Merge pull request #179 from MaxUlysse/master_htslib_tabix

Replace bcftools/tabix by tabix/tabix

.github/filters.yml

@@ -26,10 +26,6 @@ bcftools_stats:
   - software/bcftools/stats/**
   - tests/software/bcftools/stats/**
 
-bcftools_tabix:
-  - software/bcftools/tabix/**
-  - tests/software/bcftools/tabix/**
-
 bedtools_complement:
   - software/bedtools/complement/**
   - tests/software/bedtools/complement/**
@@ -140,6 +136,10 @@ gunzip:
   - software/gunzip/**
   - tests/software/gunzip/**
 
+tabix_tabix:
+  - software/tabix/tabix/**
+  - tests/software/tabix/tabix/**
+
 ivar_consensus:
   - software/ivar/consensus/**
   - tests/software/ivar/consensus/**

deprecated/htslib/tabix/main.nf

@@ -1,16 +0,0 @@
-process htslib_tabix {
-    tag "$vcf"
-
-    container 'quay.io/biocontainers/tabix:0.2.6--ha92aebf_0'
-
-    input:
-        path vcf
-
-    output:
-        path "${vcf}.tbi"
-
-    script:
-    """
-    tabix -p vcf ${vcf}
-    """
-}

deprecated/htslib/tabix/meta.yml

@@ -1,26 +0,0 @@
-name: htslib tabix
-description: create tabix index from a bgzip vcf file
-keywords:
-    - index
-    - tabix
-tools:
-    - bwa:
-        description: |
-            Generic indexer for TAB-delimited genome position files.
-        homepage: https://www.htslib.org/
-        documentation: https://www.htslib.org/doc/tabix.1.html
-        doi: 10.1093/bioinformatics/btq671
-input:
-    -
-        - input:
-            type: file
-            description: Input vcf.gz file
-            pattern: "*.{vcf.gz}"
-output:
-    -
-        - index:
-            type: file
-            description: tabix index file
-            pattern: "*.{vcf.gz.tbi}"
-authors:
-    - "@maxulysse"

deprecated/htslib/tabix/test/main.nf

@@ -1,13 +0,0 @@
-#!/usr/bin/env nextflow
-nextflow.preview.dsl = 2
-include '../../../tests/functions/check_process_outputs.nf' params(params)
-include '../main.nf' params(params)
-
-// Define input channels
-input = '../../../test-datasets/tools/file.vcf.gz'
-
-// Run the workflow
-workflow {
-    tabix_index(ch_read_files)
-    // .check_output()
-}

deprecated/htslib/tabix/test/nextflow.config

@@ -1,2 +0,0 @@
-docker.enabled = true
-params.outdir = './results'

software/bcftools/tabix/meta.yml

@@ -1,63 +0,0 @@
-name: bcftools_tabix
-description: Index GFF/BED/SAM/VCF file
-keywords:
-    - index
-    - vcf
-    - bed
-    - sam
-    - gff
-tools:
-    - stats:
-        description: |
-          Indexes a TAB-delimited genome position file.
-        homepage: http://samtools.github.io/bcftools/bcftools.html
-        documentation: http://www.htslib.org/doc/bcftools.html
-        doi: 10.1093/bioinformatics/btp352
-params:
-    - outdir:
-        type: string
-        description: |
-          The pipeline's output directory. By default, the module will
-          output files into `$params.outdir/<SOFTWARE>`
-    - publish_dir_mode:
-        type: string
-        description: |
-          Value for the Nextflow `publishDir` mode parameter.
-          Available: symlink, rellink, link, copy, copyNoFollow, move.
-    - enable_conda:
-        type: boolean
-        description: |
-          Run the module with Conda using the software specified
-          via the `conda` directive
-    - singularity_pull_docker_container:
-        type: boolean
-        description: |
-          Instead of directly downloading Singularity images for use with Singularity,
-          force the workflow to pull and convert Docker containers instead.
-input:
-    - meta:
-        type: map
-        description: |
-          Groovy Map containing sample information
-          e.g. [ id:'test', single_end:false ]
-    - vcf:
-        type: file
-        description: TAB-delimited genome position file compressed with bgzip
-        pattern: "*.{vcf.gz,bed.gz,sam.gz,gff.gz}"
-output:
-    - meta:
-        type: map
-        description: |
-          Groovy Map containing sample information
-          e.g. [ id:'test', single_end:false ]
-    - tbi:
-        type: file
-        description:  Index file
-        pattern: "*.{tbi}"
-    - version:
-        type: file
-        description: File containing software version
-        pattern: "*.{version.txt}"
-authors:
-    - "@joseespinosa"
-    - "@drpatelh"

software/tabix/tabix/functions.nf

@@ -1,4 +1,3 @@
-
 /*
  * -----------------------------------------------------
  *  Utility functions used in nf-core DSL2 module files
@@ -31,7 +30,7 @@ def initOptions(Map args) {
  */
 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
+    paths = paths.collect { it.trim().replaceAll("^[/]+|[/]+\$", "") } // Trim whitespace and trailing slashes
     return paths.join('/')
 }
 

software/tabix/tabix/main.nf

@@ -4,21 +4,22 @@ include { initOptions; saveFiles; getSoftwareName } from './functions'
 params.options = [:]
 def options    = initOptions(params.options)
 
-process BCFTOOLS_TABIX {
+process TABIX_TABIX {
     tag "$meta.id"
+    label 'process_medium'
     publishDir "${params.outdir}",
         mode: params.publish_dir_mode,
         saveAs: { filename -> saveFiles(filename:filename, options:params.options, publish_dir:getSoftwareName(task.process), publish_id:meta.id) }
 
-    conda (params.enable_conda ? "bioconda::bcftools=1.11=h7c999a4_0" : null)
+    conda (params.enable_conda ? "bioconda::tabix=0.2.6" : null)
     if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
-        container "https://depot.galaxyproject.org/singularity/bcftools:1.11--h7c999a4_0"
+        container "https://depot.galaxyproject.org/singularity/tabix:0.2.6--ha92aebf_0"
     } else {
-        container "quay.io/biocontainers/bcftools:1.11--h7c999a4_0"
+        container "quay.io/biocontainers/tabix:0.2.6--ha92aebf_0"
     }
 
     input:
-    tuple val(meta), path(vcf)
+    tuple val(meta), path(tab)
 
     output:
     tuple val(meta), path("*.tbi"), emit: tbi
@@ -27,7 +28,8 @@ process BCFTOOLS_TABIX {
     script:
     def software = getSoftwareName(task.process)
     """
-    tabix $options.args $vcf
-    echo \$(bcftools --version 2>&1) | sed 's/^.*bcftools //; s/ .*\$//' > ${software}.version.txt
+    tabix $options.args $tab
+
+    echo \$(tabix -h 2>&1) | sed 's/^.*Version: //; s/(.*\$//' > ${software}.version.txt
     """
 }

software/tabix/tabix/meta.yml

@@ -0,0 +1,61 @@
+name: tabix_tabix
+description: create tabix index from a sorted bgzip tab-delimited genome file
+keywords:
+    - index
+    - tabix
+    - vcf
+tools:
+    - tabix:
+        description: Generic indexer for TAB-delimited genome position files.
+        homepage: https://www.htslib.org/doc/tabix.html
+        documentation: https://www.htslib.org/doc/tabix.1.html
+        doi: 10.1093/bioinformatics/btq671
+params:
+    - outdir:
+        type: string
+        description: |
+            The pipeline's output directory. By default, the module will
+            output files into `$params.outdir/<SOFTWARE>`
+    - publish_dir_mode:
+        type: string
+        description: |
+            Value for the Nextflow `publishDir` mode parameter.
+            Available: symlink, rellink, link, copy, copyNoFollow, move.
+    - enable_conda:
+        type: boolean
+        description: |
+            Run the module with Conda using the software specified
+            via the `conda` directive
+    - singularity_pull_docker_container:
+        type: boolean
+        description: |
+            Instead of directly downloading Singularity images for use with Singularity,
+            force the workflow to pull and convert Docker containers instead.
+input:
+    - meta:
+        type: map
+        description: |
+          Groovy Map containing sample information
+          e.g. [ id:'test', single_end:false ]
+    - tab:
+        type: file
+        description: TAB-delimited genome position file compressed with bgzip
+        pattern: "*.{bed.gz,gff.gz,sam.gz,vcf.gz}"
+output:
+    - meta:
+        type: map
+        description: |
+          Groovy Map containing sample information
+          e.g. [ id:'test', single_end:false ]
+    - tbi:
+        type: file
+        description: tabix index file
+        pattern: "*.{tbi}"
+    - version:
+        type: file
+        description: File containing software version
+        pattern: "*.{version.txt}"
+authors:
+    - "@joseespinosa"
+    - "@drpatelh"
+    - "@maxulysse"

tests/software/bcftools/tabix/main.nf

@@ -1,14 +0,0 @@
-#!/usr/bin/env nextflow
-
-nextflow.enable.dsl = 2
-
-include { BCFTOOLS_TABIX } from '../../../../software/bcftools/tabix/main.nf' addParams( options: [:] )
-
-workflow test_bcftools_tabix {
-
-    def input = []
-    input = [ [ id:'test' ], // meta map
-              [ file("${launchDir}/tests/data/vcf/test.vcf.gz", checkIfExists: true) ] ]
-
-    BCFTOOLS_TABIX ( input )
-}

tests/software/bcftools/tabix/test.yml

@@ -1,8 +0,0 @@
-- name: bcftools tabix
-  command: nextflow run ./tests/software/bcftools/tabix -entry test_bcftools_tabix -c tests/config/nextflow.config
-  tags:
-    - bcftools
-    - bcftools_tabix
-  files:
-    - path: output/bcftools/test.vcf.gz.tbi
-      md5sum: 06d52177f819730dd409157914534e8d

tests/software/tabix/tabix/main.nf

@@ -0,0 +1,31 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { TABIX_TABIX as TABIX_BED } from '../../../../software/tabix/tabix/main.nf' addParams( options: ['args': '-p bed'] )
+include { TABIX_TABIX as TABIX_GFF } from '../../../../software/tabix/tabix/main.nf' addParams( options: ['args': '-p gff'] )
+include { TABIX_TABIX as TABIX_VCF } from '../../../../software/tabix/tabix/main.nf' addParams( options: ['args': '-p vcf'] )
+
+workflow test_tabix_tabix_bed {
+    def input = []
+    input = [ [ id:'B.bed' ], // meta map
+              [ file("${launchDir}/tests/data/bed/B.bed.gz", checkIfExists: true) ] ]
+
+    TABIX_BED ( input )
+}
+
+workflow test_tabix_tabix_gff {
+    def input = []
+    input = [ [ id:'a.gff3' ], // meta map
+              [ file("${launchDir}/tests/data/gff/a.gff3.gz", checkIfExists: true) ] ]
+
+    TABIX_GFF ( input )
+}
+
+workflow test_tabix_tabix_vcf {
+    def input = []
+    input = [ [ id:'test.vcf' ], // meta map
+              [ file("${launchDir}/tests/data/vcf/test.vcf.gz", checkIfExists: true) ] ]
+
+    TABIX_VCF ( input )
+}

tests/software/tabix/tabix/test.yml

@@ -0,0 +1,27 @@
+- name: tabix tabix bed
+  command: nextflow run ./tests/software/tabix/tabix -entry test_tabix_tabix_bed -c tests/config/nextflow.config
+  tags:
+    - tabix
+    - tabix_tabix
+    - bed
+  files:
+    - path: output/tabix/B.bed.gz.tbi
+      md5sum: 6d657159c283200109f2ed082296f853
+- name: tabix tabix gff
+  command: nextflow run ./tests/software/tabix/tabix -entry test_tabix_tabix_gff -c tests/config/nextflow.config
+  tags:
+    - tabix
+    - tabix_tabix
+    - gff
+  files:
+    - path: output/tabix/a.gff3.gz.tbi
+      md5sum: 77f666b63da45c6af8a394ab5642602e
+- name: tabix tabix vcf
+  command: nextflow run ./tests/software/tabix/tabix -entry test_tabix_tabix_vcf -c tests/config/nextflow.config
+  tags:
+    - tabix
+    - tabix_tabix
+    - vcf
+  files:
+    - path: output/tabix/test.vcf.gz.tbi
+      md5sum: a03f56d3e968f32256ffb9f6b7d01812