From a2bcb5c36a555d18d13c4d560d277c5ef91b878c Mon Sep 17 00:00:00 2001
From: Gregor Sturm <mail@gregor-sturm.de>
Date: Wed, 15 Jul 2020 13:00:10 +0200
Subject: [PATCH] Update module according to discussion on Slack

---
 software/fastqc/main.nf      | 43 ++++++++++++++++++++++--------------
 software/fastqc/meta.yml     | 18 +++++++++++++--
 software/fastqc/test/main.nf |  3 ++-
 3 files changed, 45 insertions(+), 19 deletions(-)

diff --git a/software/fastqc/main.nf b/software/fastqc/main.nf
index 01c21a62..35753657 100644
--- a/software/fastqc/main.nf
+++ b/software/fastqc/main.nf
@@ -1,29 +1,40 @@
-nextflow.preview.dsl = 2
 def MODULE = "fastqc"
 params.publish_dir = MODULE
 params.publish_results = "default"
 
 process FASTQC {
-    input:
-        tuple val(name), path(reads)
-
-    output:
-        tuple val(name), path ("*fastqc*"), emit: all
-        path "*.zip", emit: report // e.g. for MultiQC later
-        path "*.version.txt", emit: version
-
-    container "docker.pkg.github.com/nf-core/$MODULE"
-    conda "${moduleDir}/environment.yml"
-
-    publishDir "${params.out_dir}/${params.publish_dir}/$name",
+    publishDir "${params.out_dir}/${params.publish_dir}",
         mode: params.publish_dir_mode,
         saveAs: { filename ->
-                        if(params.publish_results == "none") null
-                        else filename }
+                    if (params.publish_results == "none") null
+                    else filename }
+
+    container "docker.pkg.github.com/nf-core/$MODULE"
+
+    conda "${moduleDir}/environment.yml"
+
+    input:
+    tuple val(name), val(single_end), path(reads)
+
+    output:
+    tuple val(name), val(single_end), path("*.html"), emit: html
+    tuple val(name), val(single_end), path("*.zip"), emit: zip
+    path "*.version.txt", emit: version
 
     script:
+    // Add soft-links to original FastQs for consistent naming in pipeline
+    if (single_end) {
         """
-        fastqc ${params.fastqc_args} -t ${task.cpus} $reads
+        [ ! -f  ${name}.fastq.gz ] && ln -s $reads ${name}.fastq.gz
+        fastqc ${params.fastqc_args} --threads $task.cpus ${name}.fastq.gz
         fastqc --version | sed -n "s/.*\\(v.*\$\\)/\\1/p" > fastqc.version.txt
         """
+    } else {
+        """
+        [ ! -f  ${name}_1.fastq.gz ] && ln -s ${reads[0]} ${name}_1.fastq.gz
+        [ ! -f  ${name}_2.fastq.gz ] && ln -s ${reads[1]} ${name}_2.fastq.gz
+        fastqc ${params.fastqc_args} --threads $task.cpus ${name}_1.fastq.gz ${name}_2.fastq.gz
+        fastqc --version | sed -n "s/.*\\(v.*\$\\)/\\1/p" > fastqc.version.txt
+        """
+    }
 }
diff --git a/software/fastqc/meta.yml b/software/fastqc/meta.yml
index 94f3e320..fcbfeab3 100644
--- a/software/fastqc/meta.yml
+++ b/software/fastqc/meta.yml
@@ -29,14 +29,28 @@ params:
         Append to the path for the standard output directory provided by `$out_dir`.
   - publish_dir_mode:
       type: string
-      description: Provide a value for the Nextflow `publishDir` mode parameter (e.g. copy, link, ...)
+      description: |
+        Provide a value for the Nextflow `publishDir` mode parameter
+        (e.g. copy, link, ...)
+  - publish_results:
+      type: string
+      description: |
+        Whether or not to publish results into `publish_dir`. Set to `none` to not
+        publish any files at all; to `default` to publish all relevant files.
 input:
   - name:
       type: string
       description: Sample identifier
+  - single_end:
+      type: boolean
+      description: |
+        Boolean indicating whether the corresponding sample is single-end (true)
+        or paired-end (false).
   - reads:
       type: file
-      description: List of input FastQ files of size 1 and 2 for single-end and paired-end data, respectively.
+      description: |
+        List of input FastQ files of size 1 and 2 for single-end and paired-end data,
+        respectively.
 output:
   - report:
       type: file
diff --git a/software/fastqc/test/main.nf b/software/fastqc/test/main.nf
index f9eb4a75..0b0038d7 100755
--- a/software/fastqc/test/main.nf
+++ b/software/fastqc/test/main.nf
@@ -12,7 +12,7 @@ include { FASTQC } from '../main.nf'
  */
 workflow test_single_end {
     input_files = Channel.fromPath("data/test_single_end.fastq.gz")
-                    .map {f -> [f.baseName, f]}
+                    .map {f -> [f.baseName, true, f]}
     FASTQC(input_files)
 }
 
@@ -21,6 +21,7 @@ workflow test_single_end {
  */
 workflow test_paired_end {
     input_files = Channel.fromFilePairs("data/test_R{1,2}.fastq.gz")
+                    .map {f -> [f[0], false, f[1]]}
     FASTQC(input_files)
 }