From fe3e18eaeee1b5784bc1197cb01e9e7204de5da9 Mon Sep 17 00:00:00 2001
From: "James A. Fellows Yates" <jfy133@gmail.com>
Date: Tue, 6 Sep 2022 19:32:30 +0200
Subject: [PATCH] Add krakentools/combinekreports (#2018)

* Add krakentools_combinekreports

* Update test.yml

* Update main.nf

* Update tests/modules/krakentools/combinekreports/test.yml

* Prettier

* Update modules/krakentools/combinekreports/main.nf

Co-authored-by: Robert A. Petit III <robbie.petit@gmail.com>

Co-authored-by: Robert A. Petit III <robbie.petit@gmail.com>
---
 modules/krakentools/combinekreports/main.nf   | 34 +++++++++++++++++++
 modules/krakentools/combinekreports/meta.yml  | 33 ++++++++++++++++++
 tests/config/pytest_modules.yml               |  4 +++
 .../krakentools/combinekreports/main.nf       | 23 +++++++++++++
 .../combinekreports/nextflow.config           |  5 +++
 .../krakentools/combinekreports/test.yml      |  9 +++++
 6 files changed, 108 insertions(+)
 create mode 100644 modules/krakentools/combinekreports/main.nf
 create mode 100644 modules/krakentools/combinekreports/meta.yml
 create mode 100644 tests/modules/krakentools/combinekreports/main.nf
 create mode 100644 tests/modules/krakentools/combinekreports/nextflow.config
 create mode 100644 tests/modules/krakentools/combinekreports/test.yml

diff --git a/modules/krakentools/combinekreports/main.nf b/modules/krakentools/combinekreports/main.nf
new file mode 100644
index 00000000..c50b7dc0
--- /dev/null
+++ b/modules/krakentools/combinekreports/main.nf
@@ -0,0 +1,34 @@
+process KRAKENTOOLS_COMBINEKREPORTS {
+    label 'process_low'
+
+    conda (params.enable_conda ? "bioconda::krakentools=1.2" : null)
+    container "${ workflow.containerEngine == 'singularity' && !task.ext.singularity_pull_docker_container ?
+        'https://depot.galaxyproject.org/singularity/krakentools:1.2--pyh5e36f6f_0':
+        'quay.io/biocontainers/krakentools:1.2--pyh5e36f6f_0' }"
+
+    input:
+    path kreports
+
+    output:
+    path "*.txt", emit: txt
+    path "versions.yml", emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args = task.ext.args ?: ''
+    def prefix = task.ext.prefix ?: "kreports_combined.txt"
+    def VERSION = '1.2' // WARN: Version information not provided by tool on CLI. Please update this string when bumping container versions.
+    """
+    combine_kreports.py \\
+        -r ${kreports} \\
+        -o ${prefix} \\
+        ${args}
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        combine_kreports.py: ${VERSION}
+    END_VERSIONS
+    """
+}
diff --git a/modules/krakentools/combinekreports/meta.yml b/modules/krakentools/combinekreports/meta.yml
new file mode 100644
index 00000000..275897cb
--- /dev/null
+++ b/modules/krakentools/combinekreports/meta.yml
@@ -0,0 +1,33 @@
+name: krakentools_combinekreports
+description: Takes a Kraken report file and prints out a krona-compatible TEXT file
+keywords:
+  - kraken
+  - krakentools
+  - metagenomics
+  - table
+  - combining
+  - merging
+tools:
+  - krakentools:
+      description: KrakenTools is a suite of scripts to be used for post-analysis of Kraken/KrakenUniq/Kraken2/Bracken results. Please cite the relevant paper if using KrakenTools with any of the listed programs.
+      homepage: https://github.com/jenniferlu717/KrakenTools
+      licence: ["GPL v3"]
+
+input:
+  - kreports:
+      type: file
+      description: List of kraken-style report files
+      pattern: "*.{txt,kreport}"
+
+output:
+  - versions:
+      type: file
+      description: File containing software versions
+      pattern: "versions.yml"
+  - txt:
+      type: file
+      description: Combined kreport file of all input files
+      pattern: "*.txt"
+
+authors:
+  - "@jfy133"
diff --git a/tests/config/pytest_modules.yml b/tests/config/pytest_modules.yml
index 7640b018..2c9f3c77 100644
--- a/tests/config/pytest_modules.yml
+++ b/tests/config/pytest_modules.yml
@@ -1302,6 +1302,10 @@ kraken2/kraken2:
   - modules/untar/**
   - tests/modules/kraken2/kraken2/**
 
+krakentools/combinekreports:
+  - modules/krakentools/combinekreports/**
+  - tests/modules/krakentools/combinekreports/**
+
 krakentools/kreport2krona:
   - modules/krakentools/kreport2krona/**
   - tests/modules/krakentools/kreport2krona/**
diff --git a/tests/modules/krakentools/combinekreports/main.nf b/tests/modules/krakentools/combinekreports/main.nf
new file mode 100644
index 00000000..a35fe913
--- /dev/null
+++ b/tests/modules/krakentools/combinekreports/main.nf
@@ -0,0 +1,23 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { UNTAR } from '../../../../modules/untar/main'
+include { KRAKEN2_KRAKEN2 } from '../../../../modules/kraken2/kraken2/main'
+include { KRAKENTOOLS_COMBINEKREPORTS } from '../../../../modules/krakentools/combinekreports/main.nf'
+
+workflow test_krakentools_combinekreports {
+
+    input = Channel.of(
+        [[ id:'test', single_end:false ], [ file(params.test_data['sarscov2']['illumina']['test_1_fastq_gz'], checkIfExists: true), file(params.test_data['sarscov2']['illumina']['test_2_fastq_gz'], checkIfExists: true) ]],
+        [[ id:'test2', single_end:false ], [ file(params.test_data['sarscov2']['illumina']['test_1_fastq_gz'], checkIfExists: true), file(params.test_data['sarscov2']['illumina']['test_2_fastq_gz'], checkIfExists: true) ]],
+    )
+
+
+    db    = [ [], file(params.test_data['sarscov2']['genome']['kraken2_tar_gz'], checkIfExists: true) ]
+
+    UNTAR ( db )
+    KRAKEN2_KRAKEN2 ( input, UNTAR.out.untar.map{ it[1] }, false, false )
+
+    KRAKENTOOLS_COMBINEKREPORTS ( KRAKEN2_KRAKEN2.out.report.map{ [ it[1] ] }.collect().dump() )
+}
diff --git a/tests/modules/krakentools/combinekreports/nextflow.config b/tests/modules/krakentools/combinekreports/nextflow.config
new file mode 100644
index 00000000..50f50a7a
--- /dev/null
+++ b/tests/modules/krakentools/combinekreports/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/krakentools/combinekreports/test.yml b/tests/modules/krakentools/combinekreports/test.yml
new file mode 100644
index 00000000..ff4cdc69
--- /dev/null
+++ b/tests/modules/krakentools/combinekreports/test.yml
@@ -0,0 +1,9 @@
+- name: krakentools combinekreports test_krakentools_combinekreports
+  command: nextflow run ./tests/modules/krakentools/combinekreports -entry test_krakentools_combinekreports -c ./tests/config/nextflow.config  -c ./tests/modules/krakentools/combinekreports/nextflow.config
+  tags:
+    - krakentools/combinekreports
+    - krakentools
+  files:
+    - path: output/krakentools/kreports_combined.txt
+      contains:
+        - "#Number of Samples: 2"