From 23f9284aaecf77c8f3ddd3f447109a06bd588917 Mon Sep 17 00:00:00 2001
From: Michael L Heuer <heuermh@acm.org>
Date: Thu, 21 Jan 2021 16:34:49 -0600
Subject: [PATCH] Added dsh-bio split-bed as dsh_splitbed

---
 .github/workflows/dsh_splitbed.yml | 40 ++++++++++++++++++++
 software/dsh/splitbed/functions.nf | 59 ++++++++++++++++++++++++++++++
 software/dsh/splitbed/main.nf      | 42 +++++++++++++++++++++
 software/dsh/splitbed/meta.yml     | 59 ++++++++++++++++++++++++++++++
 tests/software/dsh/main.nf         | 10 +++++
 tests/software/dsh/test.yml        | 11 ++++++
 6 files changed, 221 insertions(+)
 create mode 100644 .github/workflows/dsh_splitbed.yml
 create mode 100644 software/dsh/splitbed/functions.nf
 create mode 100644 software/dsh/splitbed/main.nf
 create mode 100644 software/dsh/splitbed/meta.yml

diff --git a/.github/workflows/dsh_splitbed.yml b/.github/workflows/dsh_splitbed.yml
new file mode 100644
index 00000000..f2ebbd88
--- /dev/null
+++ b/.github/workflows/dsh_splitbed.yml
@@ -0,0 +1,40 @@
+name: dsh_splitbed
+on:
+  push:
+    paths:
+      - software/dsh/splitbed/**
+      - .github/workflows/dsh_splitbed.yml
+      - tests/software/dsh/**
+  pull_request:
+    paths:
+      - software/dsh/splitbed/**
+      - .github/workflows/dsh_splitbed.yml
+      - tests/software/dsh/**
+
+jobs:
+  ci_test:
+    runs-on: ubuntu-latest
+    strategy:
+      matrix:
+        nxf_version: [20.11.0-edge]
+    env:
+      NXF_ANSI_LOG: false
+    steps:
+      - uses: actions/checkout@v2
+
+      - name: Install Nextflow
+        env:
+          NXF_VER: ${{ matrix.nxf_version }}
+        run: |
+          wget -qO- get.nextflow.io | bash
+          sudo mv nextflow /usr/local/bin/
+
+      - name: Set up Python
+        uses: actions/setup-python@v2
+        with:
+          python-version: "3.x"
+      - name: Install dependencies
+        run: python -m pip install --upgrade pip pytest-workflow
+
+      # Test the module
+      - run: pytest --tag dsh_splitbed --symlink --wt 2
diff --git a/software/dsh/splitbed/functions.nf b/software/dsh/splitbed/functions.nf
new file mode 100644
index 00000000..d25eea86
--- /dev/null
+++ b/software/dsh/splitbed/functions.nf
@@ -0,0 +1,59 @@
+/*
+ * -----------------------------------------------------
+ *  Utility functions used in nf-core DSL2 module files
+ * -----------------------------------------------------
+ */
+
+/*
+ * Extract name of software tool from process name using $task.process
+ */
+def getSoftwareName(task_process) {
+    return task_process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()
+}
+
+/*
+ * Function to initialise default values and to generate a Groovy Map of available options for nf-core modules
+ */
+def initOptions(Map args) {
+    def Map options = [:]
+    options.args          = args.args ?: ''
+    options.args2         = args.args2 ?: ''
+    options.publish_by_id = args.publish_by_id ?: false
+    options.publish_dir   = args.publish_dir ?: ''
+    options.publish_files = args.publish_files
+    options.suffix        = args.suffix ?: ''
+    return options
+}
+
+/*
+ * Tidy up and join elements of a list to return a path string
+ */
+def getPathFromList(path_list) {
+    def paths = path_list.findAll { item -> !item?.trim().isEmpty() }  // Remove empty entries
+    paths = paths.collect { it.trim().replaceAll("^[/]+|[/]+\$", "") } // Trim whitespace and trailing slashes
+    return paths.join('/')
+}
+
+/*
+ * Function to save/publish module results
+ */
+def saveFiles(Map args) {
+    if (!args.filename.endsWith('.version.txt')) {
+        def ioptions = initOptions(args.options)
+        def path_list = [ ioptions.publish_dir ?: args.publish_dir ]
+        if (ioptions.publish_by_id) {
+            path_list.add(args.publish_id)
+        }
+        if (ioptions.publish_files instanceof Map) {
+            for (ext in ioptions.publish_files) {
+                if (args.filename.endsWith(ext.key)) {
+                    def ext_list = path_list.collect()
+                    ext_list.add(ext.value)
+                    return "${getPathFromList(ext_list)}/$args.filename"
+                }
+            }
+        } else if (ioptions.publish_files == null) {
+            return "${getPathFromList(path_list)}/$args.filename"
+        }
+    }
+}
diff --git a/software/dsh/splitbed/main.nf b/software/dsh/splitbed/main.nf
new file mode 100644
index 00000000..713b4215
--- /dev/null
+++ b/software/dsh/splitbed/main.nf
@@ -0,0 +1,42 @@
+// Import generic module functions
+include { initOptions; saveFiles; getSoftwareName } from './functions'
+
+params.options = [:]
+def options    = initOptions(params.options)
+
+process DSH_SPLITBED {
+    tag "${meta.id}"
+    label 'process_medium'
+    publishDir "${params.outdir}",
+        mode: params.publish_dir_mode,
+        saveAs: { filename -> saveFiles(filename:filename, options:params.options, publish_dir:getSoftwareName(task.process), publish_id:meta.id) }
+
+    conda (params.enable_conda ? "bioconda::dsh-bio=1.4" : null)
+
+    if (workflow.containerEngine == 'singularity' && !params.singularity_pull_docker_container) {
+        container "https://depot.galaxyproject.org/singularity/dsh-bio:1.4--0"
+    } else {
+        container "quay.io/biocontainers/dsh-bio:1.4--0"
+    }
+
+    input:
+    tuple val(meta), path(features)
+
+    output:
+    tuple val(meta), path("*.bed.gz"), emit: features
+    path "*.version.txt"             , emit: version
+
+    script:
+    def software = getSoftwareName(task.process)
+    def prefix   = options.suffix ? "${meta.id}${options.suffix}" : "${meta.id}"
+
+    """
+    dsh-split-bed \\
+        $options.args \\
+        -p $prefix \\
+        -s '.bed.gz' \\
+        -i $features
+
+    echo \$(dsh-bio --version 2>&1) | grep -o 'dsh-bio-tools .*' | cut -f2 -d ' ' > ${software}.version.txt
+    """
+}
diff --git a/software/dsh/splitbed/meta.yml b/software/dsh/splitbed/meta.yml
new file mode 100644
index 00000000..91444648
--- /dev/null
+++ b/software/dsh/splitbed/meta.yml
@@ -0,0 +1,59 @@
+name: dsh_splitbed
+description: Split features in gzipped BED format
+keywords:
+    - bed
+tools:
+    - dsh:
+        description: |
+            Reads, features, variants, assemblies, alignments, genomic range trees, pangenome
+            graphs, and a bunch of random command line tools for bioinformatics. LGPL version 3
+            or later.
+        homepage: https://github.com/heuermh/dishevelled-bio
+        documentation: https://github.com/heuermh/dishevelled-bio
+params:
+    - outdir:
+        type: string
+        description: |
+            The pipeline's output directory. By default, the module will
+            output files into `$params.outdir/<SOFTWARE>`
+    - publish_dir_mode:
+        type: string
+        description: |
+            Value for the Nextflow `publishDir` mode parameter.
+            Available: symlink, rellink, link, copy, copyNoFollow, move.
+    - enable_conda:
+        type: boolean
+        description: |
+            Run the module with Conda using the software specified
+            via the `conda` directive
+    - singularity_pull_docker_container:
+        type: boolean
+        description: |
+            Instead of directly downloading Singularity images for use with Singularity,
+            force the workflow to pull and convert Docker containers instead.
+input:
+    - meta:
+        type: map
+        description: |
+            Groovy Map containing sample information
+            e.g. [ id:'test', single_end:false ]
+    - features:
+        type: file
+        description: Features in gzipped BED format to split
+        pattern: "*.{bed.gz}"
+output:
+    - meta:
+        type: map
+        description: |
+            Groovy Map containing sample information
+            e.g. [ id:'test', single_end:false ]
+    - features:
+        type: file
+        description: Features in split gzipped BED formatted files
+        pattern: "*.{bed.gz}"
+    - version:
+        type: file
+        description: File containing software version
+        pattern: "*.{version.txt}"
+authors:
+    - "@heuermh"
diff --git a/tests/software/dsh/main.nf b/tests/software/dsh/main.nf
index e22ae054..e007e3e1 100644
--- a/tests/software/dsh/main.nf
+++ b/tests/software/dsh/main.nf
@@ -3,6 +3,7 @@
 nextflow.enable.dsl = 2
 
 include { DSH_FILTERBED } from '../../../software/dsh/filterbed/main.nf' addParams( options: [suffix: '.filtered', args: '--range chr1:0-1000'] )
+include { DSH_SPLITBED } from '../../../software/dsh/splitbed/main.nf' addParams( options: [suffix: '.', args: '--records 2'] )
 
 workflow test_dsh_filterbed {
 
@@ -12,3 +13,12 @@ workflow test_dsh_filterbed {
 
     DSH_FILTERBED ( input )
 }
+
+workflow test_dsh_splitbed {
+
+    def input = []
+    input = [ [ id:'A' ], // meta map
+              [ file("${launchDir}/tests/data/bed/A.bed.gz", checkIfExists: true) ] ]
+
+    DSH_SPLITBED ( input )
+}
diff --git a/tests/software/dsh/test.yml b/tests/software/dsh/test.yml
index e6dde73b..3e4261b7 100644
--- a/tests/software/dsh/test.yml
+++ b/tests/software/dsh/test.yml
@@ -6,3 +6,14 @@
   files:
     - path: output/dsh/A.filtered.bed.gz
       md5sum: 91cee4392d6034793f99f6c53891d866
+
+- name: Run dsh splitbed
+  command: nextflow run ./tests/software/dsh -profile docker -entry test_dsh_splitbed -c tests/config/nextflow.config
+  tags:
+    - dsh
+    - dsh_splitbed
+  files:
+    - path: output/dsh/A.0.bed.gz
+      md5sum: 619e51c435d0e7138ca41b660ed07e3a
+    - path: output/dsh/A.1.bed.gz
+      md5sum: 223bba025a0bb859de9a1bf385550a7e