mascot-dat 0.1.2 → 0.1.3

data/.gitignore

@@ -7,4 +7,5 @@ Gemfile.lock
 test.rb
 gh.yml
 doc/*
-.yardoc
+.yardoc
+scratch.rb

data/Rakefile

@@ -8,11 +8,10 @@ Rake::TestTask.new do |t|
   t.verbose = true
 end
 
-
 # documentation
 require 'yard'
 require 'yard/rake/yardoc_task'
 YARD::Rake::YardocTask.new do |t|
-  t.files   = ['lib/**/*.rb']   # optional
-  t.options = ["--files README.rdoc LICENSE -r README.rdoc"] # optional
-end
+  t.files   = ['lib/**/*.rb', '-' , "README.rdoc","LICENSE"]
+  t.options =  ["-r","README.rdoc"]
+end

data/lib/mascot/dat/masses.rb

@@ -1,23 +1,60 @@
 module Mascot
   class DAT
+
+    # A Hash of the mass table section. See the {#masstable masstable} instance method for details.
     class Masses
+      # The main table of masses. Given the following examples from a DAT file:
+      #
+      #     W=186.079313
+      #     X=111.000000
+      #     Y=163.063329
+      #     Z=128.550590
+      #     Hydrogen=1.007825
+      #     Carbon=12.000000
+      #
+      # You can access the value for Hydrogen as:
+      #     mydat.masses.masstable[:Hydorgen] # => 1.007825
+      # or
+      #     mydat.masses.m[:Hydrogen] # => 1.007825
+      # or
+      #    mydat.masses.hydorgen # =>  1.007825
+      attr_reader :masstable
+      # def masstable
+      #   @masstable
+      # end
+      alias_method :m, :masstable
 
-      # An array of mass deltas. Order corresponds to parameter IT_MODS
+      # A subset of the mass table defining the variable modications. For
+      # example, given the following delta in a DAT file:
+      #
+      #     delta1=15.994915,Oxidation (M)
+      #
+      # Then the following gets defined:
+      #
+      #     @deltas = [[15.994915,"Oxidation (M)"], ... ]
+      #
       attr_reader :deltas
+      alias_method :mods, :deltas
+      alias_method :d, :deltas
 
-      # An array of mass deltas. Order corresponds to parameter MODS
-      attr_reader :fixed_mods
-
-      # A Hash of the mass table. E.g.
+      # A subset of the mass table defining the fixed modifications. For
+      # example:
       #
-      #     self.masstable[:Hydorgen] #=> 1.007825
+      #     FixedMod1=57.021464,Carbamidomethyl (C)
+      #     FixedModResidues1=C
       #
-      attr_reader :masstable
+      # Then the following gets defined:
+      #
+      #     @fixed_modifications = [[57.021464, "Carbamidomethyl (C)", "C"], ...]
+      #
+      attr_reader :fixed_modifications
+      alias_method :fixed_mods, :fixed_modifications
+      alias_method :f, :fixed_modifications
 
       def initialize masses_section
         @masstable = {}
         @deltas = []
-        @fixed_mods = []
+        @fixed_modifications = []
 
         masses_section.split("\n").each do |l|
           k,v = l.split("=")
@@ -32,16 +69,19 @@ module Mascot
             idx = $2.to_i - 1
             if $1.empty?
               # new fixed mod record
-              @fixed_mods[idx] = v.split(",")
-              @fixed_mods[idx][0] = @fixed_mods[idx][0].to_f
-              @masstable[k.to_sym] = @fixed_mods[idx][0]
+              @fixed_modifications[idx] = v.split(",")
+              @fixed_modifications[idx][0] = @fixed_modifications[idx][0].to_f
+              @masstable[k.to_sym] = @fixed_modifications[idx][0]
             else
               # append the modified residue to the array
-              @fixed_mods[idx] << v
+              @fixed_modifications[idx] << v
             end
           else
             @masstable[k.to_sym] = v.to_f
           end
+          @masstable.keys.each do |m|
+            self.class.send(:define_method,m, lambda { @masstable[m] })
+          end
         end
       end
     end

data/lib/mascot/dat/psm.rb

@@ -39,10 +39,8 @@ module Mascot
           k,v = l.split "="
           case k
           when /^q(\d+)_p(\d+)$/
-            # just put query & rank in the beginning, because $1, $2 return zero after the split
             psm_result.query = $1.to_i
             psm_result.rank = $2.to_i
-            #main result, must split value
             psm_vals, prots  = v.split(";")
             psm_vals = psm_vals.split(',')
             psm_result.missed_cleavages= psm_vals[0].to_i

data/lib/mascot/dat/version.rb

@@ -1,6 +1,6 @@
 module Mascot
   class DAT
-    VERSION = "0.1.2"
+    VERSION = "0.1.3"
   end
 end
 

data/lib/mascot/dat.rb

@@ -24,6 +24,7 @@ module Mascot
   class DAT
     attr_reader :idx
     attr_reader :boundary
+    attr_reader :boundary_string
     attr_reader :dat_file
     SECTIONS = ["summary", "decoy_summary", "et_summary", "parameters",
                 "peptides", "decoy_peptides", "et_peptides",
@@ -34,6 +35,7 @@ module Mascot
       @dat_file = File.open(dat_file_path)
       @idx = {}
       @boundary = nil
+      @boundary_string = nil
       @cache_index = cache_index
       parse_index
     end
@@ -53,11 +55,10 @@ module Mascot
     def query(n)
       # search index for this
       bytepos = @idx["query#{n}".to_sym]
-      @dat_file.pos = bytepos
-      @dat_file.readline # ADDED
+      @dat_file.pos = bytepos + @boundary_string.length
       att_rx = /(\w+)\=(.+)/
       q = {}
-      each do |l|
+      @dat_file.each do |l|
         l.chomp
         case l
         when att_rx
@@ -65,9 +66,7 @@ module Mascot
           case k
           when "title"
             q[k.to_sym] = URI.decode(v)
-            # when "Ions1"
-            #   q[k.to_sym] = v.split(",").collect {|e| e.split(":").collect {|ee| ee.to_f}}
-          when "Ions1" # CHANGED POSITION TO GET IN THE BLOCK AND DO THE PARSE TO GET THE ARRAY
+          when "Ions1"
             q[:peaks] = parse_mzi(v)
           else
             q[k.to_sym] = v
@@ -82,15 +81,7 @@ module Mascot
     end
 
     alias_method :spectrum, :query
-    def parse_mzi(ions_str)
-      mzi  = [[],[]]
-      ions_str.split(",").collect do |mzpair|
-        tmp = mzpair.split(":").collect {|e| e.to_f}
-        mzi[0] << tmp[0]
-        mzi[1] << tmp[1]
-      end
-      mzi
-    end
+
 
     # Go to a section of the Mascot DAT file
     def goto(key)
@@ -104,7 +95,7 @@ module Mascot
     # Read a section of the DAT file into memory. THIS IS NOT
     # RECOMMENDED UNLESS YOU KNOW WHAT YOU ARE DOING.
     #
-    # @param [String or Symbol] The section name
+    # @param key [String or Symbol] The section name
     # @return [String] The section of the DAT file as a String. The section
     #                  includes the MIME boundary and  content type
     #                  definition lines.
@@ -165,6 +156,7 @@ module Mascot
         idxf = File.open(idxfn)
         @idx = ::Marshal.load(idxf.read)
         @boundary = @idx[:boundary]
+        @boundary_string = @idx[:boundary_string]
         idxf.close
       else
         create_index()
@@ -178,9 +170,10 @@ module Mascot
       # MIME header line, to parse out boundary
       @dat_file.readline
       @dat_file.readline =~/boundary=(\w+)$/
-      boundary_string = "--#{$1}"
-      @boundary = /#{boundary_string}/
+      @boundary_string = "--#{$1}"
+      @boundary = /#{@boundary_string}/
       @idx[:boundary] = @boundary
+      @idx[:boundary_string] = @boundary_string
       @dat_file.grep(@boundary) do |l|
         break if @dat_file.eof?
         section_position = @dat_file.pos - l.length
@@ -195,5 +188,16 @@ module Mascot
       end
       @dat_file.rewind
     end
+
+    # Parse the ion string of mz/intensity peaks in Ions section
+    # Peaks are not ordered, so we must account for that.
+    def parse_mzi(ions_str)
+      mzi_tmp = []
+      ions_str.split(",").collect do |mzpair|
+        mzi_tmp <<  mzpair.split(":").collect {|e| e.to_f}
+      end
+      # now sort the mz_tmp array as ascending m/z, and return the array
+      mzi_tmp.sort
+    end
   end
 end

data/test/fixtures/query23_peaks.dmp

@@ -0,0 +1 @@
+[&[f59.044502f1.951e5[f76.396653f1.951e5[f88.063115f1.951e5[f92.727062f1.951e5[f111.734216f1.951e5[f114.091341f1.951e5[f122.082957f1.951e5[f138.586954f1.951e5[f160.757021f1.951e5[f167.097686f1.951e5[f171.105762f1.951e5[f175.118953f1.951e5[f182.620797f1.951e5[f190.112916f1.951e5[f206.443325f1.951e5[f223.795476f1.951e5[f227.175405f1.951e5[f240.631893f1.951e5[f244.138013f1.951e5[f256.155004f1.951e5[f276.166632f1.951e5[f284.665736f1.951e5[f309.16135f1.951e5[f333.188096f1.951e5[f335.189576f1.951e5[f364.234317f1.951e5[f365.703382f1.951e5[f480.256511f1.951e5[f511.302732f1.951e5[f568.324196f1.951e5[f617.315423f1.951e5[f669.371875f1.951e5[f730.399487f1.951e5

data/test/test_mascot-dat-helper.rb

@@ -9,4 +9,7 @@ class TestMascotDatHelper < Test::Unit::TestCase
     pass
   end
 
+  # def teardown
+  #   File.unlink("test/fixtures/example.dat.idx")
+  # end
 end

data/test/test_mascot-dat-masses.rb

@@ -17,6 +17,7 @@ class TestMascotDatMasses < TestMascotDatHelper
   def test_masses_delta1
     # delta1=15.994915,Oxidation (M)
     assert_equal(15.994915,@masses.masstable[:delta1])
+    assert_equal(15.994915,@masses.delta1)
   end
   def test_masses_var_mod_is_delta1
     assert_equal(15.994915,@masses.deltas[0][0])
@@ -24,6 +25,7 @@ class TestMascotDatMasses < TestMascotDatHelper
   end
   def test_masses_FixedMod1_mass
     assert_equal(57.021464,@masses.masstable[:FixedMod1])
+    assert_equal(57.021464,@masses.FixedMod1)
   end
 
   def test_masses_fixed_mod_is_FixedMod1

data/test/test_mascot-dat.rb

@@ -35,10 +35,17 @@ class TestMascotDat < TestMascotDatHelper
     assert_equal(expected_section, @dat.read_section("enzyme"))
     assert_equal(expected_section, @dat.read_section(:enzyme))
   end
-    def test_read_section_masses
+
+  def test_read_section_masses
     expected_section = File.read("test/fixtures/masses_section.txt")
     assert_equal(expected_section, @dat.read_section("masses"))
     assert_equal(expected_section, @dat.read_section(:masses))
   end
 
+  def test_peaks
+    expected_peaks = Marshal.load(File.read("test/fixtures/query23_peaks.dmp"))
+    query23 = @dat.query(23)
+    assert_equal(expected_peaks,query23[:peaks])
+  end
+
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: mascot-dat
 version: !ruby/object:Gem::Version
-  version: 0.1.2
+  version: 0.1.3
   prerelease: 
 platform: ruby
 authors:
@@ -9,11 +9,11 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2012-06-25 00:00:00.000000000 Z
+date: 2012-07-03 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: rake
-  requirement: &70321258178380 !ruby/object:Gem::Requirement
+  requirement: &70330121470000 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -21,10 +21,10 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *70321258178380
+  version_requirements: *70330121470000
 - !ruby/object:Gem::Dependency
   name: yard
-  requirement: &70321258193380 !ruby/object:Gem::Requirement
+  requirement: &70330121469460 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -32,7 +32,7 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *70321258193380
+  version_requirements: *70330121469460
 description: Mascot DAT file format parser
 email:
 - angel@upenn.edu
@@ -61,6 +61,7 @@ files:
 - test/fixtures/example.dat
 - test/fixtures/header_section.txt
 - test/fixtures/masses_section.txt
+- test/fixtures/query23_peaks.dmp
 - test/test_mascot-dat-enzyme.rb
 - test/test_mascot-dat-header_info.rb
 - test/test_mascot-dat-helper.rb
@@ -85,7 +86,7 @@ required_ruby_version: !ruby/object:Gem::Requirement
       version: '0'
       segments:
       - 0
-      hash: -2702393217074327474
+      hash: -998556290879411337
 required_rubygems_version: !ruby/object:Gem::Requirement
   none: false
   requirements:
@@ -94,7 +95,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
       segments:
       - 0
-      hash: -2702393217074327474
+      hash: -998556290879411337
 requirements: []
 rubyforge_project: 
 rubygems_version: 1.8.11
@@ -106,6 +107,7 @@ test_files:
 - test/fixtures/example.dat
 - test/fixtures/header_section.txt
 - test/fixtures/masses_section.txt
+- test/fixtures/query23_peaks.dmp
 - test/test_mascot-dat-enzyme.rb
 - test/test_mascot-dat-header_info.rb
 - test/test_mascot-dat-helper.rb