From 2926ec0abb192e29888a0b040ec776e99826891f Mon Sep 17 00:00:00 2001
From: Mahesh Binzer-Panchal <mahesh.binzer-panchal@nbis.se>
Date: Fri, 20 May 2022 13:23:11 +0000
Subject: [PATCH] Add mash screen

---
 modules/mash/screen/main.nf               | 38 ++++++++++++++++++
 modules/mash/screen/meta.yml              | 47 +++++++++++++++++++++++
 tests/config/pytest_modules.yml           |  4 ++
 tests/modules/mash/screen/main.nf         | 21 ++++++++++
 tests/modules/mash/screen/nextflow.config |  5 +++
 tests/modules/mash/screen/test.yml        | 12 ++++++
 6 files changed, 127 insertions(+)
 create mode 100644 modules/mash/screen/main.nf
 create mode 100644 modules/mash/screen/meta.yml
 create mode 100644 tests/modules/mash/screen/main.nf
 create mode 100644 tests/modules/mash/screen/nextflow.config
 create mode 100644 tests/modules/mash/screen/test.yml

diff --git a/modules/mash/screen/main.nf b/modules/mash/screen/main.nf
new file mode 100644
index 00000000..3e7859ed
--- /dev/null
+++ b/modules/mash/screen/main.nf
@@ -0,0 +1,38 @@
+process MASH_SCREEN {
+    tag "$meta.id"
+    label 'process_medium'
+
+    conda (params.enable_conda ? "bioconda::mash=2.3" : null)
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/mash:2.3--he348c14_1':
+        'quay.io/biocontainers/mash:2.3--he348c14_1' }"
+
+    input:
+    tuple val(meta), path(query_sketch)
+    path fastx_db
+
+    output:
+    tuple val(meta), path("*.screen"), emit: screen
+    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}"
+    """
+    mash \\
+        screen \\
+        $args \\
+        -p $task.cpus \\
+        $query_sketch \\
+        $fastx_db \\
+        > ${prefix}.screen
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        mash: \$( mash --version )
+    END_VERSIONS
+    """
+}
diff --git a/modules/mash/screen/meta.yml b/modules/mash/screen/meta.yml
new file mode 100644
index 00000000..81e455bc
--- /dev/null
+++ b/modules/mash/screen/meta.yml
@@ -0,0 +1,47 @@
+name: "mash_screen"
+description: Screens query sequences against large sequence databases
+keywords:
+  - screen
+  - containment
+  - contamination
+  - taxonomic assignment
+tools:
+  - "mash":
+      description: Fast sequence distance estimator that uses MinHash
+      homepage: https://github.com/marbl/Mash
+      documentation: https://mash.readthedocs.io/en/latest/sketches.html
+      tool_dev_url: https://github.com/marbl/Mash
+      doi: "10.1186/s13059-016-0997-x"
+      licence: ["https://github.com/marbl/Mash/blob/master/LICENSE.txt"]
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - query_sketch:
+      type: file
+      description: MinHash sketch of query sequences
+      pattern: "*.msh"
+  - fastx_db:
+      type: file
+      description: Sequence files to match against
+
+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"
+  - screen:
+      type: file
+      description: List of sequences from fastx_db similar to query sequences
+      pattern: "*.screen"
+
+authors:
+  - "@mahesh-panchal"
diff --git a/tests/config/pytest_modules.yml b/tests/config/pytest_modules.yml
index 16411798..c1aeb0c6 100644
--- a/tests/config/pytest_modules.yml
+++ b/tests/config/pytest_modules.yml
@@ -1214,6 +1214,10 @@ mash/dist:
   - modules/mash/dist/**
   - tests/modules/mash/dist/**
 
+mash/screen:
+  - modules/mash/screen/**
+  - tests/modules/mash/screen/**
+
 mash/sketch:
   - modules/mash/sketch/**
   - tests/modules/mash/sketch/**
diff --git a/tests/modules/mash/screen/main.nf b/tests/modules/mash/screen/main.nf
new file mode 100644
index 00000000..7f0b8bd9
--- /dev/null
+++ b/tests/modules/mash/screen/main.nf
@@ -0,0 +1,21 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { MASH_SKETCH } from '../../../../modules/mash/sketch/main.nf'
+include { MASH_SCREEN } from '../../../../modules/mash/screen/main.nf'
+
+workflow test_mash_screen {
+
+    input = [
+        [ id:'test', single_end:false], // meta map
+        [
+            file(params.test_data['sarscov2']['illumina']['test_1_fastq_gz'], checkIfExists: true),
+            file(params.test_data['sarscov2']['illumina']['test_2_fastq_gz'], checkIfExists: true)
+        ]
+    ]
+    fastx_db = file(params.test_data['sarscov2']['genome']['genome_fasta'], checkIfExists: true)
+
+    MASH_SKETCH ( input )
+    MASH_SCREEN ( MASH_SKETCH.out.mash, fastx_db )
+}
diff --git a/tests/modules/mash/screen/nextflow.config b/tests/modules/mash/screen/nextflow.config
new file mode 100644
index 00000000..50f50a7a
--- /dev/null
+++ b/tests/modules/mash/screen/nextflow.config
@@ -0,0 +1,5 @@
+process {
+
+    publishDir = { "${params.outdir}/${task.process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()}" }
+    
+}
\ No newline at end of file
diff --git a/tests/modules/mash/screen/test.yml b/tests/modules/mash/screen/test.yml
new file mode 100644
index 00000000..f622817a
--- /dev/null
+++ b/tests/modules/mash/screen/test.yml
@@ -0,0 +1,12 @@
+- name: mash screen test_mash_screen
+  command: nextflow run ./tests/modules/mash/screen -entry test_mash_screen -c ./tests/config/nextflow.config  -c ./tests/modules/mash/screen/nextflow.config
+  tags:
+    - mash
+    - mash/screen
+  files:
+    - path: output/mash/test.mash_stats
+      md5sum: 2a6f297d8e69a5e4160243bc6c89129c
+    - path: output/mash/test.msh
+      md5sum: d747145a43dad5f82342036f8f5d9133
+    - path: output/mash/test.screen
+      md5sum: d3c871dccd5cd57ab54781fa5c5d7278