: https://github.com/ncbi/sra-tools
+      licence: ['Public Domain']
+
+input:
+  - meta:
+      type: map
+      description: >
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - id:
+      type: val
+      description: >
+        A string denoting an SRA id.
+
+output:
+  - meta:
+      type: map
+      description: >
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - sra:
+      type: directory
+      description: >
+        Directory containing the ETL data for the given SRA id.
+      pattern: "*/*.sra"
+  - versions:
+      type: file
+      description: File containing software versions
+      pattern: "versions.yml"
+
+authors:
+  - "@Midnighter"

tests/config/pytest_modules.yml

@@ -529,6 +529,10 @@ iqtree:
   - modules/iqtree/**
   - tests/modules/iqtree/**
 
+ismapper:
+  - modules/ismapper/**
+  - tests/modules/ismapper/**
+
 isoseq3/cluster:
   - modules/isoseq3/cluster/**
   - tests/modules/isoseq3/cluster/**
@@ -537,10 +541,6 @@ isoseq3/refine:
   - modules/isoseq3/refine/**
   - tests/modules/isoseq3/refine/**
 
-ismapper:
-  - modules/ismapper/**
-  - tests/modules/ismapper/**
-
 ivar/consensus:
   - modules/ivar/consensus/**
   - tests/modules/ivar/consensus/**
@@ -979,6 +979,10 @@ spatyper:
   - modules/spatyper/**
   - tests/modules/spatyper/**
 
+sratools/prefetch:
+  - modules/sratools/prefetch/**
+  - tests/modules/sratools/prefetch/**
+
 staphopiasccmec:
   - modules/staphopiasccmec/**
   - tests/modules/staphopiasccmec/**

tests/modules/sratools/prefetch/main.nf

@@ -0,0 +1,15 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { SRATOOLS_PREFETCH } from '../../../../modules/sratools/prefetch/main.nf' addParams( options: [:] )
+
+workflow test_sratools_prefetch {
+
+    input = [
+        [ id:'test', single_end:false ], // meta map
+        'ERR2815334'
+    ]
+
+    SRATOOLS_PREFETCH ( input )
+}

tests/modules/sratools/prefetch/test.yml

@@ -0,0 +1,8 @@
+- name: sratools prefetch test_sratools_prefetch
+  command: nextflow run tests/modules/sratools/prefetch -entry test_sratools_prefetch -c tests/config/nextflow.config
+  tags:
+    - sratools/prefetch
+    - sratools
+  files:
+    - path: output/sratools/ERR2815334/ERR2815334.sra
+      md5sum: 9a98c7f6f4774b7ef94aa915b92a54ea