dirseq 0.2.1 → 0.3.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: f1af55dda3a8cfe508bf3cf49fb877dddfbda7b6
-  data.tar.gz: e2334d82dfcaf555255234cb61e409a0c3bf262d
+  metadata.gz: 48124cb9d1c64721dce10b66a4ae9c6d1773ae5c
+  data.tar.gz: 7f7ee356e059956cc204be699d67fab65fa6c41e
 SHA512:
-  metadata.gz: 43d4538cc5bd983468642056f80d3386663c93b7b7a64ecf2c2917fd84592c75787851c83c8657606f228dc1008e15e67733a48ac713e45bb9b2a95556ca37df
-  data.tar.gz: 155d7a12624068d8cc8cf20ce2b672c027601a6cc39741ebf78d670cef22779899d26f1073ef2450c2f200bac31f3349262c4d8994094e5138430be692b63596
+  metadata.gz: 64263f1b13de9816f978c79961a188f4ab306ff1d7abd0f94d1bfdb6d9c4c807bec7b96b7651afbf1158b2dac38135b6970e273e018bb7f9fdf1fbc9cc736b5e
+  data.tar.gz: 6031c1e7952072c0c379a744bf067ae7f2aa9d7328dfec698ed4af84194ac407eb4b18948a3717ef7538f569cdd787a9f0ad4455025ac8148604834c7bb33988

data/README.md

@@ -37,7 +37,8 @@ Optional parameters:
         --measure-type TYPE          what to count for each gene [options: count, coverage][default: coverage]
         --accepted-feature-types TYPE
                                      Print only features of these type(s) [default CDS]
-
+        --comment-fields             Print elements from the comments in the GFF file [default ID]
+        
 Verbosity:
 
     -q, --quiet                      Run quietly, set logging to ERROR level [default INFO]

data/VERSION

@@ -1 +1 @@
-0.2.1
+0.3.0

data/bin/dirseq

@@ -24,6 +24,7 @@ options = {
   :count_type => COVERAGE_COUNT_TYPE,
   :forward_read_only => false,
   :accepted_feature_types => ['CDS'],
+  :comment_fields_to_print => ['ID'],
 }
 o = OptionParser.new do |opts|
   opts.banner = "
@@ -52,6 +53,10 @@ o = OptionParser.new do |opts|
           "Print only features of these type(s) [default #{options[:accepted_feature_types].join(',')}]") do |arg|
     options[:accepted_feature_types] = Set.new(arg)
   end
+  opts.on("--comment-fields", Array,
+          "Print elements from the comments in the GFF file [default #{options[:comment_fields_to_print].join(',')}]") do |arg|
+    options[:comment_fields_to_print] = arg
+  end
 
   # logger options
   opts.separator "\nVerbosity:\n\n"
@@ -246,6 +251,9 @@ else
   raise
 end
 headers.push 'annotation'
+options[:comment_fields_to_print].each do |field|
+  headers.push field
+end
 puts headers.join("\t")
 
 covs_fwd.each do |feature, cov_fwd|
@@ -268,5 +276,13 @@ covs_fwd.each do |feature, cov_fwd|
   ]
   to_print.push cov_rev unless options[:ignore_directions]
   to_print.push product
+  options[:comment_fields_to_print].each do |field|
+    answer1 = record.attributes.select{|a| a[0] == field}
+    if answer1.empty?
+      to_print.push ''
+    else
+      to_print.push answer1[0][1]
+    end
+  end
   puts to_print.join("\t")
 end

data/spec/script_spec.rb

@@ -6,9 +6,9 @@ describe 'script' do
 
   it "should regular mode" do
     answer = %w(
-    contig	type	start	end	strand	forward_average_coverage	reverse_average_coverage	annotation
+    contig	type	start	end	strand	forward_average_coverage	reverse_average_coverage	annotation	ID
     ).join("\t")+"\n"+%w(
-    contig_100	CDS	2	127	+	0.0	1.1428571428571428	unannotated
+    contig_100	CDS	2	127	+	0.0	1.1428571428571428	unannotated	40_1
     ).join("\t")+"\n"
 
     found = Bio::Commandeer.run "#{path_to_script} --bam #{data_dir}/eg.bam --gff #{data_dir}/eg.gff -q"
@@ -20,9 +20,9 @@ describe 'script' do
     found = Bio::Commandeer.run "#{path_to_script} --bam #{data_dir}/eg.bam --gff #{data_dir}/eg.gff -q --ignore-direction"
 
     answer = %w(
-    contig	type	start	end	strand	average_coverage	annotation
+    contig	type	start	end	strand	average_coverage	annotation	ID
     ).join("\t")+"\n"+%w(
-    contig_100	CDS	2	127	+	1.1428571428571428	unannotated
+    contig_100	CDS	2	127	+	1.1428571428571428	unannotated	40_1
     ).join("\t")+"\n"
 
     found.should == answer
@@ -30,9 +30,9 @@ describe 'script' do
 
   it 'should not fail when the GFF has a FASTA section' do
     answer = %w(
-    contig	type	start	end	strand	forward_average_coverage	reverse_average_coverage	annotation
+    contig	type	start	end	strand	forward_average_coverage	reverse_average_coverage	annotation	ID
     ).join("\t")+"\n"+%w(
-    contig_100	CDS	2	127	+	0.0	1.1428571428571428	unannotated
+    contig_100	CDS	2	127	+	0.0	1.1428571428571428	unannotated	40_1
     ).join("\t")+"\n"
 
     found = Bio::Commandeer.run "#{path_to_script} --bam #{data_dir}/eg.bam --gff #{data_dir}/eg_with_fasta.gff -q"
@@ -42,10 +42,10 @@ describe 'script' do
 
   it 'should print annotation out properly' do
     answer = %w(
-    contig	type	start	end	strand	forward_average_coverage	reverse_average_coverage	annotation
+    contig	type	start	end	strand	forward_average_coverage	reverse_average_coverage	annotation	ID
     ).join("\t")+"\n"+%w(
     contig_100	CDS	2	127	+	0.0	1.1428571428571428	putative
-    ).join("\t")+" methyltransferase YcgJ\n"
+    ).join("\t")+" methyltransferase YcgJ	PROKKA_00001\n"
 
     found = Bio::Commandeer.run "#{path_to_script} --bam #{data_dir}/eg.bam --gff #{data_dir}/realer.gff -q"
 
@@ -54,10 +54,10 @@ describe 'script' do
 
   it 'should print counts correctly' do
     answer = %w(
-    contig	type	start	end	strand	forward_read_count	reverse_read_count	annotation
+    contig	type	start	end	strand	forward_read_count	reverse_read_count	annotation	ID
     ).join("\t")+"\n"+%w(
     contig_100	CDS	2	127	+	0.0	2.0	putative
-    ).join("\t")+" methyltransferase YcgJ\n"
+    ).join("\t")+" methyltransferase YcgJ	PROKKA_00001\n"
 
     found = Bio::Commandeer.run "#{path_to_script} --bam #{data_dir}/eg.bam --gff #{data_dir}/realer.gff -q --measure-type count"
 
@@ -66,10 +66,10 @@ describe 'script' do
 
   it 'should count only the forward read when asked' do
     answer = %w(
-    contig	type	start	end	strand	forward_read_count	reverse_read_count	annotation
+    contig	type	start	end	strand	forward_read_count	reverse_read_count	annotation	ID
     ).join("\t")+"\n"+%w(
     contig_100	CDS	2	127	+	0.0	1.0	putative
-    ).join("\t")+" methyltransferase YcgJ\n"
+    ).join("\t")+" methyltransferase YcgJ	PROKKA_00001\n"
 
     found = Bio::Commandeer.run "#{path_to_script} --bam #{data_dir}/eg.bam --gff #{data_dir}/realer.gff -q --measure-type count --forward-read-only"
 

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: dirseq
 version: !ruby/object:Gem::Version
-  version: 0.2.1
+  version: 0.3.0
 platform: ruby
 authors:
 - Ben J. Woodcroft
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2018-09-20 00:00:00.000000000 Z
+date: 2018-10-01 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bio-commandeer