From bf09fdf8921797428d462e44d8e495cf2029889d Mon Sep 17 00:00:00 2001
From: nvnieuwk <101190534+nvnieuwk@users.noreply.github.com>
Date: Thu, 8 Sep 2022 15:03:08 +0200
Subject: [PATCH] new module: svtk/standardize (#2031)

* new module: svtk/standardize

* linting

* fixed tests
---
 modules/svtk/standardize/main.nf              | 45 ++++++++++++++
 modules/svtk/standardize/meta.yml             | 61 +++++++++++++++++++
 tests/config/pytest_modules.yml               | 12 ++--
 tests/modules/svtk/standardize/main.nf        | 56 +++++++++++++++++
 .../modules/svtk/standardize/nextflow.config  | 12 ++++
 tests/modules/svtk/standardize/test.yml       | 35 +++++++++++
 6 files changed, 217 insertions(+), 4 deletions(-)
 create mode 100644 modules/svtk/standardize/main.nf
 create mode 100644 modules/svtk/standardize/meta.yml
 create mode 100644 tests/modules/svtk/standardize/main.nf
 create mode 100644 tests/modules/svtk/standardize/nextflow.config
 create mode 100644 tests/modules/svtk/standardize/test.yml

diff --git a/modules/svtk/standardize/main.nf b/modules/svtk/standardize/main.nf
new file mode 100644
index 00000000..104afc20
--- /dev/null
+++ b/modules/svtk/standardize/main.nf
@@ -0,0 +1,45 @@
+process SVTK_STANDARDIZE {
+    tag "$meta.id"
+    label 'process_low'
+
+    conda (params.enable_conda ? "bioconda::svtk=0.0.20190615" : null)
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/svtk:0.0.20190615--py37h73a75cf_2':
+        'quay.io/biocontainers/svtk:0.0.20190615--py37h73a75cf_2' }"
+
+    input:
+    tuple val(meta), path(vcf)
+    path fasta_fai
+
+    output:
+    tuple val(meta), path("*.std.vcf.gz"), emit: standardized_vcf
+    path "versions.yml"                  , emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args = task.ext.args ?: ''
+    def prefix = task.ext.prefix ?: "${meta.id}"
+
+    def arguments   = args.args     ?: ''
+    def caller      = args.caller   ?: 'delly'
+
+    def VERSION = '0.0.20190615' // WARN: Version information not provided by tool on CLI. Please update this string when bumping container versions.
+
+    def contigs = fasta_fai ? "--contigs ${fasta_fai}" : ""
+
+    """
+    svtk standardize \\
+        ${arguments} \\
+        ${contigs} \\
+        ${vcf} \\
+        ${prefix}.std.vcf.gz \\
+        ${caller}
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        svtk: ${VERSION}
+    END_VERSIONS
+    """
+}
diff --git a/modules/svtk/standardize/meta.yml b/modules/svtk/standardize/meta.yml
new file mode 100644
index 00000000..13d551cf
--- /dev/null
+++ b/modules/svtk/standardize/meta.yml
@@ -0,0 +1,61 @@
+name: "svtk_standardize"
+description: Convert SV calls to a standardized format.
+keywords:
+  - svtk
+  - structural variants
+  - SV
+  - vcf
+  - standardization
+tools:
+  - "svtk":
+      description: "Utilities for consolidating, filtering, resolving, and annotating structural variants."
+      homepage: "https://github.com/broadinstitute/gatk-sv/tree/master/src/svtk"
+      documentation: "https://github.com/broadinstitute/gatk-sv/tree/master/src/svtk"
+      tool_dev_url: "https://github.com/broadinstitute/gatk-sv/tree/master/src/svtk"
+      doi: ""
+      licence: "['MIT']"
+
+input:
+  - args:
+      type: map
+      description: |
+        Groovy Map containing tool parameters. MUST follow the structure/keywords below and be provided via modules.config. Parameters must be set between quotes.
+        ```
+        {
+          [
+            "args": "",
+            "caller": "delly" // Should be either delly, lumpy, manta, wham or melt
+          ]
+        }
+        ```
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - vcf:
+      type: file
+      description: A gzipped VCF file to be standardized
+      pattern: "*.vcf.gz"
+  - fasta_fai:
+      type: file
+      description: Optional fasta index file that specifies the contigs to be used in the VCF header (defaults to all contigs of GRCh37)
+      pattern: "*.fai"
+
+output:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - versions:
+      type: file
+      description: File containing software versions
+      pattern: "versions.yml"
+  - standardized_vcf:
+      type: file
+      description: A gzipped version of the standardized VCF file
+      pattern: "*.std.vcf.gz"
+
+authors:
+  - "@nvnieuwk"
diff --git a/tests/config/pytest_modules.yml b/tests/config/pytest_modules.yml
index 81c184ba..6f3c361b 100644
--- a/tests/config/pytest_modules.yml
+++ b/tests/config/pytest_modules.yml
@@ -1538,14 +1538,14 @@ metabat2/metabat2:
   - modules/metabat2/metabat2/**
   - tests/modules/metabat2/metabat2/**
 
-metaphlan3/metaphlan3:
-  - modules/metaphlan3/metaphlan3/**
-  - tests/modules/metaphlan3/metaphlan3/**
-
 metaphlan3/mergemetaphlantables:
   - modules/metaphlan3/mergemetaphlantables/**
   - tests/modules/metaphlan3/mergemetaphlantables/**
 
+metaphlan3/metaphlan3:
+  - modules/metaphlan3/metaphlan3/**
+  - tests/modules/metaphlan3/metaphlan3/**
+
 methyldackel/extract:
   - modules/methyldackel/extract/**
   - tests/modules/methyldackel/extract/**
@@ -2263,6 +2263,10 @@ svdb/query:
   - modules/svdb/query/**
   - tests/modules/svdb/query/**
 
+svtk/standardize:
+  - modules/svtk/standardize/**
+  - tests/modules/svtk/standardize/**
+
 tabix/bgzip:
   - modules/tabix/bgzip/**
   - tests/modules/tabix/bgzip/**
diff --git a/tests/modules/svtk/standardize/main.nf b/tests/modules/svtk/standardize/main.nf
new file mode 100644
index 00000000..6601edbd
--- /dev/null
+++ b/tests/modules/svtk/standardize/main.nf
@@ -0,0 +1,56 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { SVTK_STANDARDIZE } from '../../../../modules/svtk/standardize/main.nf'
+include { MANTA_GERMLINE   } from '../../../../modules/manta/germline/main.nf'
+
+workflow test_svtk_standardize {
+    
+    input = [
+        [ id:'test', single_end:false ], // meta map
+        file(params.test_data['homo_sapiens']['illumina']['test_paired_end_sorted_cram'], checkIfExists: true),
+        file(params.test_data['homo_sapiens']['illumina']['test_paired_end_sorted_cram_crai'], checkIfExists: true),
+        file(params.test_data['homo_sapiens']['genome']['genome_bed_gz'], checkIfExists: true),
+        file(params.test_data['homo_sapiens']['genome']['genome_bed_gz_tbi'], checkIfExists: true)
+    ]
+
+    fasta = file(params.test_data['homo_sapiens']['genome']['genome_fasta'], checkIfExists: true)
+    fasta_fai = file(params.test_data['homo_sapiens']['genome']['genome_fasta_fai'], checkIfExists: true)
+
+    MANTA_GERMLINE(
+        input,
+        fasta,
+        fasta_fai
+    )
+
+    SVTK_STANDARDIZE ( 
+        MANTA_GERMLINE.out.diploid_sv_vcf,
+        fasta_fai
+    )
+}
+
+workflow test_svtk_standardize_no_contigs {
+    
+    input = [
+        [ id:'test', single_end:false ], // meta map
+        file(params.test_data['homo_sapiens']['illumina']['test_paired_end_sorted_cram'], checkIfExists: true),
+        file(params.test_data['homo_sapiens']['illumina']['test_paired_end_sorted_cram_crai'], checkIfExists: true),
+        file(params.test_data['homo_sapiens']['genome']['genome_bed_gz'], checkIfExists: true),
+        file(params.test_data['homo_sapiens']['genome']['genome_bed_gz_tbi'], checkIfExists: true)
+    ]
+
+    fasta = file(params.test_data['homo_sapiens']['genome']['genome_fasta'], checkIfExists: true)
+    fasta_fai = file(params.test_data['homo_sapiens']['genome']['genome_fasta_fai'], checkIfExists: true)
+
+    MANTA_GERMLINE(
+        input,
+        fasta,
+        fasta_fai
+    )
+
+    SVTK_STANDARDIZE ( 
+        MANTA_GERMLINE.out.diploid_sv_vcf,
+        []
+    )
+}
diff --git a/tests/modules/svtk/standardize/nextflow.config b/tests/modules/svtk/standardize/nextflow.config
new file mode 100644
index 00000000..f432a5e6
--- /dev/null
+++ b/tests/modules/svtk/standardize/nextflow.config
@@ -0,0 +1,12 @@
+process {
+
+    publishDir = { "${params.outdir}/${task.process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()}" }
+    
+    withName: SVTK_STANDARDIZE {
+        ext.args = [
+            args   : '',
+            caller : 'manta'
+        ]
+    }
+
+}
\ No newline at end of file
diff --git a/tests/modules/svtk/standardize/test.yml b/tests/modules/svtk/standardize/test.yml
new file mode 100644
index 00000000..764cfe2f
--- /dev/null
+++ b/tests/modules/svtk/standardize/test.yml
@@ -0,0 +1,35 @@
+- name: svtk standardize test_svtk_standardize
+  command: nextflow run ./tests/modules/svtk/standardize -entry test_svtk_standardize -c ./tests/config/nextflow.config -c ./tests/modules/svtk/standardize/nextflow.config
+  tags:
+    - svtk
+    - svtk/standardize
+  files:
+    - path: output/manta/test.candidate_small_indels.vcf.gz
+    - path: output/manta/test.candidate_small_indels.vcf.gz.tbi
+      md5sum: 4cb176febbc8c26d717a6c6e67b9c905
+    - path: output/manta/test.candidate_sv.vcf.gz
+    - path: output/manta/test.candidate_sv.vcf.gz.tbi
+      md5sum: 4cb176febbc8c26d717a6c6e67b9c905
+    - path: output/manta/test.diploid_sv.vcf.gz
+    - path: output/manta/test.diploid_sv.vcf.gz.tbi
+      md5sum: 4cb176febbc8c26d717a6c6e67b9c905
+    - path: output/svtk/test.std.vcf.gz
+      md5sum: f7530f3bc7e6020e758cc996adc8ae35
+
+- name: svtk standardize test_svtk_standardize_no_contigs
+  command: nextflow run ./tests/modules/svtk/standardize -entry test_svtk_standardize_no_contigs -c ./tests/config/nextflow.config -c ./tests/modules/svtk/standardize/nextflow.config
+  tags:
+    - svtk
+    - svtk/standardize
+  files:
+    - path: output/manta/test.candidate_small_indels.vcf.gz
+    - path: output/manta/test.candidate_small_indels.vcf.gz.tbi
+      md5sum: 4cb176febbc8c26d717a6c6e67b9c905
+    - path: output/manta/test.candidate_sv.vcf.gz
+    - path: output/manta/test.candidate_sv.vcf.gz.tbi
+      md5sum: 4cb176febbc8c26d717a6c6e67b9c905
+    - path: output/manta/test.diploid_sv.vcf.gz
+    - path: output/manta/test.diploid_sv.vcf.gz.tbi
+      md5sum: 4cb176febbc8c26d717a6c6e67b9c905
+    - path: output/svtk/test.std.vcf.gz
+      md5sum: ea67be7438a0fb72db8096e1bc1bb957