Add centrifuge_kreport module (#1514)

modules/centrifuge/kreport/main.nf

@@ -0,0 +1,33 @@
+process CENTRIFUGE_KREPORT {
+    tag "$meta.id"
+    label 'process_low'
+
+    conda (params.enable_conda ? "bioconda::centrifuge=1.0.4_beta" : null)
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/centrifuge:1.0.4_beta--h9a82719_6':
+        'quay.io/biocontainers/centrifuge:1.0.4_beta--h9a82719_6' }"
+
+    input:
+    tuple val(meta), path(results)
+    path db
+
+    output:
+    tuple val(meta), path('*.txt')                , emit: kreport
+    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}"
+    """
+    db_name=`find -L ${db} -name "*.1.cf" -not -name "._*"  | sed 's/.1.cf//'`
+    centrifuge-kreport -x \$db_name ${results} > ${prefix}.txt
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        centrifuge: \$( centrifuge --version  | sed -n 1p | sed 's/^.*centrifuge-class version //')
+    END_VERSIONS
+    """
+}

modules/centrifuge/kreport/meta.yml

@@ -0,0 +1,41 @@
+name: "centrifuge_kreport"
+description: Creates Kraken-style reports from centrifuge out files
+keywords:
+  - metagenomics
+tools:
+  - centrifuge:
+      description: Centrifuge is a classifier for metagenomic sequences.
+      homepage: https://ccb.jhu.edu/software/centrifuge/
+      documentation: https://ccb.jhu.edu/software/centrifuge/manual.shtml
+      doi: 10.1101/gr.210641.116
+      licence: ["GPL v3"]
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - results:
+      type: file
+      description: File containing the centrifuge classification results
+      pattern: "*.{txt}"
+
+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"
+  - kreport:
+      type: file
+      description: |
+        File containing kraken-style report from centrifuge
+        out files.
+      pattern: "*.{txt}"
+authors:
+  - "@sofstam"
+  - "@jfy133"

tests/modules/centrifuge/centrifuge/main.nf

@@ -25,7 +25,6 @@ workflow test_centrifuge_centrifuge_paired_end {
                 file(params.test_data['sarscov2']['illumina']['test_2_fastq_gz'], checkIfExists: true) ]
             ]
      db    =  [ [], file('https://raw.githubusercontent.com/nf-core/test-datasets/modules/data/delete_me/minigut_cf.tar.gz', checkIfExists: true) ]
-     //db_name = "minigut_cf"
      save_unaligned = true
      save_aligned = false
      sam_format = false

tests/modules/centrifuge/kreport/main.nf

@@ -0,0 +1,32 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { UNTAR                } from '../../../../modules/untar/main.nf'
+include { CENTRIFUGE_CENTRIFUGE } from '../../../../modules/centrifuge/centrifuge/main.nf'
+include { CENTRIFUGE_KREPORT    } from '../../../../modules/centrifuge/kreport/main.nf'
+
+workflow test_centrifuge_kreport_single_end {
+
+    input = [ [ id:'test', single_end:true ], // meta map
+              [ file(params.test_data['sarscov2']['illumina']['test_1_fastq_gz'], checkIfExists: true) ]
+            ]
+    db    =  [ [], file('https://raw.githubusercontent.com/nf-core/test-datasets/modules/data/delete_me/minigut_cf.tar.gz', checkIfExists: true) ]
+
+    ch_db = UNTAR ( db )
+    CENTRIFUGE_CENTRIFUGE ( input, ch_db.untar.map{ it[1] }, false, false, false )
+    CENTRIFUGE_KREPORT ( CENTRIFUGE_CENTRIFUGE.out.results, ch_db.untar.map{ it[1] } )
+}
+
+workflow test_centrifuge_kreport_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) ]
+            ]
+     db    =  [ [], file('https://raw.githubusercontent.com/nf-core/test-datasets/modules/data/delete_me/minigut_cf.tar.gz', checkIfExists: true) ]
+
+    ch_db = UNTAR ( db )
+    CENTRIFUGE_CENTRIFUGE ( input, ch_db.untar.map{ it[1] }, false, false, false )
+    CENTRIFUGE_KREPORT ( CENTRIFUGE_CENTRIFUGE.out.results, ch_db.untar.map{ it[1] } )
+}
+

tests/modules/centrifuge/kreport/nextflow.config

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

tests/modules/centrifuge/kreport/test.yml

@@ -0,0 +1,21 @@
+- name: centrifuge kreport test_centrifuge_kreport_single_end
+  command: nextflow run tests/modules/centrifuge/kreport -entry test_centrifuge_kreport_single_end -c tests/config/nextflow.config
+  tags:
+    - centrifuge
+    - centrifuge/kreport
+  files:
+    - path: output/centrifuge/test.txt
+      md5sum: af1a51fe57eb6d428350ff4a4bf759d4
+      contains: ["unclassified"]
+    - path: output/centrifuge/versions.yml
+
+- name: centrifuge kreport test_centrifuge_kreport_paired_end
+  command: nextflow run tests/modules/centrifuge/kreport -entry test_centrifuge_kreport_paired_end -c tests/config/nextflow.config
+  tags:
+    - centrifuge
+    - centrifuge/kreport
+  files:
+    - path: output/centrifuge/test.txt
+      md5sum: af1a51fe57eb6d428350ff4a4bf759d4
+      contains: ["unclassified"]
+    - path: output/centrifuge/versions.yml