mzml 0.1.0 → 0.2.0

data/Rakefile

@@ -10,10 +10,10 @@ begin
     gem.email = "angel@delagoya.com"
     gem.homepage = "http://github.com/delagoya/mzml"
     gem.authors = ["Angel Pizarro"]
-    gem.add_development_dependency "rspec", ">= 1.2.9"
+    gem.add_development_dependency "rspec", "1.3.0"
     # gem is a Gem::Specification... see http://www.rubygems.org/read/chapter/20 for additional settings
-    gem.add_dependency  "nokogiri", ">= 1.3.3"
-    
+    gem.add_dependency  "nokogiri", "1.4.1"
+
   end
   Jeweler::GemcutterTasks.new
 rescue LoadError
@@ -39,6 +39,6 @@ task :default => :spec
 require 'yard'
 YARD::Rake::YardocTask.new do |yardoc|
   version = File.exist?('VERSION') ? File.read('VERSION') : ""
-  yardoc.options = ["--title", "mzml #{version}", "-r", "README.rdoc"] 
+  yardoc.options = ["--title", "mzml #{version}", "-r", "README.rdoc"]
   yardoc.files = ['README*','lib/**/*.rb']
 end

data/VERSION

@@ -1 +1 @@
-0.1.0
+0.2.0

data/bin/mzML2mgf.rb

@@ -0,0 +1,53 @@
+#!/opt/local/bin/ruby
+
+################################
+####
+##
+#     David Austin - UPENN
+#     converts mzML to MGF format
+#     set up to replicate msconvert but muuchh slower
+#
+
+require 'rubygems'
+require 'mzml'
+
+
+#first load nokogiri document
+
+mzml =  MzML::Doc.new(ARGV[0])
+
+#now loop through each spectrum.. sort first to be the same as msconvert
+
+sorted_keys = mzml.parse_index_list[:spectrum].keys.sort{ |x,y| x.split('=')[3].to_i <=> y.split('=')[3].to_i }
+
+sorted_keys.each do |k|
+  
+  s = mzml.spectrum(k)
+  unless  s.precursor_list.nil? || s.precursor_list.empty?
+  
+ 
+    #now we print!
+
+    puts "BEGIN IONS"
+    puts "TITLE=#{s.id}"
+    puts "RTINSECONDS=#{s.retention_time}"
+    puts "PEPMASS=#{s.precursor_mass} #{s.precursor_intensity}"
+
+    0.upto(s.mz.length-1) do |i|
+
+      puts "#{sprintf('%5.7f', s.mz[i])} #{sprintf('%4.9f', s.intensity[i])}"
+
+    end
+
+
+    puts "END IONS"
+
+    
+  end
+  
+
+  
+end
+
+
+

data/lib/mzml.rb

@@ -147,13 +147,15 @@ module MzML
     attr_accessor :id, :default_array_length, :spot_id, :type,\
     :charge, :precursor, :base_peak_mz, :base_peak_intensity, :ms_level, \
     :high_mz, :low_mz, :title, :tic, :polarity, :representation, :mz_node, :intensity_node, \
-    :mz, :intensity, :precursor_list, :scan_list, :retention_time
+    :mz, :intensity, :precursor_list, :scan_list, :retention_time, :precursor_mass, :precursor_intensity
+    
     attr_reader :node, :params
 
     # mz & intensity arrays will be don by proper methods maybe.
     def initialize(spectrum_node)
       @node = spectrum_node
       @params = {}
+      @precursor_list = []
       parse_element()
     end
 
@@ -161,11 +163,11 @@ module MzML
     # This method pulls out all of the annotation from the XML node
     def parse_element
       # id
-      @id = @node[:id]
-      @default_array_length = @node[:defaultArrayLength]
-      @spot_id = @node[:spotID]
+      @id = @node.xpath("spectrum")[0][:id]
+      @default_array_length = @node.xpath("spectrum")[0][:defaultArrayLength]
+      @spot_id = @node.xpath("spectrum")[0][:spotID]
       # now reaching into params
-      @params = @node.xpath("cvParam").inject({}) do  |memo,prm|
+      @params = @node.xpath("spectrum/cvParam").inject({}) do  |memo,prm|
         memo[prm[:name]] = prm[:value]
         memo
       end
@@ -178,13 +180,14 @@ module MzML
       # polarity
       # representation
       # precursor list
-      if (@node.xpath("precursorList")[0])
+      if (! @node.xpath("spectrum/precursorList")[0].nil?)
         parse_precursor_list()
+        get_parent_info()
       else
-        @precursor_list = nil
+        @precursor_list = []
       end
       # scan list
-      if (@node.xpath("scanList")[0])
+      if (@node.xpath("spectrum/scanList")[0])
         @scan_list = parse_scan_list()
       else
         @scan_list = nil
@@ -194,14 +197,27 @@ module MzML
     end
 
     def parse_precursor_list
-      @precursor_list = @node.css("precursorList > precursor").each do |p|
+      @node.css("precursorList > precursor").each do |p|
         [p[:spectrumRef], p]
+        @precursor_list << p
       end
     end
 
+    def get_parent_info
+      
+      unless @precursor_list.empty?
+        
+        @precursor_mass = @precursor_list[0].xpath("selectedIonList/selectedIon/cvParam[@accession='MS:1000744']")[0][:value] unless @precursor_list[0].xpath("selectedIonList/selectedIon/cvParam[@accession='MS:1000744']")[0].nil?
+        @precursor_intensity = @precursor_list[0].xpath("selectedIonList/selectedIon/cvParam[@accession='MS:1000042']")[0][:value] unless @precursor_list[0].xpath("selectedIonList/selectedIon/cvParam[@accession='MS:1000042']")[0].nil?
+        
+      end
+        
+      
+    end
+
     def parse_scan_list
-      @scan_list = @node.xpath("scanList/scan")
-      @retention_time = @node.xpath("scanList/scan/cvParam[@accesion='MS:1000016']")[0]
+      @scan_list = @node.xpath("spectrum/scanList/scan")
+      @retention_time = @node.xpath("spectrum/scanList/scan/cvParam[@accession='MS:1000016']")[0][:value] unless @node.xpath("spectrum/scanList/scan/cvParam[@accession='MS:1000016']")[0].nil?
     end
 
     def parse_binary_data
@@ -214,6 +230,7 @@ module MzML
       # 64-bit floats? default is 32-bit
       dtype = @mz_node.xpath("cvParam[@accession='MS:1000523']")[0] ? "E*" : "e*"
       @mz = data.unpack(dtype)
+      
       @intensity_node = @node.xpath("spectrum/binaryDataArrayList/binaryDataArray/cvParam[@accession='MS:1000515']").first.parent
       data = Base64.decode64(@intensity_node.xpath("binary").text)
       if @intensity_node.xpath("cvParam[@accession='MS:1000574']")[0]

data/mzml.gemspec

@@ -5,13 +5,15 @@
 
 Gem::Specification.new do |s|
   s.name = %q{mzml}
-  s.version = "0.1.0"
+  s.version = "0.2.0"
 
   s.required_rubygems_version = Gem::Requirement.new(">= 0") if s.respond_to? :required_rubygems_version=
   s.authors = ["Angel Pizarro"]
-  s.date = %q{2009-12-06}
+  s.date = %q{2010-04-23}
+  s.default_executable = %q{mzML2mgf.rb}
   s.description = %q{A non-validating mzML parser. MzML is a standard data format for representing mass spectrometry data.}
   s.email = %q{angel@delagoya.com}
+  s.executables = ["mzML2mgf.rb"]
   s.extra_rdoc_files = [
     "LICENSE",
      "README.rdoc"
@@ -24,6 +26,7 @@ Gem::Specification.new do |s|
      "README.rdoc",
      "Rakefile",
      "VERSION",
+     "bin/mzML2mgf.rb",
      "lib/mzml.rb",
      "mzml.gemspec",
      "spec/mzml_spec.rb",
@@ -36,7 +39,7 @@ Gem::Specification.new do |s|
   s.homepage = %q{http://github.com/delagoya/mzml}
   s.rdoc_options = ["--charset=UTF-8"]
   s.require_paths = ["lib"]
-  s.rubygems_version = %q{1.3.5}
+  s.rubygems_version = %q{1.3.6}
   s.summary = %q{A non-validating mzML parser}
   s.test_files = [
     "spec/mzml_spec.rb",
@@ -48,15 +51,15 @@ Gem::Specification.new do |s|
     s.specification_version = 3
 
     if Gem::Version.new(Gem::RubyGemsVersion) >= Gem::Version.new('1.2.0') then
-      s.add_development_dependency(%q<rspec>, [">= 1.2.9"])
-      s.add_runtime_dependency(%q<nokogiri>, [">= 1.3.3"])
+      s.add_development_dependency(%q<rspec>, ["= 1.3.0"])
+      s.add_runtime_dependency(%q<nokogiri>, ["= 1.4.1"])
     else
-      s.add_dependency(%q<rspec>, [">= 1.2.9"])
-      s.add_dependency(%q<nokogiri>, [">= 1.3.3"])
+      s.add_dependency(%q<rspec>, ["= 1.3.0"])
+      s.add_dependency(%q<nokogiri>, ["= 1.4.1"])
     end
   else
-    s.add_dependency(%q<rspec>, [">= 1.2.9"])
-    s.add_dependency(%q<nokogiri>, [">= 1.3.3"])
+    s.add_dependency(%q<rspec>, ["= 1.3.0"])
+    s.add_dependency(%q<nokogiri>, ["= 1.4.1"])
   end
 end
 

data/spec/mzml_spec.rb

@@ -26,6 +26,7 @@ describe MzML do
       s = mz.spectrum(mz.index[:spectrum].keys.first)
       s.mz.should_not be_nil
     end
+    
     it "should unmarshall the a 32 byte intensity array" do
       mz = MzML::Doc.new(@file)
       s = mz.spectrum(mz.index[:spectrum].keys.first)
@@ -42,7 +43,39 @@ describe MzML do
       i.join(", ").should be == mgf.intensity.join(", ")
       m.join(", ").should be == mgf.mz.join(", ")
     end
+
+    it "should get a spectrum's id, default array length, retention time" do
+      mz = MzML::Doc.new(@file)
+      s = mz.spectrum(mz.index[:spectrum].keys.first)
+      s.id.should_not be_nil
+      s.default_array_length.should_not be_nil
+      s.retention_time.should_not be_nil
+
+    end
+
+    it "should get a spectrum's precursor information if it has a precursor" do
+      mz = MzML::Doc.new(@file)
+      found_at_least_one_precursor = false
+      mz.index[:spectrum].keys.each do |k|
+        s = mz.spectrum(k)
+        if ! s.precursor_list.empty?
+          s.precursor_mass.should_not be_nil
+          s.precursor_intensity.should_not be_nil
+          found_at_least_one_precursor = true
+          break
+        end
+        
+      end
+      
+      found_at_least_one_precursor.should == true
+      
+
+    end
+    
+
+
   end
+  
 
   context "Given a valid mzML file that uses compression" do
     it "should unmarshall and uncompress the 64 byte mz array" do