Merge pull request #615 from MaxUlysse/installNextflowDoc

Small refactoring

Author: Szilveszter Juhos

CHANGELOG.md

@@ -5,6 +5,19 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/)
 and this project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.html).
 
 ## [Unreleased]
+### `Added`
+- [#613](https://github.com/SciLifeLab/Sarek/pull/613) - Add Issue Templates (bug report and feature request)
+- [#614](https://github.com/SciLifeLab/Sarek/pull/614) - Add PR Template
+- [#615](https://github.com/SciLifeLab/Sarek/pull/615) - Add presentation
+- [#615](https://github.com/SciLifeLab/Sarek/pull/615) - Update documentation
+
+### `Changed`
+- [#608](https://github.com/SciLifeLab/Sarek/pull/608) - Update Nextflow required version
+- [#616](https://github.com/SciLifeLab/Sarek/pull/616) - Update CHANGELOG
+- [#615](https://github.com/SciLifeLab/Sarek/pull/615) - Use `splitCsv` instead of `readlines`
+
+### `Removed`
+- [#616](https://github.com/SciLifeLab/Sarek/pull/616) - Remove old Issue Template
 
 ## [2.1.0] - Ruotes - 2018-08-14
 
@@ -25,8 +38,6 @@ and this project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.
 - [#601](https://github.com/SciLifeLab/Sarek/pull/601), [#603](https://github.com/SciLifeLab/Sarek/pull/603) - Container for GATK4
 - [#606](https://github.com/SciLifeLab/Sarek/pull/606) - Add test data as a submodule from [`Sarek-data`](https://github.com/SciLifeLab/Sarek-data)
 - [#608](https://github.com/SciLifeLab/Sarek/pull/608) - Add documentation on how to install Nextflow on `bianca`
-- [#613](https://github.com/SciLifeLab/Sarek/pull/613) - Add Issue Templates (bug report and feature request)
-- [#614](https://github.com/SciLifeLab/Sarek/pull/614) - Add PR Template
 
 ### `Changed`
 - [#557](https://github.com/SciLifeLab/Sarek/pull/557), [#583](https://github.com/SciLifeLab/Sarek/pull/583), [#585](https://github.com/SciLifeLab/Sarek/pull/585), [#588](https://github.com/SciLifeLab/Sarek/pull/588) - Update help
@@ -58,7 +69,6 @@ and this project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.
 
 ### `Removed`
 - [#607](https://github.com/SciLifeLab/Sarek/pull/607) - Remove Mutect1
-- [#616](https://github.com/SciLifeLab/Sarek/pull/616) - Remove old Issue Template
 
 ## [2.0.0] - 2018-03-23
 ### `Added`

doc/INSTALL_BIANCA.md

@@ -50,10 +50,20 @@ For more information about using Singularity with UPPMAX, follow the [Singularit
 > mv .nextflow nextflow_v[xx.yy.zz]
 > mv bin nextflow_v[xx.yy.zz]/bin
 
+# Establish permission for some files
+> chmod 755 nextflow_v[xx.yy.zz]/bin/nextflow
+> chmod 660 nextflow_v[xx.yy.zz]/framework/[xx.yy.zz]/nextflow-[xx.yy.zz]-one.jar
+
+# If you want other people to use it
+# Be sure that your group has rights to the directory as well
+
+> chown -R .[PROJECT] nextflow_v[xx.yy.zz]
+
 # Clean directory
 > rm nextflow_v[xx.yy.zz].tgz
 
 # And everytime you're launching Nextflow, don't forget to export the following ENV variables
+# Or add them to your .bashrc file
 > export NXF_HOME=/castor/project/proj/nobackup/tools/nextflow/nextflow_v[xx.yy.zz]
 > export PATH=${NXF_HOME}/bin:${PATH}
 > export NXF_TEMP=$SNIC_TMP
@@ -137,6 +147,10 @@ Wrote Sarek-[snapID].tar.gz
 # extract Sarek
 > tar xvzf Sarek-[snapID].tgz
 
+# If you want other people to use it
+# Be sure that your group has rights to the directory as well
+> chown -R .[PROJECT] Sarek-[snapID]
+
 # Make a symbolic link to the extracted repository
 > ln -s Sarek-[snapID] default
 ```

doc/Presentations/2018-07-04-MGarcia-JOBIM.pdf

@@ -8,6 +8,12 @@ class SarekUtils {
     if (!it.toString().toLowerCase().endsWith(extension.toLowerCase())) exit 1, "File: ${it} has the wrong extension: ${extension} see --help for more information"
   }
 
+  // Check if a row has the expected number of item
+  static def checkNumberOfItem(row, number) {
+    if (row.size() != number) exit 1, "Malformed row in TSV file: ${row}, see --help for more information"
+    return true
+  }
+
   // Check parameter existence
   static def checkParameterExistence(it, list) {
     if (!list.contains(it)) {
@@ -143,16 +149,16 @@ class SarekUtils {
   // Channeling the TSV file containing BAM.
   // Format is: "subject gender status sample bam bai"
   static def extractBams(tsvFile, mode) {
-    Channel
-      .from(tsvFile.readLines())
-      .map{line ->
-        def list      = SarekUtils.returnTSV(line.split(),6)
-        def idPatient = list[0]
-        def gender    = list[1]
-        def status    = SarekUtils.returnStatus(list[2].toInteger())
-        def idSample  = list[3]
-        def bamFile   = SarekUtils.returnFile(list[4])
-        def baiFile   = SarekUtils.returnFile(list[5])
+    Channel.from(tsvFile)
+      .splitCsv(sep: '\t')
+      .map { row ->
+        SarekUtils.checkNumberOfItem(row, 6)
+        def idPatient = row[0]
+        def gender    = row[1]
+        def status    = SarekUtils.returnStatus(row[2].toInteger())
+        def idSample  = row[3]
+        def bamFile   = SarekUtils.returnFile(row[4])
+        def baiFile   = SarekUtils.returnFile(row[5])
 
         SarekUtils.checkFileExtension(bamFile,".bam")
         SarekUtils.checkFileExtension(baiFile,".bai")
@@ -199,12 +205,6 @@ class SarekUtils {
     return it
   }
 
-  // Return TSV if it has the correct number of items in row
-  static def returnTSV(it, number) {
-    if (it.size() != number) exit 1, "Malformed row in TSV file: ${it}, see --help for more information"
-    return it
-  }
-
   // Sarek ascii art
   static def sarek_ascii() {
     println "    ____        _____               _    "

lib/SarekUtils.groovy

@@ -547,23 +547,23 @@ def defineStepList() {
 def extractFastq(tsvFile) {
   // Channeling the TSV file containing FASTQ.
   // Format is: "subject gender status sample lane fastq1 fastq2"
-  Channel
-    .from(tsvFile.readLines())
-    .map{line ->
-      def list       = SarekUtils.returnTSV(line.split(),7)
-      def idPatient  = list[0]
-      def gender     = list[1]
-      def status     = SarekUtils.returnStatus(list[2].toInteger())
-      def idSample   = list[3]
-      def idRun      = list[4]
-      def fastqFile1 = SarekUtils.returnFile(list[5])
-      def fastqFile2 = SarekUtils.returnFile(list[6])
-
-      SarekUtils.checkFileExtension(fastqFile1,".fastq.gz")
-      SarekUtils.checkFileExtension(fastqFile2,".fastq.gz")
-
-      [idPatient, gender, status, idSample, idRun, fastqFile1, fastqFile2]
-    }
+  Channel.from(tsvFile)
+  .splitCsv(sep: '\t')
+  .map { row ->
+    SarekUtils.checkNumberOfItem(row, 7)
+    def idPatient  = row[0]
+    def gender     = row[1]
+    def status     = SarekUtils.returnStatus(row[2].toInteger())
+    def idSample   = row[3]
+    def idRun      = row[4]
+    def fastqFile1 = SarekUtils.returnFile(row[5])
+    def fastqFile2 = SarekUtils.returnFile(row[6])
+
+    SarekUtils.checkFileExtension(fastqFile1,".fastq.gz")
+    SarekUtils.checkFileExtension(fastqFile2,".fastq.gz")
+
+    [idPatient, gender, status, idSample, idRun, fastqFile1, fastqFile2]
+  }
 }
 
 def extractFastqFromDir(pattern) {
@@ -602,24 +602,24 @@ def extractFastqFromDir(pattern) {
 def extractRecal(tsvFile) {
   // Channeling the TSV file containing Recalibration Tables.
   // Format is: "subject gender status sample bam bai recalTables"
-  Channel
-    .from(tsvFile.readLines())
-    .map{line ->
-      def list       = SarekUtils.returnTSV(line.split(),7)
-      def idPatient  = list[0]
-      def gender     = list[1]
-      def status     = SarekUtils.returnStatus(list[2].toInteger())
-      def idSample   = list[3]
-      def bamFile    = SarekUtils.returnFile(list[4])
-      def baiFile    = SarekUtils.returnFile(list[5])
-      def recalTable = SarekUtils.returnFile(list[6])
-
-      SarekUtils.checkFileExtension(bamFile,".bam")
-      SarekUtils.checkFileExtension(baiFile,".bai")
-      SarekUtils.checkFileExtension(recalTable,".recal.table")
-
-      [ idPatient, gender, status, idSample, bamFile, baiFile, recalTable ]
-    }
+  Channel.from(tsvFile)
+    .splitCsv(sep: '\t')
+    .map { row ->
+    SarekUtils.checkNumberOfItem(row, 7)
+    def idPatient  = row[0]
+    def gender     = row[1]
+    def status     = SarekUtils.returnStatus(row[2].toInteger())
+    def idSample   = row[3]
+    def bamFile    = SarekUtils.returnFile(row[4])
+    def baiFile    = SarekUtils.returnFile(row[5])
+    def recalTable = SarekUtils.returnFile(row[6])
+
+    SarekUtils.checkFileExtension(bamFile,".bam")
+    SarekUtils.checkFileExtension(baiFile,".bai")
+    SarekUtils.checkFileExtension(recalTable,".recal.table")
+
+    [ idPatient, gender, status, idSample, bamFile, baiFile, recalTable ]
+  }
 }
 
 def flowcellLaneFromFastq(path) {

main.nf

@@ -72,7 +72,7 @@ do
     shift # past value
     ;;
     --tag)
-    tag=$2
+    TAG=$2
     shift # past argument
     shift # past value
     ;;