ms-mascot 0.2.0 → 0.2.2

data/lib/ms/mascot/export.rb

@@ -1,84 +1,67 @@
-require 'tap/http/submit'
+require 'tap/mechanize/submit'
+require 'ms/mascot/validation'
 
 module Ms
   module Mascot
     # :startdoc::manifest exports results from a search
-    class Export < Tap::Http::Submit
-    
+    class Export < Tap::Mechanize::Submit
+      include Validation
+      
       # The MatrixScience public search site
-      DEFAULT_URI = "http://www.matrixscience.com/cgi/export_dat_2.pl"
+      config :uri, "http://www.matrixscience.com/cgi/export_dat_2.pl"
 
       # Parameters for a typical export
-      DEFAULT_PARAMS = {
-        "pep_expect"=>"1",
-        "prot_mass"=>"1",
-        "protein_master"=>"1",
-        "_server_mudpit_switch"=>"0.000000001",
-        "pep_exp_mz"=>"1",
-        "do_export"=>"1",
-        "pep_delta"=>"1",
-        "export_format"=>"XML",
-        "prot_acc"=>"1",
-        "pep_score"=>"1",
-        "show_format"=>"1",
-        "_showsubsets"=>"0",
-        "_show_decoy_report"=>"",
-        "pep_scan_title"=>"1",
-        "pep_miss"=>"1",
-        "pep_calc_mr"=>"1",
-        "pep_exp_mr"=>"1",
-        "prot_score"=>"1",
-        "pep_query"=>"1",
-        "peptide_master"=>"1",
-        "prot_matches"=>"1",
-        "_onlyerrortolerant"=>"",
-        "_showallfromerrortolerant"=>"",
-        "prot_hit_num"=>"1",
-        "search_master"=>"1",
-        "_sigthreshold"=>"0.05",
-        "show_params"=>"1",
-        "show_mods"=>"1",
-        "show_header"=>"1",
-        "pep_isbold"=>"1",
-        "pep_seq"=>"1",
-        "pep_exp_z"=>"1",
-        "prot_desc"=>"1",
-        "_ignoreionsscorebelow"=>"0",
-        "REPORT"=>"AUTO",
-        "pep_rank"=>"1",
-        "pep_var_mod"=>"1",
-        "_noerrortolerant"=>""
-      }
-
-      # Typical headers for an export
-      DEFAULT_HEADERS = {
-        "Keep-Alive"=>"300",
-        "Accept-Encoding"=>"gzip,deflate",
-        "Accept-Language"=>"en-us,en;q=0.5",
-        "Content-Type"=> "multipart/form-data; boundary=---------------------------168072824752491622650073",
-        "Accept-Charset"=>"ISO-8859-1,utf-8;q=0.7,*;q=0.7",
-        "Accept"=>"text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8",
-        "Connection"=>"keep-alive"
-      }
-      
-      config :uri, DEFAULT_URI                           # The uri of the mascot search site
-      config :headers, DEFAULT_HEADERS, &c.hash          # a hash of request headers
-      config :params, DEFAULT_PARAMS, &c.hash            # a hash of query parameters
-      config :request_method, 'GET'                      # the request method (get or post)
-      config :version, 1.1                               # the HTTP version
-      config :redirection_limit, nil, &c.integer_or_nil  # the redirection limit for the request
+      nest :params do
+        config "pep_expect", "1", &MASCOT_SWITCH
+        config "prot_mass", "1", &MASCOT_SWITCH
+        config "protein_master", "1", &MASCOT_SWITCH
+        config "_server_mudpit_switch", 0.000000001, &c.num
+        config "pep_exp_mz", "1", &MASCOT_SWITCH
+        config "do_export", "1", &MASCOT_SWITCH
+        config "pep_delta", "1", &MASCOT_SWITCH
+        config "export_format", "XML", &c.string
+        config "prot_acc", "1", &MASCOT_SWITCH
+        config "pep_score", "1", &MASCOT_SWITCH
+        config "show_format", "1", &MASCOT_SWITCH
+        config "_showsubsets", "0", &MASCOT_SWITCH
+        config "_show_decoy_report", ""
+        config "pep_scan_title", "1", &MASCOT_SWITCH
+        config "pep_miss", "1", &MASCOT_SWITCH
+        config "pep_calc_mr", "1", &MASCOT_SWITCH
+        config "pep_exp_mr", "1", &MASCOT_SWITCH
+        config "prot_score", "1", &MASCOT_SWITCH
+        config "pep_query", "1", &MASCOT_SWITCH
+        config "peptide_master", "1", &MASCOT_SWITCH
+        config "prot_matches", "1", &MASCOT_SWITCH
+        config "_onlyerrortolerant", ""
+        config "_showallfromerrortolerant", ""
+        config "prot_hit_num", "1", &MASCOT_SWITCH
+        config "search_master", "1", &MASCOT_SWITCH
+        config "_sigthreshold", 0.05, &c.num
+        config "show_params", "1", &MASCOT_SWITCH
+        config "show_mods", "1", &MASCOT_SWITCH
+        config "show_header", "1", &MASCOT_SWITCH
+        config "pep_isbold", "1", &MASCOT_SWITCH
+        config "pep_seq", "1", &MASCOT_SWITCH
+        config "pep_exp_z", "1", &MASCOT_SWITCH
+        config "prot_desc", "1", &MASCOT_SWITCH
+        config "_ignoreionsscorebelow", "0", &MASCOT_SWITCH
+        config "REPORT", "AUTO", &c.string
+        config "pep_rank", "1", &MASCOT_SWITCH
+        config "pep_var_mod", "1", &MASCOT_SWITCH
+        config "_noerrortolerant", ""
+      end
       
       def process(result_filepath)
-        # duplicate the configurations
-        request = {}
-        config.each_pair do |key, value|
-          request[key] = value.kind_of?(Hash) ? value.dup : value
-        end
-        
-        # set filename for export
-        request[:params]['file'] = result_filepath
-        
-        super(request)
+        params = config[:params].to_hash
+        params['file'] = result_filepath
+      
+        # submit request
+        super(
+          :request_method => 'GET',
+          :uri => uri,
+          :params => params
+        )
       end
     end 
   end

data/lib/ms/mascot/format_mgf.rb

@@ -8,20 +8,19 @@ module Ms
     # configurations specify various details of the dump, including the
     # precision and default headers.
     #
-    #   % rap fragment TVQQEL --:s format_mgf
+    #   % tap run -- fragment TVQQEL --:s format_mgf
     #
     # (note: be sure to use the splat option on the join)
     #
     class FormatMgf < Tap::Task
       
-      config :default_headers, {}, &c.hash        # a hash of default headers
-      config :min_length, 3, &c.integer_or_nil    # the minimum peptide length
-      config :mz_precision, 6, &c.integer         # the precision of mzs
-      config :intensity_precision, 0, &c.integer  # the precision of intensities
-      config :pepmass_precision, 6, &c.integer    # the precision of peptide mass
+      config :default_headers, {}, &c.hash        # A hash of default headers
+      config :mz_precision, 6, &c.integer         # The precision of mzs
+      config :intensity_precision, 0, &c.integer  # The precision of intensities
+      config :pepmass_precision, 6, &c.integer    # The precision of peptide mass
       
-      config :prefix, nil, &c.string_or_nil       # an optional prefix
-      config :suffix, "\n", &c.string_or_nil      # an optional suffix
+      config :prefix, nil, &c.string_or_nil       # An optional prefix
+      config :suffix, "\n", &c.string_or_nil      # An optional suffix
       
       # Maps header keys (typically output by a fragment task)
       # to Mgf::Entry header strings.

data/lib/ms/mascot/fragment.rb

@@ -30,7 +30,7 @@ module Ms
         {
           :charge => charge,
           :parent_ion_mass => spec.parent_ion_mass(charge),
-          :title => "#{spec.sequence} (#{series.join(', ')})"
+          :title => "#{spec.sequence.gsub(/\s+/, "")} (#{series.join(', ')})"
         }
       end
       

data/lib/ms/mascot/mgf.rb

@@ -1,9 +1,14 @@
 require 'ms/mascot/mgf/entry'
 require 'ms/mascot/mgf/archive'
+require 'set'
 
 module Ms
   module Mascot
     module Mgf
+      # see http://www.matrixscience.com/help/data_file_help.html
+      VALID_LOCAL_HEADERS = Set.new(%w(CHARGE COMP ETAG INSTRUMENT IT_MODS PEPMASS RTINSECONDS SCANS SEQ TAG TITLE TOL TOLU))
+      VALID_GLOBAL_HEADERS = Set.new(%w(ACCESSION CHARGE CLE COM CUTOUT DB DECOY ERRORTOLERANT FORMAT FRAMES INSTRUMENT IT_MODS ITOL ITOLU MASS MODS PEP_ISOTOPE_ERROR PFA PRECURSOR QUANTITATION REPORT REPTYPE SEARCH SEG TAXONOMY TOL TOLU USER00 USER01 USER02 USER03 USER04 USER05 USER06 USER07 USER08 USER09 USER10 USER11 USER12 USEREMAIL USERNAME))
+
       class << self
         # Opens the file and yields an array of entries (well, the array is
         # actually an Ms::Mascot::Mgf::Archive object that acts like an array
@@ -26,10 +31,35 @@ module Ms
         # returns each entry in the mgf file, like IO.foreach
         def foreach(file, &block)
           open(file) do |ar|
-            ar.each &block
+            ar.each( &block )
           end
         end
-      end
+
+        # yields an Ms::Mascot::Mgf::Archive object and writes the data to
+        # outfile.
+        #
+        # example of writing spetra to "out.mgf":
+        #
+        #     Ms::Mascot::Mgf.write("out.mgf") do |mgf|
+        #       # use the Query#to_mgf method:
+        #       mgf << query.to_mgf(peptide_hit)
+        #
+        #       # create your own entry object
+        #       mgf << Ms::Mascot::Dat::Mgf::Entry.new(header, data)
+        #
+        #       # push on the strings
+        #       mgf << "BEGIN IONS"
+        #       mgf << "TITLE=mytitle"
+        #       # ... the rest of the info ...
+        #       mgf << "END IONS"
+        #     end
+        def write(outfile)
+          mgf = Archive.new
+          yield mgf
+          mgf.close(outfile, nil, true)
+        end
+
+      end # end module methods
     end
   end
 end

data/lib/ms/mascot/mgf/archive.rb

@@ -8,6 +8,13 @@ module Ms
       # Provides array-like access to an mgf archival file.
       class Archive < ExternalArchive
 
+        # yields an object for writing
+        def self.write(filename)
+          mgf = self.new
+          File.open(filename, 'w') do |out|
+          end
+        end
+
         # Reindexes self to each mgf entry in io
         def reindex(&block)
           reindex_by_sep("BEGIN IONS", :entry_follows_sep => true, &block)
@@ -21,4 +28,4 @@ module Ms
       end
     end
   end
-end
+end

data/lib/ms/mascot/submit.rb

@@ -1,4 +1,5 @@
-require 'tap/http/submit'
+require 'tap/mechanize/request'
+require 'ms/mascot/validation'
 
 module Ms
   module Mascot
@@ -28,55 +29,8 @@ module Ms
     # values MUST be overridden and are only provided as a template (for
     # those that want the adventure of manually making a config file).
     #
-    class Submit < Tap::Http::Submit
-      
-      # The MatrixScience public search site
-      DEFAULT_URI = "http://www.matrixscience.com/cgi/nph-mascot.exe?1"
-      
-      # Parameters for MS/MS searching of a human sample digested with trypsin.
-      DEFAULT_PARAMS = {
-        "ErrTolRepeat"=>"0",
-        "PFA"=>"1",
-        "INSTRUMENT"=>"Default",
-        "REPTYPE"=>"peptide",
-        "COM"=>"Search Title",
-        "FORMAT"=>"Mascot generic",
-        "PEAK"=>"AUTO",
-        "CHARGE"=>"2+",
-        "INTERMEDIATE"=>"",
-        "SHOWALLMODS"=>"",
-        "PRECURSOR"=>"",
-        "USERNAME"=>"Name",
-        "TOLU"=>"ppm",
-        "USEREMAIL"=>"email@email.com",
-        "CLE"=>"Trypsin",
-        "TOL"=>"100",
-        "ITOLU"=>"Da",
-        "QUANTITATION"=>"None",
-        "SEARCH"=>"MIS",
-        "DB"=>"SwissProt",
-        "PEP_ISOTOPE_ERROR"=>"0",
-        "ITOL"=>"0.6",
-        "FORMVER"=>"1.01",
-        "IT_MODS"=> [
-          "Acetyl (Protein N-term)",
-          "Gln->pyro-Glu (N-term Q)",
-          "Oxidation (M)"],
-        "MASS"=>"Monoisotopic",
-        "REPORT"=>"AUTO",
-        "TAXONOMY"=>". . . . . . . . . . . . . . . . Homo sapiens (human)"
-      }
-      
-      # Typical headers for an MS/MS search.
-      DEFAULT_HEADERS = {
-       "Keep-Alive"=>"300",
-       "Accept-Encoding"=>"gzip,deflate",
-       "Accept-Language"=>"en-us,en;q=0.5",
-       "Content-Type"=> "multipart/form-data; boundary=---------------------------168072824752491622650073",
-       "Accept-Charset"=>"ISO-8859-1,utf-8;q=0.7,*;q=0.7",
-       "Accept"=>"text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8",
-       "Connection"=>"keep-alive"
-      }
+    class Submit < Tap::Mechanize::Request
+      include Validation
       
       # Matches a successful search response.  After the match:
       #
@@ -88,29 +42,58 @@ module Ms
       #  $1:: the failure message
       FAILURE_REGEXP = /<BR>(.*)/m
       
-      config :uri, DEFAULT_URI                           # The uri of the mascot search site
-      config :headers, DEFAULT_HEADERS, &c.hash          # a hash of request headers
-      config :params, DEFAULT_PARAMS, &c.hash            # a hash of query parameters
-      config :request_method, 'POST'                     # the request method (get or post)
-      config :version, 1.1                               # the HTTP version
-      config :redirection_limit, nil, &c.integer_or_nil  # the redirection limit for the request
+      # The MatrixScience public search site
+      config :uri, "http://www.matrixscience.com/cgi/nph-mascot.exe?1"  # The uri of the mascot search site
+      
+      # Parameters for MS/MS searching of a human sample digested with trypsin
+      nest :params do                                                   # The query parameters
+        config "ErrTolRepeat", 0, &MASCOT_SWITCH
+        config "PFA", 1, &MASCOT_SWITCH
+        config "INSTRUMENT", "Default", &c.string
+        config "REPTYPE", "peptide",  &c.string
+        config "COM", "Search Title",  &c.string
+        config "FORMAT", "Mascot generic",  &c.string
+        config "PEAK", "AUTO",  &c.string
+        config "CHARGE", "+2"
+        config "INTERMEDIATE", "",  &c.string
+        config "SHOWALLMODS", "",  &c.string
+        config "PRECURSOR", "",  &c.string
+        config "USERNAME", "Name",  &c.string
+        config "TOLU", "ppm",  &c.string
+        config "USEREMAIL", '', &c.string
+        config "CLE", "Trypsin", &c.string
+        config "TOL", 100, &c.num
+        config "ITOLU", "Da", &c.string
+        config "QUANTITATION", "None", &c.string
+        config "SEARCH", "MIS", &c.string
+        config "DB", "SwissProt", &c.string
+        config "PEP_ISOTOPE_ERROR", 0, &c.num
+        config "ITOL", 0.6, &c.float
+        config "FORMVER", 1.01, &c.float
+        config "IT_MODS", [
+          "Acetyl (Protein N-term)",
+          "Gln->pyro-Glu (N-term Q)",
+          "Oxidation (M)"], &c.list
+        config "MASS", "Monoisotopic", &c.string
+        config "REPORT", "AUTO", &c.string
+        config "TAXONOMY", ". . . . . . . . . . . . . . . . Homo sapiens (human)", &c.string
+      end
       
       def process(mgf_file)
+        File.open(mgf_file) do |io|
+          # set filename for upload
+          params = config[:params].to_hash
+          params['FILE'] = io
         
-        # duplicate the configurations
-        request = {}
-        config.each_pair do |key, value|
-          request[key] = value.kind_of?(Hash) ? value.dup : value
+          # submit request
+          page = super(
+            :request_method => 'POST',
+            :uri => uri,
+            :params => params
+          )
+          
+          parse_response_body(page.body)
         end
-        
-        # set filename for upload
-        file = request[:params]['FILE'] ||= {}
-        file['Filename'] = mgf_file
-        file['Content-Type'] = 'application/octet-stream'
-        file.delete('Content')
-        
-        # submit request
-        parse_response_body super(request)
       end
       
       # Processes the response body.  Returns the result file if the body

data/lib/ms/mascot/validation.rb

@@ -0,0 +1,17 @@
+module Ms
+  module Mascot
+    module Validation
+      MASCOT_SWITCH = lambda do |input|
+        input = case input
+        when true, 1, '1', /true/i   then '1'
+        when false, 0, '0', /false/i then '0'
+        else input
+        end
+        
+        Configurable::Validation::validate(input, ['1', '0'])
+      end
+      
+      Configurable::DEFAULT_ATTRIBUTES[MASCOT_SWITCH] = {:type => :switch}
+    end
+  end
+end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification 
 name: ms-mascot
 version: !ruby/object:Gem::Version 
-  version: 0.2.0
+  version: 0.2.2
 platform: ruby
 authors: 
 - Simon Chiang
@@ -10,7 +10,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2009-02-24 00:00:00 -07:00
+date: 2009-03-31 00:00:00 -06:00
 default_executable: 
 dependencies: 
 - !ruby/object:Gem::Dependency 
@@ -21,17 +21,17 @@ dependencies:
     requirements: 
     - - ">="
       - !ruby/object:Gem::Version 
-        version: "0.12"
+        version: 0.12.4
     version: 
 - !ruby/object:Gem::Dependency 
-  name: tap-http
+  name: tap-mechanize
   type: :runtime
   version_requirement: 
   version_requirements: !ruby/object:Gem::Requirement 
     requirements: 
     - - ">="
       - !ruby/object:Gem::Version 
-        version: 0.3.2
+        version: 0.5.1
     version: 
 - !ruby/object:Gem::Dependency 
   name: external
@@ -51,7 +51,7 @@ dependencies:
     requirements: 
     - - ">="
       - !ruby/object:Gem::Version 
-        version: 0.2.2
+        version: 0.2.3
     version: 
 description: 
 email: simon.a.chiang@gmail.com
@@ -85,6 +85,7 @@ files:
 - lib/ms/mascot/mgf/entry.rb
 - lib/ms/mascot/spectrum.rb
 - lib/ms/mascot/submit.rb
+- lib/ms/mascot/validation.rb
 - tap.yml
 - README
 - MIT-LICENSE