Seqtk seq (#1340)

* Initial commit of seqtk/seq module files

* pytest.yml

* updated module and tests code, need to finish modules/main.nf

* Initial commit of seqtk/seq module files

* pytest.yml

* updated module and tests code, need to finish modules/main.nf

* Adding code and configs for seqtk/seq module

* Re-tested module following minor code update

* removed trailing whitespace errors

* Changed variable name  to  following reviewer suggestions

Co-authored-by: Sateesh <33637490+sateeshperi@users.noreply.github.com>

modules/seqtk/seq/main.nf

@@ -0,0 +1,40 @@
+process SEQTK_SEQ {
+    tag "$meta.id"
+    label 'process_low'
+
+    conda (params.enable_conda ? "bioconda::seqtk=1.3" : null)
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/seqtk:1.3--h5bf99c6_3' :
+        'quay.io/biocontainers/seqtk:1.3--h5bf99c6_3' }"
+
+    input:
+    tuple val(meta), path(fastx)
+
+    output:
+    tuple val(meta), path("*.gz")     , emit: fastx
+    path "versions.yml"               , emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args = task.ext.args ?: ''
+    def prefix = task.ext.prefix ?: "${meta.id}"
+
+    def extension = "fastq"
+    if ("$fastx" ==~ /.+\.fasta|.+\.fasta.gz|.+\.fa|.+\.fa.gz|.+\.fas|.+\.fas.gz|.+\.fna|.+\.fna.gz/ || "$args" ==~ /\-[aA]/ ) {
+        extension = "fasta"
+    }
+    """
+    seqtk \\
+        seq \\
+        $args \\
+        $fastx | \\
+        gzip -c > ${prefix}.seqtk-seq.${extension}.gz
+
+    cat <<-END_VERSIONS > versions.yml
+        "${task.process}":
+            seqtk: \$(echo \$(seqtk 2>&1) | sed 's/^.*Version: //; s/ .*\$//')
+    END_VERSIONS
+    """
+}

modules/seqtk/seq/meta.yml

@@ -0,0 +1,42 @@
+name: seqtk_seq
+description: Common transformation operations on FASTA or FASTQ files.
+keywords:
+  - seq
+tools:
+  - seqtk:
+      description: Seqtk is a fast and lightweight tool for processing sequences in the FASTA or FASTQ format. The seqtk seq command enables common transformation operations on FASTA or FASTQ files.
+      homepage: https://github.com/lh3/seqtk
+      documentation: https://docs.csc.fi/apps/seqtk/
+      tool_dev_url: https://github.com/lh3/seqtk
+      licence: ['MIT']
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test' ]
+  - sequences:
+      type: file
+      description: A FASTQ or FASTA file
+      pattern: "*.{fastq.gz, fastq, fq, fq.gz, fasta, fastq.gz, fa, fa.gz, fas, fas.gz, fna, fna.gz}"
+
+output:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test' ]
+  - versions:
+      type: file
+      description: File containing software versions
+      pattern: "versions.yml"
+  - sequences:
+      type: file
+      description: FASTQ/FASTA file containing renamed sequences
+      pattern: "*.{fastq.gz, fasta.gz}"
+
+authors:
+  - "@hseabolt"
+  - "@mjcipriano"
+  - "@sateeshperi"

tests/config/pytest_modules.yml

@@ -1461,6 +1461,10 @@ seqtk/sample:
   - modules/seqtk/sample/**
   - tests/modules/seqtk/sample/**
 
+seqtk/seq:
+  - modules/seqtk/seq/**
+  - tests/modules/seqtk/seq/**
+
 seqtk/subseq:
   - modules/seqtk/subseq/**
   - tests/modules/seqtk/subseq/**

tests/modules/seqtk/seq/main.nf

@@ -0,0 +1,19 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { SEQTK_SEQ } from '../../../../modules/seqtk/seq/main.nf'
+
+workflow test_seqtk_seq {
+    input = [ [ id:'test' ],   // meta map
+              [ file(params.test_data['sarscov2']['genome']['genome_fasta'], checkIfExists: true) ]
+            ]
+    SEQTK_SEQ ( input )
+}
+
+workflow test_seqtk_seq_fq {
+    input = [ [ id:'test' ], // meta map
+              [ file(params.test_data['sarscov2']['illumina']['test_1_fastq_gz'], checkIfExists: true) ]
+            ]
+    SEQTK_SEQ ( input )
+}

tests/modules/seqtk/seq/nextflow.config

@@ -0,0 +1,7 @@
+process {
+    // Testing ext.args for passing arguments into seqtk seq
+    withName: 'SEQTK_SEQ' {
+        ext.args = '-A'
+        publishDir = { "${params.outdir}/${task.process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()}" }
+    }
+}

tests/modules/seqtk/seq/test.yml

@@ -0,0 +1,21 @@
+- name: seqtk seq test_seqtk_seq
+  command: nextflow run tests/modules/seqtk/seq -entry test_seqtk_seq -c tests/config/nextflow.config
+  tags:
+    - seqtk/seq
+    - seqtk
+  files:
+    - path: output/seqtk/test.seqtk-seq.fasta.gz
+      md5sum: 50d73992c8c7e56dc095ef47ec52a754
+    - path: output/seqtk/versions.yml
+      md5sum: 2b89cd4a6e28f35fcfbbd2188384f944
+
+- name: seqtk seq test_seqtk_seq_fq
+  command: nextflow run tests/modules/seqtk/seq -entry test_seqtk_seq_fq -c tests/config/nextflow.config
+  tags:
+    - seqtk/seq
+    - seqtk
+  files:
+    - path: output/seqtk/test.seqtk-seq.fasta.gz
+      md5sum: 2f009f1647971a97b4edec726a99dc1a
+    - path: output/seqtk/versions.yml
+      md5sum: 3467a76d3540bee8f58de050512bddaa