New Snpsites module (#416)

* new gubbins module

* new gubbins module

* new gubbins module

* Update software/gubbins/main.nf

Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com>

* Update software/gubbins/main.nf

Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com>

* Update software/gubbins/main.nf

Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com>

* Update software/gubbins/meta.yml

Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com>

* new gubbins module

* new gubbins module

* new gubbins module

* new gubbins module

* new gubbins module

* new gubbins module

* new gubbins module

* Update tests/config/test_data.config

Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com>

* Update tests/software/gubbins/main.nf

Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com>

* new gubbins module

* new gubbins module

* new snpsites module

* new snpsites module

* new snpsites module

* new snpsites module

* adding fasttree module

* correct trailing whitespace

* using sarscov2 as a test dir

* use sars-cov-2 alignment

* remove old test alignment

* new snpsites module

* new snpsites module

* new snpsites module

* updated test file names

* new snpsites module

* revert to gubbins test on upstream

* add new lines

* renove whitespace

* Apply suggestions from code review

Co-authored-by: avantonder <avt@sanger.ac.uk>
Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com>
Co-authored-by: avantonder <ajv37@cam.ac.uk>

software/snpsites/functions.nf

@@ -0,0 +1,60 @@
+/*
+ * -----------------------------------------------------
+ *  Utility functions used in nf-core DSL2 module files
+ * -----------------------------------------------------
+ */
+
+/*
+ * Extract name of software tool from process name using $task.process
+ */
+def getSoftwareName(task_process) {
+    return task_process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()
+}
+
+/*
+ * Function to initialise default values and to generate a Groovy Map of available options for nf-core modules
+ */
+def initOptions(Map args) {
+    def Map options = [:]
+    options.args          = args.args ?: ''
+    options.args2         = args.args2 ?: ''
+    options.args3         = args.args3 ?: ''
+    options.publish_by_id = args.publish_by_id ?: false
+    options.publish_dir   = args.publish_dir ?: ''
+    options.publish_files = args.publish_files
+    options.suffix        = args.suffix ?: ''
+    return options
+}
+
+/*
+ * Tidy up and join elements of a list to return a path string
+ */
+def getPathFromList(path_list) {
+    def paths = path_list.findAll { item -> !item?.trim().isEmpty() }  // Remove empty entries
+    paths = paths.collect { it.trim().replaceAll("^[/]+|[/]+\$", "") } // Trim whitespace and trailing slashes
+    return paths.join('/')
+}
+
+/*
+ * Function to save/publish module results
+ */
+def saveFiles(Map args) {
+    if (!args.filename.endsWith('.version.txt')) {
+        def ioptions = initOptions(args.options)
+        def path_list = [ ioptions.publish_dir ?: args.publish_dir ]
+        if (ioptions.publish_by_id) {
+            path_list.add(args.publish_id)
+        }
+        if (ioptions.publish_files instanceof Map) {
+            for (ext in ioptions.publish_files) {
+                if (args.filename.endsWith(ext.key)) {
+                    def ext_list = path_list.collect()
+                    ext_list.add(ext.value)
+                    return "${getPathFromList(ext_list)}/$args.filename"
+                }
+            }
+        } else if (ioptions.publish_files == null) {
+            return "${getPathFromList(path_list)}/$args.filename"
+        }
+    }
+}

software/snpsites/main.nf

@@ -0,0 +1,39 @@
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName } from './functions'
+
+params.options = [:]
+options        = initOptions(params.options)
+
+process SNPSITES {
+    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:'') }
+
+    conda (params.enable_conda ? "bioconda::snp-sites=2.5.1" : null)
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container "https://depot.galaxyproject.org/singularity/snp-sites:2.5.1--hed695b0_0"
+    } else {
+        container "quay.io/biocontainers/snp-sites:2.5.1--hed695b0_0"
+    }
+
+    input:
+    path alignment
+
+    output:
+    path "*.fas"        , emit: fasta
+    path "*.sites.txt"  , emit: constant_sites
+    path "*.version.txt", emit: version
+
+    script:
+    def software = getSoftwareName(task.process)
+    """
+    snp-sites -c \\
+        $alignment \\
+        > filtered_alignment.fas
+
+    echo \$(snp-sites -C $alignment) > constant.sites.txt
+
+    echo \$(snp-sites -V 2>&1) | sed 's/snp-sites //' > ${software}.version.txt
+    """
+}

software/snpsites/meta.yml

@@ -0,0 +1,31 @@
+name: snpsites
+description: Rapidly extracts SNPs from a multi-FASTA alignment.
+keywords:
+  - SNPs
+  - invariant
+  - constant
+tools:
+  - snpsites:
+      description: Rapidly extracts SNPs from a multi-FASTA alignment.
+      homepage: https://www.sanger.ac.uk/tool/snp-sites/
+      documentation: https://github.com/sanger-pathogens/snp-sites
+input:
+  - alignment:
+      type: file
+      description: fasta alignment file
+      pattern: "*.{fasta,fas,fa,aln}"
+output:
+  - version:
+      type: file
+      description: File containing software version
+      pattern: "*.{version.txt}"
+  - fasta:
+      type: file
+      description: Variant fasta file
+      pattern: "*.{fas}"
+  - constant_sites:
+      type: file
+      description: Text file containing counts of constant sites
+      pattern: "*.{sites.txt}"
+authors:
+  - "@avantonder"

tests/config/pytest_software.yml

@@ -458,6 +458,10 @@ shovill:
   - software/shovill/**
   - tests/software/shovill/**
 
+snpsites:
+  - software/snpsites/**
+  - tests/software/snpsites/**
+
 spades:
   - software/spades/**
   - tests/software/spades/**

tests/software/snpsites/main.nf

@@ -0,0 +1,12 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { SNPSITES } from '../../../software/snpsites/main.nf' addParams( options: [:] )
+
+workflow test_snpsites {
+    
+    input = file(params.test_data['sarscov2']['genome']['all_sites_fas'], checkIfExists: true)
+
+    SNPSITES ( input )
+}

tests/software/snpsites/test.yml

@@ -0,0 +1,9 @@
+- name: snpsites
+  command: nextflow run ./tests/software/snpsites -entry test_snpsites -c tests/config/nextflow.config
+  tags:
+    - snpsites
+  files:
+    - path: output/snpsites/filtered_alignment.fas
+      md5sum: 34a3d2ae4439447eabe80282f4625dba
+    - path: output/snpsites/constant.sites.txt
+      md5sum: 8b9b226e3787f7baaefce07405af22c9