From 4575e5455c62560e0cabdf128ec3adcc81f84a2c Mon Sep 17 00:00:00 2001
From: aleksandrabliznina
 <34429012+aleksandrabliznina@users.noreply.github.com>
Date: Tue, 25 May 2021 04:15:57 +0900
Subject: [PATCH] New last/split module to find split alignments. (#511)

* New last/split module to find split alignments.

The `last-split` tool distributed with [LAST](https://gitlab.com/mcfrith/last)
finds split or spliced alignments in a MAF file that is produced with, for
example, LAST `lastal` command.

This new module is part of the work discribed in Issue #464. During this
development, we fix the versiob of LAST to 1219 to ensure consistency. We will
upgrade it later.

* Update software/last/split/main.nf

* Apply suggestions from code review

Co-authored-by: Harshil Patel <drpatelh@users.noreply.github.com>
---
 software/last/split/functions.nf   | 70 ++++++++++++++++++++++++++++++
 software/last/split/main.nf        | 36 +++++++++++++++
 software/last/split/meta.yml       | 45 +++++++++++++++++++
 tests/config/pytest_software.yml   | 12 +++--
 tests/software/last/split/main.nf  | 13 ++++++
 tests/software/last/split/test.yml |  8 ++++
 6 files changed, 180 insertions(+), 4 deletions(-)
 create mode 100644 software/last/split/functions.nf
 create mode 100644 software/last/split/main.nf
 create mode 100644 software/last/split/meta.yml
 create mode 100644 tests/software/last/split/main.nf
 create mode 100644 tests/software/last/split/test.yml

diff --git a/software/last/split/functions.nf b/software/last/split/functions.nf
new file mode 100644
index 00000000..9d0137e3
--- /dev/null
+++ b/software/last/split/functions.nf
@@ -0,0 +1,70 @@
+/*
+ * -----------------------------------------------------
+ *  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)}/$args.filename"
+        }
+    }
+}
diff --git a/software/last/split/main.nf b/software/last/split/main.nf
new file mode 100644
index 00000000..04b7c8d3
--- /dev/null
+++ b/software/last/split/main.nf
@@ -0,0 +1,36 @@
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName } from './functions'
+
+params.options = [:]
+options        = initOptions(params.options)
+
+process LAST_SPLIT {
+    tag "$meta.id"
+    label 'process_high'
+    publishDir "${params.outdir}",
+        mode: params.publish_dir_mode,
+        saveAs: { filename -> saveFiles(filename:filename, options:params.options, publish_dir:getSoftwareName(task.process), meta:meta, publish_by_meta:['id']) }
+
+    conda (params.enable_conda ? "bioconda::last=1219" : null)
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container "https://depot.galaxyproject.org/singularity/last:1219--h2e03b76_0"
+    } else {
+        container "quay.io/biocontainers/last:1219--h2e03b76_0"
+    }
+
+    input:
+    tuple val(meta), path(maf)
+
+    output:
+    tuple val(meta), path("*.maf.gz"), emit: maf
+    path "*.version.txt"             , emit: version
+
+    script:
+    def software = getSoftwareName(task.process)
+    def prefix   = options.suffix ? "${meta.id}${options.suffix}" : "${meta.id}"
+    """
+    zcat $maf | last-split $options.args | gzip --no-name > ${prefix}.maf.gz
+
+    echo \$(last-split --version 2>&1) | sed 's/last-split //' > ${software}.version.txt
+    """
+}
diff --git a/software/last/split/meta.yml b/software/last/split/meta.yml
new file mode 100644
index 00000000..73f37784
--- /dev/null
+++ b/software/last/split/meta.yml
@@ -0,0 +1,45 @@
+name: last_split
+description: Find split or spliced alignments in a MAF file
+keywords:
+  - LAST
+  - split
+  - spliced
+  - alignment
+  - MAF
+tools:
+  - last:
+      description: LAST finds & aligns related regions of sequences.
+      homepage: https://gitlab.com/mcfrith/last
+      documentation: https://gitlab.com/mcfrith/last/-/blob/main/doc/
+      tool_dev_url: https://gitlab.com/mcfrith/last
+      doi: ""
+      licence: ['GPL v3-or-later']
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - maf:
+      type: file
+      description: Multiple Aligment Format (MAF) file, compressed with gzip
+      pattern: "*.{maf.gz}"
+
+output:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - version:
+      type: file
+      description: File containing software version
+      pattern: "*.{version.txt}"
+  - maf:
+      type: file
+      description: Multiple Aligment Format (MAF) file, compressed with gzip
+      pattern: "*.{maf.gz}"
+
+authors:
+  - "@aleksandrabliznina"
diff --git a/tests/config/pytest_software.yml b/tests/config/pytest_software.yml
index a7a3ab17..b7665654 100644
--- a/tests/config/pytest_software.yml
+++ b/tests/config/pytest_software.yml
@@ -262,6 +262,10 @@ gatk4/haplotypecaller:
   - software/gatk4/haplotypecaller/**
   - tests/software/gatk4/haplotypecaller/**
 
+gatk4/intervallisttools:
+  - software/gatk4/intervallisttools/**
+  - tests/software/gatk4/intervallisttools/**
+
 gatk4/markduplicates:
   - software/gatk4/markduplicates/**
   - tests/software/gatk4/markduplicates/**
@@ -290,10 +294,6 @@ gatk4/variantfiltration:
   - software/gatk4/variantfiltration/**
   - tests/software/gatk4/variantfiltration/**
 
-gatk4/intervallisttools:
-  - software/gatk4/intervallisttools/**
-  - tests/software/gatk4/intervallisttools/**
-
 gffread:
   - software/gffread/**
   - tests/software/gffread/**
@@ -378,6 +378,10 @@ last/mafswap:
   - software/last/mafswap/**
   - tests/software/last/mafswap/**
 
+last/split:
+  - software/last/split/**
+  - tests/software/last/split/**
+
 last/train:
   - software/last/train/**
   - tests/software/last/train/**
diff --git a/tests/software/last/split/main.nf b/tests/software/last/split/main.nf
new file mode 100644
index 00000000..03bfd11c
--- /dev/null
+++ b/tests/software/last/split/main.nf
@@ -0,0 +1,13 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { LAST_SPLIT } from '../../../../software/last/split/main.nf' addParams( options: ['suffix':'.split'] )
+
+workflow test_last_split {
+    
+    input = [ [ id:'contigs.genome' ], // meta map
+              file(params.test_data['sarscov2']['genome']['contigs_genome_maf_gz'], checkIfExists: true) ]
+
+    LAST_SPLIT ( input )
+}
diff --git a/tests/software/last/split/test.yml b/tests/software/last/split/test.yml
new file mode 100644
index 00000000..01ac8348
--- /dev/null
+++ b/tests/software/last/split/test.yml
@@ -0,0 +1,8 @@
+- name: last split test_last_split
+  command: nextflow run tests/software/last/split -entry test_last_split -c tests/config/nextflow.config
+  tags:
+    - last
+    - last/split
+  files:
+    - path: output/last/contigs.genome.split.maf.gz
+      md5sum: 2a177444f63c9294767a67a0247f0f05