Add bases2fastq (#2054)

* feat(bases2fastq): nf-core modules create

* test(bases2fastq): Update tests

* docs(bases2fastq): Update meta.yml

* test: Add untar rundir

modules/bases2fastq/main.nf

@@ -0,0 +1,45 @@
+process BASES2FASTQ {
+    tag "$meta.id"
+    label 'process_high'
+
+    if (params.enable_conda) {
+        exit 1, "Conda environments cannot be used when using bases2fastq. Please use docker or singularity containers."
+    }
+    container "elembio/bases2fastq:1.1.0"
+
+    input:
+    tuple val(meta), path(run_manifest), path(run_dir)
+
+    output:
+    tuple val(meta), path('output/Samples/*/*_R*.fastq.gz'), emit: sample_fastq
+    tuple val(meta), path('output/Samples/*/*.json')       , emit: sample_json
+    tuple val(meta), path('output/*.html')                 , emit: qc_report
+    tuple val(meta), path('output/RunStats.json')          , emit: run_stats
+    tuple val(meta), path('output/RunManifest.json')       , emit: generated_run_manifest
+    tuple val(meta), path('output/Metrics.csv')            , emit: metrics
+    tuple val(meta), path('output/UnassignedSequences.csv'), emit: unassigned
+    path "versions.yml"                             , emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args = task.ext.args ?: ''
+    def prefix = task.ext.prefix ?: "${meta.id}"
+    def runManifest = run_manifest ? "-r ${run_manifest}" : ""
+    """
+    ls
+
+    bases2fastq \\
+        -p $task.cpus \\
+        $runManifest \\
+        $args \\
+        $run_dir \\
+        output
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        bases2fastq: \$(bases2fastq --version | sed -e "s/bases2fastq version //g")
+    END_VERSIONS
+    """
+}

modules/bases2fastq/meta.yml

@@ -0,0 +1,68 @@
+name: "bases2fastq"
+description: Demultiplex Element Biosciences bases files
+keywords:
+  - demultiplex
+  - element
+  - fastq
+tools:
+  - "bases2fastq":
+      description: "Demultiplexes sequencing data and converts base calls into FASTQ files for secondary analysis"
+      homepage: "https://go.elementbiosciences.com/bases2fastq-download"
+      documentation: "https://www.elementbiosciences.com/resources/user-guides/workflow/bases2fastq"
+      licence: "http://go.elembio.link/eula"
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - run_manifest:
+      type: file
+      description: RunManifest file
+      pattern: "*.{csv}"
+  - run_dir:
+      type: directory
+      description: "Input run directory containing optionally containing a RunManifest.json if run_manifest is not supplied"
+
+output:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - versions:
+      type: file
+      description: File containing software versions
+      pattern: "versions.yml"
+  - sample_fastq:
+      type: file
+      description: Demultiplexed sample FASTQ files
+      pattern: "output/Samples/*/*_R*.fastq.gz"
+  - sample_json:
+      type: file
+      description: Demultiplexed sample stats
+      pattern: "output/Samples/*/*.json"
+  - qc_report:
+      type: file
+      description: QC HTML report
+      pattern: "output/*.html"
+  - run_stats:
+      type: file
+      description: QC HTML report
+      pattern: "output/*.html"
+  - generated_run_manifest:
+      type: file
+      description: Updated Run Manifest JSON from the run_manifest csv
+      pattern: "output/RunManifest.json"
+  - metrics:
+      type: file
+      description: Sample metrics
+      pattern: "output/Metrics.csv"
+  - unassigned:
+      type: file
+      description: Unassigned Sequences
+      pattern: "output/UnassignedSequences.csv"
+
+authors:
+  - "@Emiller88"

tests/config/pytest_modules.yml

@@ -134,6 +134,10 @@ bandage/image:
   - modules/bandage/image/**
   - tests/modules/bandage/image/**
 
+bases2fastq:
+  - modules/bases2fastq/**
+  - tests/modules/bases2fastq/**
+
 bbmap/align:
   - modules/bbmap/align/**
   - tests/modules/bbmap/align/**

tests/modules/bases2fastq/main.nf

@@ -0,0 +1,23 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { BASES2FASTQ } from '../../../modules/bases2fastq/main.nf'
+include { UNTAR      } from '../../../modules/untar/main.nf'
+
+workflow test_bases2fastq {
+    
+    input = Channel.value([
+        [ id:'sim-data' ], // meta map
+        file("https://raw.githubusercontent.com/nf-core/test-datasets/demultiplex/testdata/sim-data/RunManifest.csv", checkIfExists: true),
+    ])
+
+    ch_input = input.join(
+        UNTAR ( [
+            [ id:'sim-data' ],
+            file("https://github.com/nf-core/test-datasets/raw/demultiplex/testdata/sim-data/sim-data.tar.gz", checkIfExists: true)
+        ]).untar
+    )
+
+    BASES2FASTQ ( ch_input )
+}

tests/modules/bases2fastq/nextflow.config

@@ -0,0 +1,9 @@
+process {
+
+    publishDir = { "${params.outdir}/${task.process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()}" }
+
+    withName: UNTAR {
+        publishDir = [ enabled: false ]
+    }
+    
+}

tests/modules/bases2fastq/test.yml

@@ -0,0 +1,18 @@
+- name: "bases2fastq"
+  command: nextflow run ./tests/modules/bases2fastq -entry test_bases2fastq -c ./tests/config/nextflow.config -c ./tests/modules/bases2fastq/nextflow.config
+  tags:
+    - "bases2fastq"
+  files:
+    - path: output/bases2fastq/output/Bases2Fastq-Sim_QC.html
+    - path: output/bases2fastq/output/Metrics.csv
+      md5sum: 0ec6da2b82e191098283474356024abf
+    - path: output/bases2fastq/output/RunManifest.json
+      md5sum: a07dce8ee25c2a6f9355b677c26b53e2
+    - path: output/bases2fastq/output/RunStats.json
+    - path: output/bases2fastq/output/UnassignedSequences.csv
+      md5sum: 11c1693830ce941b8cfb8d2431a59097
+    - path: output/bases2fastq/output/Samples/DefaultSample/DefaultSample_R1.fastq.gz
+      md5sum: 218abc70f61e8e8199a68f83ae836184
+    - path: output/bases2fastq/output/Samples/DefaultSample/DefaultSample_R2.fastq.gz
+      md5sum: b95109bfb204daa150b61239d3368ee2
+    - path: output/bases2fastq/output/Samples/DefaultSample/DefaultSample_stats.json