module fastani (#695)

* initiate agrvate module

* remove todos [ci skip]

* initiate fastani draft [ci skip]

* clean stubs [ci skip]

* interim commit [ci skip]

* accomodate the batch/per-sample processing [ci skip]

* use the meta map [ci skip]

* run first test [ci skip]

* remove extra spaces [ci skip]

* change output file name [ci skip]

* update the expected output [ci skip]

* update the files used for test [ci skip]

* fix typo [ci skip]

* test passing [ci skip]

* update the description

* remove extra files

* accomodate CR suggestions [ci skip]

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

* accomodate CR suggestions [ci skip]

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

* accomodate CR suggestions [ci skip]

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

* use meta map for batch analysis

* fix the tests

* rely upon tuples

* Apply suggestions from code review

* Update main.nf

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

modules/fastani/functions.nf

@@ -0,0 +1,68 @@
+//
+//  Utility functions used in nf-core DSL2 module files
+//
+
+//
+// Extract name of software tool from process name using $task.process
+//
+def getSoftwareName(task_process) {
+    return task_process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()
+}
+
+//
+// Function to initialise default values and to generate a Groovy Map of available options for nf-core modules
+//
+def initOptions(Map args) {
+    def Map options = [:]
+    options.args            = args.args ?: ''
+    options.args2           = args.args2 ?: ''
+    options.args3           = args.args3 ?: ''
+    options.publish_by_meta = args.publish_by_meta ?: []
+    options.publish_dir     = args.publish_dir ?: ''
+    options.publish_files   = args.publish_files
+    options.suffix          = args.suffix ?: ''
+    return options
+}
+
+//
+// Tidy up and join elements of a list to return a path string
+//
+def getPathFromList(path_list) {
+    def paths = path_list.findAll { item -> !item?.trim().isEmpty() }      // Remove empty entries
+    paths     = paths.collect { it.trim().replaceAll("^[/]+|[/]+\$", "") } // Trim whitespace and trailing slashes
+    return paths.join('/')
+}
+
+//
+// Function to save/publish module results
+//
+def saveFiles(Map args) {
+    if (!args.filename.endsWith('.version.txt')) {
+        def ioptions  = initOptions(args.options)
+        def path_list = [ ioptions.publish_dir ?: args.publish_dir ]
+        if (ioptions.publish_by_meta) {
+            def key_list = ioptions.publish_by_meta instanceof List ? ioptions.publish_by_meta : args.publish_by_meta
+            for (key in key_list) {
+                if (args.meta && key instanceof String) {
+                    def path = key
+                    if (args.meta.containsKey(key)) {
+                        path = args.meta[key] instanceof Boolean ? "${key}_${args.meta[key]}".toString() : args.meta[key]
+                    }
+                    path = path instanceof String ? path : ''
+                    path_list.add(path)
+                }
+            }
+        }
+        if (ioptions.publish_files instanceof Map) {
+            for (ext in ioptions.publish_files) {
+                if (args.filename.endsWith(ext.key)) {
+                    def ext_list = path_list.collect()
+                    ext_list.add(ext.value)
+                    return "${getPathFromList(ext_list)}/$args.filename"
+                }
+            }
+        } else if (ioptions.publish_files == null) {
+            return "${getPathFromList(path_list)}/$args.filename"
+        }
+    }
+}

modules/fastani/main.nf

@@ -0,0 +1,52 @@
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName } from './functions'
+
+params.options = [:]
+options        = initOptions(params.options)
+
+process FASTANI {
+    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), meta:meta, publish_by_meta:['id']) }
+
+    conda (params.enable_conda ? "bioconda::fastani=1.32" : null)
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container "https://depot.galaxyproject.org/singularity/fastani:1.32--he1c1bb9_0"
+    } else {
+        container "quay.io/biocontainers/fastani:1.32--he1c1bb9_0"
+    }
+
+    input:
+    tuple val(meta), path(query)
+    path reference
+
+    output:
+    tuple val(meta), path("*.ani.txt"), emit: ani
+    path "*.version.txt"              , emit: version
+
+    script:
+    def software = getSoftwareName(task.process)
+    def prefix   = options.suffix ? "${meta.id}${options.suffix}" : "${meta.id}"
+
+    if (meta.batch_input) {
+        """
+        fastANI \\
+            -ql $query \\
+            -rl $reference \\
+            -o ${prefix}.ani.txt
+
+        echo \$(fastANI --version 2>&1) | sed 's/version//;' > ${software}.version.txt
+        """
+    } else {
+        """
+        fastANI \\
+            -q $query \\
+            -r $reference \\
+            -o ${prefix}.ani.txt
+
+        echo \$(fastANI --version 2>&1) | sed 's/version//;' > ${software}.version.txt
+        """
+    }
+}

modules/fastani/meta.yml

@@ -0,0 +1,43 @@
+name: fastani
+description: write your description here
+keywords:
+  - fastani
+tools:
+  - fastani:
+      description: FastANI is developed for fast alignment-free computation of whole-genome Average Nucleotide Identity (ANI).
+      homepage: https://github.com/ParBLiSS/FastANI
+      documentation: https://github.com/ParBLiSS/FastANI
+      tool_dev_url: https://github.com/ParBLiSS/FastANI
+      doi: 10.1038/s41467-018-07641-9
+      licence: ['Apache-2.0']
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - query:
+      type: file
+      description: Fasta file(s) to be queried
+      pattern: "*.fasta"
+  - reference:
+      type: file
+      description: Fasta file(s) to be used as reference for the query
+      pattern: "*.fasta"
+output:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - ani:
+      type: file
+      description: Results of the query
+      pattern: "*.ani.txt"
+  - version:
+      type: file
+      description: File containing software version
+      pattern: "*.{version.txt}"
+authors:
+  - "@abhi18av"

tests/config/pytest_modules.yml

@@ -318,6 +318,10 @@ expansionhunter:
   - modules/expansionhunter/**
   - tests/modules/expansionhunter/**
 
+fastani:
+  - modules/fastani/**
+  - tests/modules/fastani/**
+
 fastp:
   - modules/fastp/**
   - tests/modules/fastp/**

tests/modules/fastani/main.nf

@@ -0,0 +1,13 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { FASTANI } from '../../../modules/fastani/main.nf' addParams( options: [:] )
+
+workflow test_fastani {
+
+    query = file(params.test_data['sarscov2']['genome']['genome_fasta'], checkIfExists: true)
+    reference = file(params.test_data['sarscov2']['illumina']['contigs_fasta'], checkIfExists: true)
+
+    FASTANI ( [[ id:'test' ], query], reference )
+}

tests/modules/fastani/test.yml

@@ -0,0 +1,7 @@
+- name: fastani
+  command: nextflow run ./tests/modules/fastani -entry test_fastani -c tests/config/nextflow.config
+  tags:
+    - fastani
+  files:
+    - path: output/fastani/test.ani.txt
+      md5sum: 31d4f04e8cffe13080c86db3f9f3a589