Merge pull request #7 from CCBR/khmer

feat: khmer/uniquekmers module w/ stubs

CHANGELOG.md

@@ -1,6 +1,7 @@
 ## development version
 
 - new modules:
-  - cutadapt
   - bwa/index
   - bwa/mem
+  - cutadapt
+  - khmer/uniquekmers

README.md

@@ -37,6 +37,6 @@ Want to **contribute** to this project? Check out the [contributing guidelines](
 Many of the modules and subworkflows in this project reuse and adapt code from [nf-core/modules](https://github.com/nf-core/modules).
 In those cases, credit is noted in the `meta.yml` file of the module/subworkflow and also listed here:
 
+- [bwa](modules/CCBR/bwa) adapts the [nf-core bwa module](https://github.com/nf-core/chipseq/tree/51eba00b32885c4d0bec60db3cb0a45eb61e34c5/modules/nf-core/modules/bwa)
 - [cutadapt](modules/CCBR/cutadapt) adapts the [nf-core cutadapt module](https://github.com/nf-core/modules/tree/master/modules/nf-core/cutadapt)
-- [bwa_mem](modules/CCBR/bwa/mem) adapts the [nf-core bwa mem module](https://github.com/nf-core/chipseq/tree/51eba00b32885c4d0bec60db3cb0a45eb61e34c5/modules/nf-core/modules/bwa/mem)
-- [bwa_index](modules/CCBR/bwa/index) adapts the [nf-core bwa index module](https://github.com/nf-core/chipseq/tree/51eba00b32885c4d0bec60db3cb0a45eb61e34c5/modules/nf-core/modules/bwa/index)
+- [khmer](modules/CCBR/khmer) adapts the [nf-core khmer module](https://github.com/nf-core/modules/tree/b48a1efc8e067502e1a9bafbac788c1e0abdfc6a/modules/nf-core/khmer)

modules/CCBR/khmer/uniquekmers/main.nf

@@ -0,0 +1,45 @@
+process KHMER_UNIQUEKMERS {
+    tag "$fasta"
+    label 'process_low'
+
+    container 'nciccbr/ccbr_khmer_3.0.0:v1'
+
+    input:
+    path fasta
+    val  kmer_size
+
+    output:
+    path "report.txt"  , emit: report
+    path "kmers.txt"   , emit: kmers
+    path "versions.yml", emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args = task.ext.args ?: ''
+    """
+    unique-kmers.py \\
+        -k $kmer_size \\
+        -R report.txt \\
+        $args \\
+        $fasta
+
+    grep ^number report.txt | sed 's/^.*:.[[:blank:]]//g' > kmers.txt
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        khmer: \$( unique-kmers.py --version 2>&1 | grep ^khmer | sed 's/^khmer //;s/ .*\$//' )
+    END_VERSIONS
+    """
+
+    stub:
+    """
+    touch report.txt kmers.txt
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        khmer: \$( unique-kmers.py --version 2>&1 | grep ^khmer | sed 's/^khmer //;s/ .*\$//' )
+    END_VERSIONS
+    """
+}

modules/CCBR/khmer/uniquekmers/meta.yml

@@ -0,0 +1,42 @@
+name: "khmer_uniquekmers"
+description: In-memory nucleotide sequence k-mer counting, filtering, graph traversal and more. Adapted from nf-core/modules khmer/uniquekhmers.
+keywords:
+  - khmer
+  - k-mer
+  - effective genome size
+
+tools:
+  - "khmer":
+      description: khmer k-mer counting library
+      homepage: https://github.com/dib-lab/khmer
+      documentation: https://khmer.readthedocs.io/en/latest/
+      tool_dev_url: https://github.com/dib-lab/khmer
+      doi: "10.12688/f1000research.6924.1"
+      licence: ["BSD License"]
+
+input:
+  - fasta:
+      type: file
+      description: fasta file
+      pattern: "*.{fa,fasta}"
+  - kmer_size:
+      type: integer
+      description: k-mer size to use
+      pattern: "[0-9]+"
+
+output:
+  - report:
+      type: file
+      description: Text file containing unique-kmers.py execution report
+      pattern: "report.txt"
+  - kmers:
+      type: file
+      description: Text file containing number of kmers
+      pattern: "kmers.txt"
+  - versions:
+      type: file
+      description: File containing software versions
+      pattern: "versions.yml"
+
+authors:
+  - "@kelly-sovacool"

tests/config/pytest_modules.yml

@@ -9,3 +9,7 @@ bwa/mem:
 cutadapt:
   - modules/CCBR/cutadapt/**
   - tests/modules/CCBR/cutadapt/**
+
+khmer/uniquekmers:
+  - modules/CCBR/khmer/uniquekmers/**
+  - tests/CCBR/khmer/uniquekmers/**

tests/modules/CCBR/khmer/uniquekmers/main.nf

@@ -0,0 +1,12 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { KHMER_UNIQUEKMERS } from '../../../../../modules/CCBR/khmer/uniquekmers/main.nf'
+
+workflow test_khmer_uniquekmers {
+
+    input = file(params.test_data['sarscov2']['genome']['genome_fasta'], checkIfExists: true)
+
+    KHMER_UNIQUEKMERS ( input, 50 )
+}

tests/modules/CCBR/khmer/uniquekmers/nextflow.config

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

tests/modules/CCBR/khmer/uniquekmers/test.yml

@@ -0,0 +1,11 @@
+- name: khmer uniquekmers test_khmer_uniquekmers
+  command: nextflow run ./tests/modules/CCBR/khmer/uniquekmers -entry test_khmer_uniquekmers -c ./tests/config/nextflow.config
+  tags:
+    - khmer
+    - khmer/uniquekmers
+  files:
+    - path: output/khmer/kmers.txt
+      md5sum: 496ebf23653a01c7a42d743e47c19f65
+    - path: output/khmer/report.txt
+      md5sum: ee489abd3b244dea3640649e1790d55e
+    - path: output/khmer/versions.yml