Merge pull request #61 from drpatelh/master

Add docs and tests for BWA index

.github/workflows/bwa_index.yml

@@ -0,0 +1,30 @@
+name: bwa_index
+on:
+  push:
+    paths:
+      - software/bwa/index/**
+      - .github/workflows/bwa_index.yml
+      - tests
+  pull_request:
+    paths:
+      - software/bwa/index/**
+      - .github/workflows/bwa_index.yml
+      - tests
+
+jobs:
+  ci_test:
+    runs-on: ubuntu-latest
+    env:
+      NXF_ANSI_LOG: false
+    steps:
+
+    - uses: actions/checkout@v2
+
+    - name: Install Nextflow
+      run: |
+        export NXF_VER="20.07.1"
+        wget -qO- get.nextflow.io | bash
+        sudo mv nextflow /usr/local/bin/
+
+    # Test the module
+    - run: nextflow run ./software/bwa/index/test/ -profile docker

software/bwa/index/meta.yml

@@ -1,25 +1,52 @@
-name: bwa index
+name: bwa_index
-description: create indexes for BWA from a fasta file
+description: Create BWA index for reference genome
 keywords:
   - index
+  - fasta
+  - genome
 tools:
   - bwa:
       description: |
-            BWA is a software package for mapping DNA sequences against a large reference genome, such as the human genome.
+        BWA is a software package for mapping DNA sequences against
+        a large reference genome, such as the human genome.
       homepage: http://bio-bwa.sourceforge.net/
       documentation: http://www.htslib.org/doc/samtools.html
       arxiv: arXiv:1303.3997
+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.
+  - conda:
+      type: boolean
+      description: |
+        Run the module with Conda using the software specified
+        via the `conda` directive
 input:
-    -
+  - fasta:
-        - input:
       type: file
-            description: Input fasta file
+      description: |
-            pattern: "*.{fasta,fa}"
+        Input genome fasta file
+  - options:
+      type: map
+      description: |
+        Groovy Map containing module options for passing command-line arguments and
+        output file paths.
 output:
-    -
   - index:
       type: file
-            description: bwa indexes file
+      description: BWA genome index files
-            pattern: "*.{fasta,fa}.{amb,ann,bwt,pac,sa}"
+      pattern: "*.{fasta}.{amb,ann,bwt,pac,sa}"
+  - version:
+      type: file
+      description: File containing software version
+      pattern: "*.version.txt"
 authors:
+  - "@drpatelh"
   - "@maxulysse"

software/bwa/index/test/input/NC_010473.fa

@@ -0,0 +1 @@
+../../../../../tests/data/fasta/E_coli/NC_010473.fa

software/bwa/index/test/main.nf

@@ -1,16 +1,13 @@
 #!/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
+nextflow.enable.dsl = 2
-input = '../../../test-datasets/tools/bwa/index/input/reference.fasta'
-Channel
-  .from(input)
-  .set { ch_input }
 
-// Run the workflow
+include { BWA_INDEX } from '../main.nf'
-workflow {
+
-    fastqc(ch_input)
+workflow test {
-    // .check_output()
+    BWA_INDEX ( file("${baseDir}/input/NC_010473.fa", checkIfExists: true), [:] )
+}
+
+workflow {
+    test()
 }

software/bwa/index/test/nextflow.config

@@ -1,2 +1,20 @@
-docker.enabled = true
+
-params.outdir = './results'
+params {
+  outdir = "output/"
+  publish_dir_mode = "copy"
+  conda = false
+}
+
+profiles {
+  conda  {
+    params.conda = true
+  }
+  docker {
+    docker.enabled = true
+    docker.runOptions = '-u \$(id -u):\$(id -g)'
+  }
+  singularity {
+    singularity.enabled = true
+    singularity.autoMounts = true
+  }
+}

software/bwa/index/test/output/bwa/NC_010473.fa.amb

@@ -0,0 +1,3 @@
+4686137 1 2
+20895 1 Y
+142347 1 R

software/bwa/index/test/output/bwa/NC_010473.fa.ann

@@ -0,0 +1,3 @@
+4686137 1 11
+0 gi|170079663|ref|NC_010473.1| Escherichia coli str. K-12 substr. DH10B, complete genome
+0 4686137 2

software/fastqc/meta.yml

@@ -1,9 +1,9 @@
 name: fastqc
 description: Run FastQC on sequenced reads
 keywords:
-  - Quality Control
+  - quality control
-  - QC
+  - qc
-  - Adapters
+  - adapters
 tools:
   - fastqc:
       description: |