Mskcc custom/splitfastqbylane

modules/nf-core/custom/splitfastqbylane/main.nf

@@ -0,0 +1,35 @@
+process CUSTOM_SPLITFASTQBYLANE {
+    tag "$meta.id"
+    label 'process_single'
+
+    conda "anaconda::gawk=5.1.0"
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/gawk:5.1.0' :
+        'biocontainers/gawk:5.1.0' }"
+
+    input:
+    tuple val(meta), path(reads)
+
+    output:
+    tuple val(meta), path("*.split.fastq.gz"), emit: reads
+    path "versions.yml"                      , emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    shell:
+    args = task.ext.args ?: ''
+    prefix = task.ext.prefix ?: "${meta.id}"
+    read1 = [reads].flatten()[0]
+    read2 = [reads].flatten().size() > 1 ? reads[1] : null
+    template 'split_lanes_awk.sh'
+
+    stub:
+    """
+    touch out.split.fastq.gz
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        gawk: \$(awk -Wversion | sed '1!d; s/.*Awk //; s/,.*//')
+    END_VERSIONS
+    """
+}

modules/nf-core/custom/splitfastqbylane/meta.yml

@@ -0,0 +1,46 @@
+name: "custom_splitfastqbylane"
+description: |
+  Splits fastq files with multiple or unknown number of merged flowcell+lane sources into separate fastq files, each with single flowcell+line sources.
+keywords:
+  - splitlanesbyfastq
+  - awk
+  - custom
+  - fastq
+  - split
+  - lanes
+tools:
+  - "gawk":
+      description: "The awk utility interprets a special-purpose programming language that makes it easy to handle simple data-reformatting jobs."
+      homepage: "https://www.gnu.org/software/gawk/manual/gawk.html"
+      documentation: "https://www.gnu.org/software/gawk/manual/gawk.html"
+      tool_dev_url: "http://savannah.gnu.org/projects/gawk/"
+      licence: "['GPL v3']"
+
+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,fastq.gz}"
+
+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"
+  - reads:
+      type: file
+      description: List of output FastQ files of size of multiples of 1 and 2, for single-end and paired-end data, respectively.
+      pattern: "*.split.fastq.gz"
+
+authors:
+  - "@anoronh4"

modules/nf-core/custom/splitfastqbylane/templates/split_lanes_awk.sh

@@ -0,0 +1,36 @@
+#!/bin/bash
+
+if [[ "!{read1}" == *gz ]] ; then
+    cat_="zcat"
+else
+    cat_="cat"
+fi
+
+function a() {
+    awk \
+        -v prefix=!{prefix} \
+        -v readnumber=$1 \
+        '
+        BEGIN {FS = ":"}
+        {
+            lane=$(NF-3)
+            flowcell=$(NF-4)
+            outfastq=prefix"@"flowcell"_L00"lane"_R"readnumber".split.fastq.gz"
+            print | "gzip > "outfastq
+            for (i = 1; i <= 3; i++) {
+                getline
+                print | "gzip > "outfastq
+            }
+        }
+        ' <( eval "$cat_ $2")
+}
+
+a 1 !{read1}
+if [ ! -z !{read2} ] ; then
+    a 2 !{read2}
+fi
+
+cat <<-END_VERSIONS > versions.yml
+"!{task.process}":
+    gawk: $(awk -Wversion | sed '1!d; s/.*Awk //; s/,.*//')
+END_VERSIONS

tests/config/pytest_modules.yml

@@ -866,6 +866,10 @@ custom/matrixfilter:
   - modules/nf-core/custom/matrixfilter/**
   - tests/modules/nf-core/custom/matrixfilter/**
 
+custom/splitfastqbylane:
+  - modules/nf-core/custom/splitfastqbylane/**
+  - tests/modules/nf-core/custom/splitfastqbylane/**
+
 custom/sratoolsncbisettings:
   - modules/nf-core/custom/sratoolsncbisettings/**
   - tests/modules/nf-core/custom/sratoolsncbisettings/**

tests/modules/nf-core/custom/splitfastqbylane/main.nf

@@ -0,0 +1,27 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { CUSTOM_SPLITFASTQBYLANE } from '../../../../../modules/nf-core/custom/splitfastqbylane/main.nf'
+
+workflow test_custom_splitfastqbylane {
+
+    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)
+	]
+    ]
+
+    CUSTOM_SPLITFASTQBYLANE ( input )
+}
+
+workflow test_custom_splitfastqbylane_single_end { 
+
+    input = [
+         [ id:'test', single_end:true ],
+	 file(params.test_data['sarscov2']['illumina']['test_1_fastq_gz'], checkIfExists: true)
+    ]
+    CUSTOM_SPLITFASTQBYLANE ( input )
+}

tests/modules/nf-core/custom/splitfastqbylane/nextflow.config

@@ -0,0 +1,5 @@
+process {
+
+    publishDir = { "${params.outdir}/${task.process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()}" }
+
+}

tests/modules/nf-core/custom/splitfastqbylane/test.yml

@@ -0,0 +1,31 @@
+- name: custom splitfastqbylane test_custom_splitfastqbylane
+  command: nextflow run ./tests/modules/nf-core/custom/splitfastqbylane -entry test_custom_splitfastqbylane -c ./tests/config/nextflow.config -c ./tests/modules/nf-core/custom/splitfastqbylane/nextflow.config
+  tags:
+    - custom
+    - custom/splitfastqbylane
+  files:
+    - path: output/custom/test@HK3MMAFX2_L001_R1.split.fastq.gz
+    - path: output/custom/test@HK3MMAFX2_L001_R2.split.fastq.gz
+    - path: output/custom/test@HK3MMAFX2_L002_R1.split.fastq.gz
+    - path: output/custom/test@HK3MMAFX2_L002_R2.split.fastq.gz
+    - path: output/custom/test@HK3MMAFX2_L003_R1.split.fastq.gz
+    - path: output/custom/test@HK3MMAFX2_L003_R2.split.fastq.gz
+    - path: output/custom/test@HK3MMAFX2_L004_R1.split.fastq.gz
+    - path: output/custom/test@HK3MMAFX2_L004_R2.split.fastq.gz
+    - path: output/custom/versions.yml
+      contains:
+        - "gawk"
+
+- name: custom splitfastqbylane test_custom_splitfastqbylane_single_end
+  command: nextflow run ./tests/modules/nf-core/custom/splitfastqbylane -entry test_custom_splitfastqbylane_single_end -c ./tests/config/nextflow.config -c ./tests/modules/nf-core/custom/splitfastqbylane/nextflow.config
+  tags:
+    - custom
+    - custom/splitfastqbylane
+  files:
+    - path: output/custom/test@HK3MMAFX2_L001_R1.split.fastq.gz
+    - path: output/custom/test@HK3MMAFX2_L002_R1.split.fastq.gz
+    - path: output/custom/test@HK3MMAFX2_L003_R1.split.fastq.gz
+    - path: output/custom/test@HK3MMAFX2_L004_R1.split.fastq.gz
+    - path: output/custom/versions.yml
+      contains:
+        - "gawk"