ngoto-bio 1.2.9.9001

data/lib/bio/appl/blast/rexml.rb

@@ -0,0 +1,144 @@
+#
+# = bio/appl/blast/rexml.rb - BLAST XML output (-m 7) parser by REXML
+# 
+# Copyright::  Copyright (C) 2002, 2003 Toshiaki Katayama <k@bioruby.org>
+# License::    The Ruby License
+#
+# $Id:$
+#
+# == Note
+# 
+# This file is automatically loaded by bio/appl/blast/report.rb
+# 
+
+begin
+  require 'rexml/document'
+rescue LoadError
+end
+
+module Bio
+  class Blast
+    class Report
+
+      private
+
+      def rexml_parse(xml)
+        dom = REXML::Document.new(xml)
+        rexml_parse_program(dom)
+        dom.elements.each("*//Iteration") do |e|
+          @iterations.push(rexml_parse_iteration(e))
+        end
+      end
+
+      def rexml_parse_program(dom)
+        hash = {}
+        dom.root.each_element_with_text do |e|
+          name, text = e.name, e.text
+          case name
+          when 'BlastOutput_param'
+            e.elements["Parameters"].each_element_with_text do |p|
+              xml_set_parameter(p.name, p.text)
+            end
+          else
+            hash[name] = text if text.strip.size > 0
+          end
+        end
+        @program	= hash['BlastOutput_program']
+        @version	= hash['BlastOutput_version']
+        @reference	= hash['BlastOutput_reference']
+        @db		= hash['BlastOutput_db']
+        @query_id	= hash['BlastOutput_query-ID']
+        @query_def	= hash['BlastOutput_query-def']
+        @query_len	= hash['BlastOutput_query-len'].to_i
+      end
+
+      def rexml_parse_iteration(e)
+        iteration = Iteration.new
+        e.elements.each do |i|
+          case i.name
+          when 'Iteration_iter-num'
+            iteration.num = i.text.to_i
+          when 'Iteration_hits'
+            i.elements.each("Hit") do |h|
+              iteration.hits.push(rexml_parse_hit(h))
+            end
+          when 'Iteration_message'
+            iteration.message = i.text
+          when 'Iteration_stat'
+            i.elements["Statistics"].each_element_with_text do |s|
+              k = s.name.sub(/Statistics_/, '')
+              v = s.text =~ /\D/ ? s.text.to_f : s.text.to_i
+              iteration.statistics[k] = v
+            end
+
+          # for new BLAST XML format
+          when 'Iteration_query-ID'
+            iteration.query_id = i.text
+          when 'Iteration_query-def'
+            iteration.query_def = i.text
+          when 'Iteration_query-len'
+            iteration.query_len = i.text.to_i
+          end
+        end #case i.name
+
+        return iteration
+      end
+
+      def rexml_parse_hit(e)
+        hit = Hit.new
+        hash = {}
+        hit.query_id = @query_id
+        hit.query_def = @query_def
+        hit.query_len = @query_len
+        e.elements.each do |h|
+          case h.name
+          when 'Hit_hsps'
+            h.elements.each("Hsp") do |s|
+              hit.hsps.push(rexml_parse_hsp(s))
+            end
+          else
+            hash[h.name] = h.text
+          end
+        end
+        hit.num		= hash['Hit_num'].to_i
+        hit.hit_id	= hash['Hit_id']
+        hit.len		= hash['Hit_len'].to_i
+        hit.definition	= hash['Hit_def']
+        hit.accession	= hash['Hit_accession']
+        return hit
+      end
+
+      def rexml_parse_hsp(e)
+        hsp = Hsp.new
+        hash = {}
+        e.each_element_with_text do |h|
+          hash[h.name] = h.text
+        end
+        hsp.num			= hash['Hsp_num'].to_i
+        hsp.bit_score		= hash['Hsp_bit-score'].to_f
+        hsp.score		= hash['Hsp_score'].to_i
+        hsp.evalue		= hash['Hsp_evalue'].to_f
+        hsp.query_from		= hash['Hsp_query-from'].to_i
+        hsp.query_to		= hash['Hsp_query-to'].to_i
+        hsp.hit_from		= hash['Hsp_hit-from'].to_i
+        hsp.hit_to		= hash['Hsp_hit-to'].to_i
+        hsp.pattern_from	= hash['Hsp_pattern-from'].to_i
+        hsp.pattern_to		= hash['Hsp_pattern-to'].to_i
+        hsp.query_frame		= hash['Hsp_query-frame'].to_i
+        hsp.hit_frame		= hash['Hsp_hit-frame'].to_i
+        hsp.identity		= hash['Hsp_identity'].to_i
+        hsp.positive		= hash['Hsp_positive'].to_i
+        hsp.gaps		= hash['Hsp_gaps'].to_i
+        hsp.align_len		= hash['Hsp_align-len'].to_i
+        hsp.density		= hash['Hsp_density'].to_i
+        hsp.qseq		= hash['Hsp_qseq']
+        hsp.hseq		= hash['Hsp_hseq']
+        hsp.midline		= hash['Hsp_midline']
+        return hsp
+      end
+
+    end
+  end
+end
+
+

data/lib/bio/appl/blast/rpsblast.rb

@@ -0,0 +1,277 @@
+#
+# = bio/appl/blast/rpsblast.rb - NCBI RPS Blast default output parser
+# 
+# Copyright::  Copyright (C) 2008 Naohisa Goto <ng@bioruby.org>
+# License::    The Ruby License
+#
+# $Id:$
+#
+# == Description
+#
+# NCBI RPS Blast (Reversed Position Specific Blast) default
+# (-m 0 option) output parser class, Bio::Blast::RPSBlast::Report
+# and related classes/modules.
+#
+# == References
+#
+# * Altschul, Stephen F., Thomas L. Madden, Alejandro A. Schaffer,
+#   Jinghui Zhang, Zheng Zhang, Webb Miller, and David J. Lipman (1997),
+#   "Gapped BLAST and PSI-BLAST: a new generation of protein database search
+#   programs", Nucleic Acids Res. 25:3389-3402.
+# * ftp://ftp.ncbi.nih.gov/blast/documents/rpsblast.html
+# * http://www.ncbi.nlm.nih.gov/Structure/cdd/cdd_help.shtml
+#
+
+require 'bio/appl/blast/format0'
+require 'bio/io/flatfile'
+
+module Bio
+class Blast
+
+  # NCBI RPS Blast (Reversed Position Specific Blast) namespace.
+  # Currently, this module is existing only for separating namespace.
+  # To parse RPSBlast results, see Bio::Blast::RPSBlast::Report documents.
+  module RPSBlast
+
+    # Flatfile splitter for RPS-BLAST reports.
+    # It is internally used when reading RPS-BLAST report.
+    # Normally, users do not need to use it directly.
+    #
+    # Note for Windows: RPS-BLAST results generated in Microsoft Windows
+    # may not be parsed correctly due to the line feed code problem.
+    # For a workaroud, convert line feed codes from Windows(DOS) to UNIX.
+    #
+    class RPSBlastSplitter < Bio::FlatFile::Splitter::Template
+
+      # Separator used to distinguish start of each report
+      ReportHead =  /\A\n*(RPS\-BLAST|Query\=)/
+        
+      # Delimiter used for IO#gets
+      Delimiter = "\n\n"
+
+      # creates a new splitter object
+      def initialize(klass, bstream)
+        super(klass, bstream)
+        @entry_head = nil
+      end
+
+      # Skips leader of the entry.
+      # In this class, only skips space characters.
+      def skip_leader
+        stream.skip_spaces
+        return nil
+      end
+
+      # Rewinds the stream
+      def rewind
+        @entry_head = nil
+        super
+      end
+        
+      # gets an entry
+      def get_entry
+        p0 = stream_pos()
+        pieces = []
+        flag_head = false # reached to start of header
+        flag_body = false # reached to start of body (Query=...)
+        while x = stream.gets(Delimiter)
+          if ReportHead =~ x then
+            case $1
+            when 'RPS-BLAST'
+              if pieces.empty? then
+                @entry_head = nil
+                flag_head = true
+              else
+                stream.ungets(x)
+                break
+              end
+            when 'Query='
+              if flag_body then
+                stream.ungets(x)
+                break
+              else
+                @entry_head = pieces.join('') if flag_head
+                flag_body = true
+              end
+            else
+              raise 'Bug: should not reach here'
+            end
+          end #if ReportHead...
+          pieces.push x
+        end #while
+        p1 = stream_pos()
+
+        self.entry_start_pos = p0
+        self.entry = 
+          if pieces.empty? then
+            nil
+          elsif !flag_head and @entry_head then
+            @entry_head + pieces.join('')
+          else
+            pieces.join('')
+          end
+        self.entry_ended_pos = p1
+        return self.entry
+      end
+    end #class RPSBlastSplitter
+
+    # NCBI RPS Blast (Reversed Position Specific Blast)
+    # default output parser.
+    #
+    # It supports defalut (-m 0 option) output of the "rpsblast" command.
+    #
+    # Because this class inherits Bio::Blast::Default::Report,
+    # almost all methods are eqaul to Bio::Blast::Default::Report.
+    # Only DELIMITER (and RS) and few methods are different.
+    #
+    # By using Bio::FlatFile, (for example, Bio::FlatFile.open),
+    # rpsblast result generated from multiple query sequences is
+    # automatically splitted into multiple 
+    # Bio::BLast::RPSBlast::Report objects corresponding to
+    # query sequences.
+    #
+    # Note for multi-fasta results WITH using Bio::FlatFile:
+    # Each splitted result is concatenated with header of the
+    # result which describes RPS-BLAST version and database
+    # information, if possible.
+    #
+    # Note for multi-fasta results WITHOUT using Bio::FlatFile:
+    # When parsing an output of rpsblast command running with
+    # multi-fasta sequences WITHOUT using Bio::FlatFile,
+    # each query's result is stored as an "iteration" of PSI-Blast.
+    # This behavior may be changed in the future.
+    #
+    # Note for nucleotide results: This class is not tested with
+    # nucleotide query and/or nucleotide databases.
+    #
+    class Report < Bio::Blast::Default::Report
+      # Delimter of each entry for RPS-BLAST.
+      DELIMITER = RS = "\nRPS-BLAST"
+
+      # (Integer) excess read size included in DELIMITER.
+      DELIMITER_OVERRUN = 9 # "RPS-BLAST"
+
+      # splitter for Bio::FlatFile support
+      FLATFILE_SPLITTER = RPSBlastSplitter
+
+      # Creates a new Report object from a string.
+      #
+      # Using Bio::FlatFile.open (or some other methods)
+      # is recommended instead of using this method directly.
+      # Refer Bio::Blast::RPSBlast::Report document for more information.
+      #
+      # Note for multi-fasta results WITHOUT using Bio::FlatFile:
+      # When parsing an output of rpsblast command running with
+      # multi-fasta sequences WITHOUT using Bio::FlatFile,
+      # each query's result is stored as an "iteration" of PSI-Blast.
+      # This behavior may be changed in the future.
+      #
+      # Note for nucleotide results: This class is not tested with
+      # nucleotide query and/or nucleotide databases.
+      #
+      def initialize(str)
+        str = str.sub(/\A\s+/, '')
+        # remove trailing entries for sure
+        str.sub!(/\n(RPS\-BLAST.*)/m, "\n") 
+        @entry_overrun = $1
+        @entry = str
+        data = str.split(/(?:^[ \t]*\n)+/)
+
+        if data[0] and /\AQuery\=/ !~ data[0] then
+          format0_split_headers(data)
+        end
+        @iterations = format0_split_search(data)
+        format0_split_stat_params(data)
+      end
+
+      # Returns definition of the query.
+      # For a result of multi-fasta input, the first query's definition
+      # is returned (The same as <tt>iterations.first.query_def</tt>).
+      def query_def
+        iterations.first.query_def
+      end
+
+      # Returns length of the query.
+      # For a result of multi-fasta input, the first query's length
+      # is returned (The same as <tt>iterations.first.query_len</tt>).
+      def query_len
+        iterations.first.query_len
+      end
+
+      private
+
+      # Splits headers into the first line, reference, query line and
+      # database line.
+      def format0_split_headers(data)
+        @f0header = data.shift
+        @f0references = []
+        while data[0] and /\ADatabase\:/ !~ data[0]
+          @f0references.push data.shift
+        end
+        @f0database = data.shift
+        # In special case, a void line is inserted after database name.
+        if /\A +[\d\,]+ +sequences\; +[\d\,]+ total +letters\s*\z/ =~ data[0] then
+          @f0database.concat "\n"
+          @f0database.concat data.shift
+        end
+      end
+
+      # Splits the search results.
+      def format0_split_search(data)
+        iterations = []
+        dummystr = 'Searching..................................................done'
+        if r = data[0] and /^Searching/ =~ r then
+          dummystr = data.shift
+        end
+        while r = data[0] and /^Query\=/ =~ r
+          iterations << Iteration.new(data, dummystr)
+        end
+        iterations
+      end
+
+      # Iteration class for RPS-Blast.
+      # Though RPS-Blast does not iterate like PSI-BLAST, 
+      # it aims to store a result of single query sequence.
+      #
+      # Normally, the instance of the class is generated
+      # by Bio::Blast::RPSBlast::Report object.
+      # 
+      class Iteration < Bio::Blast::Default::Report::Iteration
+        # Creates a new Iteration object.
+        # It is designed to be called only internally from
+        # the Bio::Blast::RPSBlast::Report class.
+        # Users shall not use the method directly.
+        def initialize(data, dummystr)
+          if /\AQuery\=/ =~ data[0] then
+            sc = StringScanner.new(data.shift)
+            sc.skip(/\s*/)
+            if sc.skip_until(/Query\= */) then
+              q = []
+              begin
+                q << sc.scan(/.*/)
+                sc.skip(/\s*^ ?/)
+              end until !sc.rest or r = sc.skip(/ *\( *([\,\d]+) *letters *\)\s*\z/)
+              @query_len = sc[1].delete(',').to_i if r
+              @query_def = q.join(' ')
+            end
+          end
+          data.unshift(dummystr)
+          
+          super(data)
+        end
+
+        # definition of the query
+        attr_reader :query_def
+
+        # length of the query sequence
+        attr_reader :query_len
+        
+      end #class Iteration
+      
+    end #class Report
+
+  end #module RPSBlast
+
+end #module Blast
+end #module Bio
+