ms-mascot 0.1.0 → 0.2.0

metadata

@@ -1,15 +1,16 @@
 --- !ruby/object:Gem::Specification 
 name: ms-mascot
 version: !ruby/object:Gem::Version 
-  version: 0.1.0
+  version: 0.2.0
 platform: ruby
 authors: 
 - Simon Chiang
+- John Prince
 autorequire: 
 bindir: bin
 cert_chain: []
 
-date: 2008-12-03 00:00:00 -07:00
+date: 2009-02-24 00:00:00 -07:00
 default_executable: 
 dependencies: 
 - !ruby/object:Gem::Dependency 
@@ -20,7 +21,7 @@ dependencies:
     requirements: 
     - - ">="
       - !ruby/object:Gem::Version 
-        version: "0.11"
+        version: "0.12"
     version: 
 - !ruby/object:Gem::Dependency 
   name: tap-http
@@ -30,7 +31,7 @@ dependencies:
     requirements: 
     - - ">="
       - !ruby/object:Gem::Version 
-        version: 0.2.0
+        version: 0.3.2
     version: 
 - !ruby/object:Gem::Dependency 
   name: external
@@ -42,16 +43,6 @@ dependencies:
       - !ruby/object:Gem::Version 
         version: 0.3.0
     version: 
-- !ruby/object:Gem::Dependency 
-  name: mspire
-  type: :runtime
-  version_requirement: 
-  version_requirements: !ruby/object:Gem::Requirement 
-    requirements: 
-    - - ">="
-      - !ruby/object:Gem::Version 
-        version: 0.5.0
-    version: 
 - !ruby/object:Gem::Dependency 
   name: ms-in_silico
   type: :runtime
@@ -60,17 +51,7 @@ dependencies:
     requirements: 
     - - ">="
       - !ruby/object:Gem::Version 
-        version: 0.2.0
-    version: 
-- !ruby/object:Gem::Dependency 
-  name: ms-testdata
-  type: :development
-  version_requirement: 
-  version_requirements: !ruby/object:Gem::Requirement 
-    requirements: 
-    - - ">="
-      - !ruby/object:Gem::Version 
-        version: 0.0.1
+        version: 0.2.2
     version: 
 description: 
 email: simon.a.chiang@gmail.com
@@ -81,26 +62,43 @@ extensions: []
 extra_rdoc_files: 
 - README
 - MIT-LICENSE
+- History
 files: 
-- cmd/generate_mgf.rb
-- cmd/generate_prospector_mgf.rb
-- cmd/reformat_mgf.rb
 - lib/ms/mascot.rb
+- lib/ms/mascot/dat.rb
+- lib/ms/mascot/dat/archive.rb
+- lib/ms/mascot/dat/header.rb
+- lib/ms/mascot/dat/index.rb
+- lib/ms/mascot/dat/masses.rb
+- lib/ms/mascot/dat/parameters.rb
+- lib/ms/mascot/dat/peptides.rb
+- lib/ms/mascot/dat/proteins.rb
+- lib/ms/mascot/dat/query.rb
+- lib/ms/mascot/dat/section.rb
+- lib/ms/mascot/dat/summary.rb
+- lib/ms/mascot/dat/summary/id.rb
 - lib/ms/mascot/export.rb
+- lib/ms/mascot/format_mgf.rb
 - lib/ms/mascot/fragment.rb
 - lib/ms/mascot/mgf.rb
 - lib/ms/mascot/mgf/archive.rb
 - lib/ms/mascot/mgf/entry.rb
-- lib/ms/mascot/predict.rb
 - lib/ms/mascot/spectrum.rb
 - lib/ms/mascot/submit.rb
+- tap.yml
 - README
 - MIT-LICENSE
+- History
 has_rdoc: true
 homepage: http://mspire.rubyforge.org/projects/ms-mascot/
 post_install_message: 
-rdoc_options: []
-
+rdoc_options: 
+- --main
+- README
+- -S
+- -N
+- --title
+- Ms-Mascot
 require_paths: 
 - lib
 required_ruby_version: !ruby/object:Gem::Requirement 

data/cmd/generate_mgf.rb

@@ -1,123 +0,0 @@
-# = Usage
-# tap generate_mgf {options} protein_sequences
-#
-# When specifying the ions to include, alternate charge states can be
-# specified using + and -, for example 'y++' or 'b-'.  The available ion
-# series are [a,b,c,x,y,z].
-#
-# = Description
-# Digests, fragments, then formats the protein sequences into mgf files.
-# Use the options to specify/modify digestion enzymes, as well as the
-# type of ions to generate.  
-#
-# = Information
-#
-# Copyright (c) 2006-2007, Regents of the University of Colorado.
-# Developer:: Simon Chiang, Biomolecular Structure Program
-# Homepage:: http://hsc-proteomics.uchsc.edu/hansen_lab
-# Support:: CU Denver School of Medicine Deans Academic Enrichment Fund
-#
-
-require 'tap/script'
-include Constants::Library
-
-app = Tap::App.instance   
-
-#
-# handle options
-#
-
-opts = Prospector::Digest.configurations.to_opts 
-opts += Mascot::Formats::Mgf::Print.configurations.to_opts
-opts += [
-  ['--charge', '-c', GetoptLong::REQUIRED_ARGUMENT, "Parent ion charge for mgf files. (default +1)"],
-  
-  ['--ions', '-i', GetoptLong::REQUIRED_ARGUMENT, "Comma-separated string of ion series to include. (default 'yb')"],
-  #['--enzyme_file', nil, GetoptLong::REQUIRED_ARGUMENT, "Specifes a Prospector-style enzyme config file."],
-  ['--residue_precision', nil, GetoptLong::REQUIRED_ARGUMENT, "The precision of residues, ex 6 for 57.021464"],
-  ['--help', '-h', GetoptLong::NO_ARGUMENT, "Print this help."],
-  ['--debug', nil, GetoptLong::NO_ARGUMENT, "Specifes debug mode."]]
-
-digest_config = {}
-print_config = {}
-series = "yb"
-charge = 1
-residue_precision = 6
-
-Tap::Script.handle_options(*opts) do |opt, value| 
-  case opt
-  when '--help'
-    puts Tap::Script.usage(__FILE__, "Usage", "Description", "Information", :keep_headers => false)
-    puts
-    puts Tap::Script.usage_options(opts)
-    exit
-    
-  when '--debug'
-    app.options.debug = true
-    
-  when '--ions'
-    series = value
-    
-  when '--charge'
-    charge = value.to_i
-    
-  when '--residue_precision'
-    residue_precision = value.to_i
-    
-  else
-    key = Prospector::Digest.configurations.opt_map(opt)
-    digest_config[key] = YAML.load(value) if key
-    
-    key = Mascot::Formats::Mgf::Print.configurations.opt_map(opt)
-    print_config[key] = YAML.load(value) if key
-  end
-end
-
-if ARGV.empty?
-  puts "no sequences specified"
-  exit
-end
-
-#
-# add your script code here
-#
-series = series.scan(/\w\-*\+*/)
-
-#loader = Prospector::LoadDigesters.new
-#loader.enq(enzyme_file)
-
-#
-digest = Prospector::Digest.new(nil, digest_config)
-
-#
-n = Molecule[digest.nterm]
-c = Molecule[digest.cterm]
-
-fragment = Tap::Task.new do |task, polypeptides|
-  polypeptides.collect do |polypeptide, start_index, end_index|
-    task.log :fragment, polypeptide.sequence[0..10] + (polypeptide.sequence.length > 10 ? "..." : "")
-    
-    f = Mascot::FragmentSpectrum.new(polypeptide.sequence, n, c, residue_precision)
-  
-    headers = {
-      :title => polypeptide.sequence,
-      :charge => charge,
-      :pepmass => (f.mass(n) + f.ladder.last + f.mass(c) + charge * f.proton_mass)/charge
-    }
-  
-    data = series.collect {|s| f.series(s)}.flatten.delete_if {|mass| mass < 0 }.sort
-    data = [data, Array.new(data.length, 1)].transpose
-  
-    Mascot::Formats::Mgf::Entry.new(headers, data)
-  end
-end
-
-#
-print = Mascot::Formats::Mgf::Print.new('generate_mgf', print_config)
-
-# workflow
-digest.enq(*ARGV)
-ARGV.clear 
-
-app.sequence(digest, fragment, print)
-app.run

data/cmd/generate_prospector_mgf.rb

@@ -1,123 +0,0 @@
-# = Usage
-# tap generate_mgf {options} protein_sequences
-#
-# When specifying the ions to include, alternate charge states can be
-# specified using + and -, for example 'y++' or 'b-'.  The available ion
-# series are [a,b,c,x,y,z].
-#
-# = Description
-# Digests, fragments, then formats the protein sequences into mgf files.
-# Use the options to specify/modify digestion enzymes, as well as the
-# type of ions to generate.  
-#
-# = Information
-#
-# Copyright (c) 2006-2007, Regents of the University of Colorado.
-# Developer:: Simon Chiang, Biomolecular Structure Program
-# Homepage:: http://hsc-proteomics.uchsc.edu/hansen_lab
-# Support:: CU Denver School of Medicine Deans Academic Enrichment Fund
-#
-
-require 'tap/script'
-include Constants::Library
-
-app = Tap::App.instance   
-
-#
-# handle options
-#
-
-opts = Prospector::Digest.configurations.to_opts 
-opts += Mascot::Formats::Mgf::Print.configurations.to_opts
-opts += [
-  ['--charge', '-c', GetoptLong::REQUIRED_ARGUMENT, "Parent ion charge for mgf files. (default +1)"],
-  
-  ['--ions', '-i', GetoptLong::REQUIRED_ARGUMENT, "Comma-separated string of ion series to include. (default 'yb')"],
-  #['--enzyme_file', nil, GetoptLong::REQUIRED_ARGUMENT, "Specifes a Prospector-style enzyme config file."],
-  ['--residue_precision', nil, GetoptLong::REQUIRED_ARGUMENT, "The precision of residues, ex 6 for 57.021464"],
-  ['--help', '-h', GetoptLong::NO_ARGUMENT, "Print this help."],
-  ['--debug', nil, GetoptLong::NO_ARGUMENT, "Specifes debug mode."]]
-
-digest_config = {}
-print_config = {}
-series = "yb"
-charge = 1
-residue_precision = 6
-
-Tap::Script.handle_options(*opts) do |opt, value| 
-  case opt
-  when '--help'
-    puts Tap::Script.usage(__FILE__, "Usage", "Description", "Information", :keep_headers => false)
-    puts
-    puts Tap::Script.usage_options(opts)
-    exit
-    
-  when '--debug'
-    app.options.debug = true
-    
-  when '--ions'
-    series = value
-    
-  when '--charge'
-    charge = value.to_i
-    
-  when '--residue_precision'
-    residue_precision = value.to_i
-    
-  else
-    key = Prospector::Digest.configurations.opt_map(opt)
-    digest_config[key] = YAML.load(value) if key
-    
-    key = Mascot::Formats::Mgf::Print.configurations.opt_map(opt)
-    print_config[key] = YAML.load(value) if key
-  end
-end
-
-if ARGV.empty?
-  puts "no sequences specified"
-  exit
-end
-
-#
-# add your script code here
-#
-series = series.scan(/\w\-*\+*/)
-
-#loader = Prospector::LoadDigesters.new
-#loader.enq(enzyme_file)
-
-#
-digest = Prospector::Digest.new(nil, digest_config)
-
-#
-n = Molecule[digest.nterm]
-c = Molecule[digest.cterm]
-
-fragment = Tap::Task.new do |task, polypeptides|
-  polypeptides.collect do |polypeptide, start_index, end_index|
-    task.log :fragment, polypeptide.sequence[0..10] + (polypeptide.sequence.length > 10 ? "..." : "")
-    
-    f = Prospector::FragmentSpectrum.new(polypeptide.sequence, n, c)
-  
-    headers = {
-      :title => polypeptide.sequence,
-      :charge => charge,
-      :pepmass => (n.mass + polypeptide.mass + c.mass + charge * (Molecule['H'].mass - Particle['Electron'].mass))/charge
-    }
-  
-    data = series.collect {|s| f.series(s)}.flatten.delete_if {|mass| mass < 0 }.sort
-    data = [data, Array.new(data.length, 1)].transpose
-  
-    Mascot::Formats::Mgf::Entry.new(headers, data)
-  end
-end
-
-#
-print = Mascot::Formats::Mgf::Print.new('generate_mgf', print_config)
-
-# workflow
-digest.enq(*ARGV)
-ARGV.clear 
-
-app.sequence(digest, fragment, print)
-app.run

data/cmd/reformat_mgf.rb

@@ -1,90 +0,0 @@
-# = Usage
-# tap reformat_mgf {options} MGF_FILES
-#
-# = Description
-# Reformats mgf files to a standard output like:
-#
-#   BEGIN IONS
-#   TITLE=7100401blank.190.190.2.dta
-#   CHARGE=2+
-#   PEPMASS=321.571138
-#   100.266 2.0
-#   111.323 2.5
-#   ...
-#   496.110 3.3
-#   601.206 1.3
-#   END IONS
-#
-# = Information
-#
-# Copyright (c) 2006-2007, Regents of the University of Colorado.
-# Developer:: Simon Chiang, Biomolecular Structure Program
-# Homepage:: http://hsc-proteomics.uchsc.edu/hansen_lab
-# Support:: CU Denver School of Medicine Deans Academic Enrichment Fund
-#
-require 'tap/script'
-
-app = Tap::App.instance   
-
-#
-# handle options
-#
-
-opts = [
-  ['--target_dir', '-t', GetoptLong::REQUIRED_ARGUMENT, "Specify an output directory."],
-  ['--mz_precision', '-m', GetoptLong::REQUIRED_ARGUMENT, "Specify the mz precision."],
-  ['--intensity_precision', '-i', GetoptLong::REQUIRED_ARGUMENT, "Specify the intensity precision."],
-  ['--pepmass_precision', '-p', GetoptLong::REQUIRED_ARGUMENT, "Specify the peptide mass precision."],
-  ['--headers', nil, GetoptLong::REQUIRED_ARGUMENT, "Specify the headers to include, separated by commas."],
-  ['--help', '-h', GetoptLong::NO_ARGUMENT, "Print this help."],
-  ['--debug', nil, GetoptLong::NO_ARGUMENT, "Specifies debug mode."]]
-
-config = {:target_dir => 'reformatted'}
-
-Tap::Script.handle_options(*opts) do |opt, value| 
-  case opt
-  when '--help'
-    puts Tap::Script.usage(__FILE__, "Usage", "Description", "Information", :keep_headers => false)
-    puts
-    puts Tap::Script.usage_options(opts)
-    exit
-    
-  when '--debug'
-    app.options.debug = true
-
-  when '--headers'
-    value = value[1..-2] if value[0] == 34 && value[-1] == 34
-    config[:headers] = value.split(/,/).collect {|header| header.strip}
-  else
-    opt =~ /--(.*)/
-    config[$1.to_sym] = value
-  
-  end
-end
-
-#
-# add your script code here
-#
-
-require 'mascot/formats/mgf'
-
-reformat = Tap::FileTask.new("", config) do |task, input|
-  target = task.filepath(task.config[:target_dir], File.basename(input))
-  task.prepare(target)
-
-  task.log_basename :reformatting, input
-  Mascot::Formats::Mgf::Archive.open(input) do |archive|
-    archive.reindex if archive.length == 0
-    
-    File.open(target, "wb") do |output|
-      archive.each do |mgf|
-        mgf.puts(output, task.config)
-        output.puts
-      end
-    end
-  end
-end
-
-args = ARGV.dup
-ARGV.clear
-app.run(reformat, *args)