From fd5aea0b76c1e5b908191f212382a55a57cee316 Mon Sep 17 00:00:00 2001
From: Mahesh Binzer-Panchal <mahesh.binzer-panchal@nbis.se>
Date: Tue, 14 Jun 2022 16:30:37 +0200
Subject: [PATCH] Add FastK/Histex (#1776)

---
 modules/fastk/histex/main.nf               | 35 ++++++++++++++++++
 modules/fastk/histex/meta.yml              | 42 ++++++++++++++++++++++
 tests/config/pytest_modules.yml            |  4 +++
 tests/modules/fastk/histex/main.nf         | 17 +++++++++
 tests/modules/fastk/histex/nextflow.config |  8 +++++
 tests/modules/fastk/histex/test.yml        |  8 +++++
 6 files changed, 114 insertions(+)
 create mode 100644 modules/fastk/histex/main.nf
 create mode 100644 modules/fastk/histex/meta.yml
 create mode 100644 tests/modules/fastk/histex/main.nf
 create mode 100644 tests/modules/fastk/histex/nextflow.config
 create mode 100644 tests/modules/fastk/histex/test.yml

diff --git a/modules/fastk/histex/main.nf b/modules/fastk/histex/main.nf
new file mode 100644
index 00000000..b1acd070
--- /dev/null
+++ b/modules/fastk/histex/main.nf
@@ -0,0 +1,35 @@
+process FASTK_HISTEX {
+    tag "$meta.id"
+    label 'process_low'
+
+    if (params.enable_conda) {
+        error "Conda environments cannot be used when using the FastK tool. Please use docker or singularity containers."
+    }
+    container 'ghcr.io/nbisweden/fastk_genescopefk_merquryfk:1.0'
+
+    input:
+    tuple val(meta), path(histogram)
+
+    output:
+    tuple val(meta), path("*.hist"), emit: hist
+    path "versions.yml"            , emit: versions
+
+    when:
+    task.ext.when == null || task.ext.when
+
+    script:
+    def args = task.ext.args ?: ''
+    def prefix = task.ext.prefix ?: "${meta.id}"
+    def FASTK_VERSION = 'f18a4e6d2207539f7b84461daebc54530a9559b0'
+    """
+    Histex \\
+        $args \\
+        $histogram \\
+        > ${prefix}.hist
+
+    cat <<-END_VERSIONS > versions.yml
+    "${task.process}":
+        fastk: $FASTK_VERSION
+    END_VERSIONS
+    """
+}
diff --git a/modules/fastk/histex/meta.yml b/modules/fastk/histex/meta.yml
new file mode 100644
index 00000000..fab381d3
--- /dev/null
+++ b/modules/fastk/histex/meta.yml
@@ -0,0 +1,42 @@
+name: "fastk_histex"
+description: A fast K-mer counter for high-fidelity shotgun datasets
+keywords:
+  - k-mer
+  - histogram
+tools:
+  - "fastk":
+      description: "A fast K-mer counter for high-fidelity shotgun datasets"
+      homepage: "https://github.com/thegenemyers/FASTK"
+      documentation: ""
+      tool_dev_url: "https://github.com/thegenemyers/FASTK"
+      doi: ""
+      licence: "https://github.com/thegenemyers/FASTK/blob/master/LICENSE"
+
+input:
+  - meta:
+      type: map
+      description: |
+        Groovy Map containing sample information
+        e.g. [ id:'test', single_end:false ]
+  - histogram:
+      type: file
+      description: A FastK histogram file
+      pattern: "*.hist"
+
+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"
+  - hist:
+      type: file
+      description: A formatted histogram file
+      pattern: "*.hist"
+
+authors:
+  - "@mahesh-panchal"
diff --git a/tests/config/pytest_modules.yml b/tests/config/pytest_modules.yml
index 0e82d01a..f98e8ec9 100644
--- a/tests/config/pytest_modules.yml
+++ b/tests/config/pytest_modules.yml
@@ -679,6 +679,10 @@ fastk/fastk:
   - modules/fastk/fastk/**
   - tests/modules/fastk/fastk/**
 
+fastk/histex:
+  - modules/fastk/histex/**
+  - tests/modules/fastk/histex/**
+
 fastp:
   - modules/fastp/**
   - tests/modules/fastp/**
diff --git a/tests/modules/fastk/histex/main.nf b/tests/modules/fastk/histex/main.nf
new file mode 100644
index 00000000..74b46378
--- /dev/null
+++ b/tests/modules/fastk/histex/main.nf
@@ -0,0 +1,17 @@
+#!/usr/bin/env nextflow
+
+nextflow.enable.dsl = 2
+
+include { FASTK_FASTK } from '../../../../modules/fastk/fastk/main.nf'
+include { FASTK_HISTEX } from '../../../../modules/fastk/histex/main.nf'
+
+workflow test_fastk_histex {
+
+    input = [
+        [ id:'test' , single_end: true ], // meta map
+        file(params.test_data['sarscov2']['illumina']['test_1_fastq_gz'], checkIfExists: true)
+    ]
+
+    FASTK_FASTK ( input )
+    FASTK_HISTEX ( FASTK_FASTK.out.hist )
+}
diff --git a/tests/modules/fastk/histex/nextflow.config b/tests/modules/fastk/histex/nextflow.config
new file mode 100644
index 00000000..038a0897
--- /dev/null
+++ b/tests/modules/fastk/histex/nextflow.config
@@ -0,0 +1,8 @@
+process {
+
+    publishDir = { "${params.outdir}/${task.process.tokenize(':')[-1].tokenize('_')[0].toLowerCase()}" }
+
+    withName: 'FASTK_FASTK' {
+        publishDir = [ enabled: false ]
+    }
+}
diff --git a/tests/modules/fastk/histex/test.yml b/tests/modules/fastk/histex/test.yml
new file mode 100644
index 00000000..8348af8f
--- /dev/null
+++ b/tests/modules/fastk/histex/test.yml
@@ -0,0 +1,8 @@
+- name: fastk histex test_fastk_histex
+  command: nextflow run ./tests/modules/fastk/histex -entry test_fastk_histex -c ./tests/config/nextflow.config  -c ./tests/modules/fastk/histex/nextflow.config
+  tags:
+    - fastk/histex
+    - fastk
+  files:
+    - path: output/fastk/test.hist
+      md5sum: e3f0c9e7641a46525123312008e4ae86