pullmerge

Merge branch 'devel' of github.com:trinityrnaseq/trinityrnaseq into devel
wangxuann1995 · Jan 9, 2019 · 3159f7c · 3159f7c
2 parents e69d897 + f833ab2
commit 3159f7c
Show file tree

Hide file tree

Showing 8 changed files with 118 additions and 48 deletions.
diff --git a/.gitignore b/.gitignore
@@ -1,4 +1,5 @@
 *.o
+*.pyc
 *.swp
 *~
 /Chrysalis/bin

diff --git a/.travis.yml b/.travis.yml
@@ -3,6 +3,9 @@ sudo: false
 git:
   depth: 1
 language: generic
+# Custom definitions
+.conditions:
+  if: &secondary type = cron OR env(SECONDARY) = true
 matrix:
   include:
     # http://www.cpan.org/src/
@@ -11,32 +14,57 @@ matrix:
     - language: perl
       perl: "5.26-shrplib"
       env: CC=gcc-8 CXX=g++-8
-    # https://docs.travis-ci.com/user/languages/cpp/
-    - language: cpp
-      compiler: clang
     # https://docs.travis-ci.com/user/languages/java/
     - language: java
       jdk: oraclejdk9
       env: CC=gcc-6 CXX=g++-6
     # https://docs.travis-ci.com/user/languages/python/
     - language: python
       python: "3.6"
+    - language: generic
+      env: CC=clang-3.9 CXX=clang++-3.9 LDFLAGS="-fopenmp=libiomp5"
+      addons:
+        apt:
+          sources:
+            - ubuntu-toolchain-r-test
+            - llvm-toolchain-precise-3.9
+          packages:
+            - clang-3.9
+      if: *secondary
+    - language: generic
+      # Open MP support is available from clang-3.7.
+      env: CC=clang-3.7 CXX=clang++-3.7 LDFLAGS="-fopenmp=libiomp5"
+      addons:
+        apt:
+          sources:
+            - ubuntu-toolchain-r-test
+            - llvm-toolchain-precise-3.7
+          packages:
+            - clang-3.7
+      if: *secondary
     # Test old version of the dependency packages.
     - language: perl
       perl: "5.18-shrplib"
       env: BOWTIE2_VERSION=2.2.6
+      if: *secondary
     # https://docs.travis-ci.com/user/reference/osx
-    - os: osx
-      language: cpp
-      # Do not have to set "compiler: gcc" as it is actually clang on osx.
-      compiler: clang
     - os: osx
       osx_image: xcode9.4
       env: CC=gcc-8 CXX=g++-8
+    - os: osx
+      osx_image: xcode9.4
+      # "clang++" installed by [email protected] instead of "clang++-3.9".
+      env: CC=clang-3.9 CXX=clang++
+      if: *secondary
+    - os: osx
+      osx_image: xcode9.4
+      env: CC=clang-3.7 CXX=clang++-3.7
+      if: *secondary
   allow_failures:
-    - language: cpp
-      compiler: clang
     - env: BOWTIE2_VERSION=2.2.6
+    - env: CC=clang-3.7 CXX=clang++-3.7 LDFLAGS="-fopenmp=libiomp5"
+    - os: osx
+      env: CC=clang-3.7 CXX=clang++-3.7
   fast_finish: true
 install:
   - |
@@ -56,21 +84,33 @@ install:
         echo "Installing ${CXX}."
         sudo apt-get install -qq "${CXX}"
       fi
+      if [[ "${CC}" =~ clang- ]]; then
+        echo "Installing ${CC}."
+        sudo apt-get install -qq "${CC}" libiomp-dev
+      fi
     else
       # osx
+      export HOMEBREW_NO_AUTO_UPDATE=1
       if [[ "${CC}" =~ gcc- ]]; then
         # Update command line tool to avoid an error:
         # "_stdio.h: No such file or directory", when building samtools.
         softwareupdate --list
-        softwareupdate --install "Command Line Tools (macOS High Sierra version 10.13) for Xcode-9.4"
+        softwareupdate --install "Command Line Tools (macOS High Sierra version 10.13) for Xcode-10.1"
 
-        # Uninstall oclint to install gcc on OSX.
-        # https://github.com/travis-ci/travis-ci/issues/8826
-        brew cask uninstall oclint
         CC_PKG="$(echo "${CC}" | sed -e 's/-/@/')"
         echo "Installing ${CC_PKG}."
         brew install "${CC_PKG}"
       fi
+      if [[ "${CC}" =~ clang- ]]; then
+        echo "Installing ${CC}."
+        CC_VERSION=$(echo "${CC}" | sed -e 's/^clang-//')
+        brew install "llvm@${CC_VERSION}"
+        brew install libomp
+        CC_PREFIX=$(brew --prefix "llvm@${CC_VERSION}")
+        export PATH="${CC_PREFIX}/bin:$PATH"
+        export LDFLAGS="-L${CC_PREFIX}/lib -Wl,-rpath,${CC_PREFIX}/lib"
+        export CPPFLAGS="-I${CC_PREFIX}/include"
+      fi
     fi
 
   - pushd "${HOME}"
@@ -93,7 +133,7 @@ install:
   - pushd samtools-1.9
   - autoheader
   - autoconf -Wno-syntax
-  - ./configure
+  - ./configure || (cat config.log; false)
   - make
   - sudo make install
   - popd
@@ -118,7 +158,7 @@ install:
   - wget https://github.com/gmarcais/Jellyfish/releases/download/v${JELLYFISH_VERSION}/jellyfish-${JELLYFISH_VERSION}.tar.gz
   - tar xzf jellyfish-${JELLYFISH_VERSION}.tar.gz
   - pushd jellyfish-${JELLYFISH_VERSION}
-  - ./configure
+  - ./configure || (cat config.log; false)
   - make
   - sudo make install
   - popd
@@ -149,10 +189,17 @@ install:
   - sudo pip install numpy
 
   - popd
-script:
+before_script:
   - |
-    export TRINITY_HOME=`pwd` && \
-    make && \
-    make plugins && \
-    sudo make install > /dev/null && \
-    make test -C sample_data/test_Trinity_Assembly
+    function show_log {
+      for file in $(find . -name config.log); do
+        echo "config.log File: ${file}"
+        cat "${file}"
+      done
+    }
+  - export TRINITY_HOME=`pwd`
+  - make || (show_log; false)
+  - make plugins || (show_log; false)
+  - sudo make install > /dev/null
+script:
+  - make test -C sample_data/test_Trinity_Assembly
diff --git a/Analysis/SuperTranscripts/AllelicVariants/run_variant_calling.py b/Analysis/SuperTranscripts/AllelicVariants/run_variant_calling.py
@@ -12,6 +12,7 @@
 import logging
 import re
 import math
+import glob
 
 sys.path.insert(0, os.path.sep.join([os.path.dirname(os.path.realpath(__file__)),
                                      "..", "..", "..", "PyLib"]))
@@ -88,7 +89,13 @@ def main():
     GATK_HOME = os.getenv("GATK_HOME")
     if not GATK_HOME:
         exit("Error, missing path to GATK in $GATK.")
-
+
+    # identify gatk4_jar file
+    gatk4_jar = glob.glob(os.path.join(GATK_HOME, "gatk-package-4.*-local.jar"))
+    if len(gatk4_jar) != 1:
+        raise RuntimeError("Error, cannot locate single gatk-package-4.*-local.jar in {}".format(GATK_HOME))
+    gatk_path = os.path.abspath(gatk4_jar[0])
+
     # get real paths before changing working directory in case they are relative paths
     if args.paired_reads:
         reads_paths = [os.path.realpath(f) for f in args.paired_reads]
@@ -231,10 +238,10 @@ def main():
         Pipeliner.Command(UTILDIR + "/clean_bam.pl dedupped.bam dedupped.bam.validation dedupped.valid.bam",
                           "make_valid_dedupped_bam.ok"),
 
-        Pipeliner.Command("java -jar " + GATK_HOME + "/GenomeAnalysisTK.jar " +
-                          "-T SplitNCigarReads -R " + st_fa_path +
-                          " -I dedupped.valid.bam -o splitNCigar.bam " +
-                          " -rf ReassignOneMappingQuality -RMQF 255 -RMQT 60 -U ALLOW_N_CIGAR_READS  --validation_strictness LENIENT",
+        Pipeliner.Command("java -jar " + gatk_path +
+                          " SplitNCigarReads -R " + st_fa_path +
+                          " -I dedupped.valid.bam -O splitNCigar.bam " +
+                          " --read-validation-stringency LENIENT",
                           "splitNCigarReads.ok")
         ])
     pipeliner.run()
@@ -245,9 +252,9 @@ def main():
     logger.info("Variant Calling using Haplotype Caller.")
 
     pipeliner.add_commands([
-        Pipeliner.Command("java -jar " + GATK_HOME + "/GenomeAnalysisTK.jar " +
-                          "-T HaplotypeCaller -R " + st_fa_path +
-                          " -I ./splitNCigar.bam -dontUseSoftClippedBases -stand_call_conf 20.0 -o output.vcf",
+        Pipeliner.Command("java -jar " + gatk_path + 
+                          " HaplotypeCaller -R " + st_fa_path +
+                          " -I ./splitNCigar.bam -dont-use-soft-clipped-bases -stand-call-conf 20.0 -O output.vcf",
                           "haplotypecaller.ok")
         ])
     pipeliner.run()
@@ -256,16 +263,16 @@ def main():
     logger.info("Doing some basic filtering of vcf.")
 
     pipeliner.add_commands([
-        Pipeliner.Command("java -jar " + GATK_HOME + "/GenomeAnalysisTK.jar " +
-                          "-T VariantFiltration -R " + st_fa_path +
+        Pipeliner.Command("java -jar " + gatk_path +
+                          " VariantFiltration -R " + st_fa_path +
                           " -V output.vcf -window 35 -cluster 3 " +
-                          "-filterName FS -filter \"FS > 30.0\" " +
-                          "-filterName QD -filter \"QD < 2.0\" -o filtered_output.vcf",
+                          "--filter-name FS -filter \"FS > 30.0\" " +
+                          "--filter-name QD -filter \"QD < 2.0\" -O filtered_output.vcf",
                           "variant_filt.ok")
         ])
 
     pipeliner.run()
-
+    
 
     logger.info("Done!")
 

diff --git a/README.md b/README.md
@@ -1,7 +1,11 @@
 trinityrnaseq
 =============
 
-[![Build Status](https://travis-ci.org/trinityrnaseq/trinityrnaseq.svg?branch=master)](https://travis-ci.org/trinityrnaseq/trinityrnaseq)
+[![Build Status](https://travis-ci.org/trinityrnaseq/trinityrnaseq.svg?branch=devel)](https://travis-ci.org/trinityrnaseq/trinityrnaseq)
 [![install with bioconda](https://img.shields.io/badge/install%20with-bioconda-brightgreen.svg?style=flat-square)](https://bioconda.github.io/recipes/trinity/README.html)
 
 Trinity RNA-Seq de novo transcriptome assembly see the main webpage [http://trinityrnaseq.github.io](http://trinityrnaseq.github.io)
+
+## Contributing
+
+We encourage you to contribute to Trinity! Please check out the [Contributing](https://github.com/trinityrnaseq/trinityrnaseq/wiki/Contributing) for the guidelines.
diff --git a/Trinity b/Trinity
@@ -1086,7 +1086,7 @@ unless ($NO_RUN_INCHWORM_FLAG || $NO_RUN_CHRYSALIS_FLAG || $bypass_java_version_
     if ($java_version =~ /(java|openjdk) version \"1\.(\d)\./) {
         $version_id = $2;
     }
-    elsif ($java_version =~ /java version \"(\d+)\" /) {
+    elsif ($java_version =~ /java version \"(\d+)(\.)?([._\d])+?\" ?/) {
         $version_id = $1;
     }
     elsif ($java_version =~ /java (\d+) /) {

diff --git a/trinity-plugins/Makefile b/trinity-plugins/Makefile
@@ -1,8 +1,8 @@
 
 OS := $(shell uname)
 
-CXX=g++
-CC=gcc
+CXX ?= g++
+CC ?= gcc
 COLLECTL_CODE="collectl-4.1.0"
 PARAFLY_CODE="ParaFly-0.1.0"
 TRIMMOMATIC_CODE=Trimmomatic-0.36

diff --git a/trinity-plugins/scaffold_iworm_contigs/Makefile b/trinity-plugins/scaffold_iworm_contigs/Makefile
@@ -1,4 +1,4 @@
-CXX    = g++
+CXX    ?= g++
 prefix = ../htslib
 
 ScaffoldIwormContigs:

diff --git a/util/misc/randomly_sample_PE_fastq.pl b/util/misc/randomly_sample_PE_fastq.pl
@@ -30,7 +30,13 @@
     srand();
     my @selected_indices = &get_random_indices($num_entries, $num_total_records);
 
-
+    if (scalar(@selected_indices) != $num_entries) {
+        die "Error, get_random_indices returned "  . scalar(@selected_indices) . " instead of $num_total_records";
+    }
+    else {
+        print STDERR "-selected  $num_entries indices. Now outputting selected records\n";
+    }
+
     my %selected = map { + $_ => 1 } @selected_indices;
 
     @selected_indices = ();
@@ -62,8 +68,6 @@
                 . "Right: " . $right_entry->get_core_read_name() . "\n";
         }
 
-        $counter++;
-
         if ($selected{$counter}) {
 
             print $left_ofh $left_entry->get_fastq_record();
@@ -73,18 +77,25 @@
             delete($selected{$counter});
 
         }
-    }
-
-    print STDERR " done.\n";
 
+        $counter++;
+        if ($counter % 100000 == 0) {
+            print STDERR "\r[$counter]   ";
+        }
+    }
+
+    print STDERR "\n\ndone.\n";
+
     close $left_ofh;
     close $right_ofh;
-
-
+
     if (%selected) {
         die "Error, missing indices: " . Dumper(\%selected);
     }
-
+    else {
+        print STDERR "-all records located and output.\n";
+    }
+
     exit(0);
 }
-Original file line number
+Diff line change
@@ -1,4 +1,5 @@
     *.o
+    *.pyc
     *.swp
     *~
     /Chrysalis/bin
@@ Expand Down @@