polished pairwise alignment code

[biopieces.git] / code_ruby / lib / maasha / align / pair.rb

diff --git a/code_ruby/lib/maasha/align/pair.rb b/code_ruby/lib/maasha/align/pair.rb
index c17dba60036bb8442e00d8d56e9bfbabfb3395c4..8bcbf25bddad3a8143e44d43a6c3a20c1e6790af 100644 (file)
--- a/code_ruby/lib/maasha/align/pair.rb
+++ b/code_ruby/lib/maasha/align/pair.rb
@@ -22,11 +22,12 @@
 
 # >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>><<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
 
-require 'maasha/align/match'
+require 'maasha/align/matches'
 require 'maasha/math_aux'
 
 FACTOR_SCORE_LENGTH =  1.0
 FACTOR_SCORE_DIAG   = -1.41
+KMER                = 32
 
 # Module with stuff to create a pairwise aligment.
 module PairAlign
@@ -46,10 +47,9 @@ module PairAlign
       @q_entry.seq.downcase!
       @s_entry.seq.downcase!
 
-      space = Space.new( 0, 0, @q_entry.length - 1, @s_entry.length - 1)
-      kmer  = 64
+      space = Space.new(0, 0, @q_entry.length - 1, @s_entry.length - 1)
 
-      align_recurse(@q_entry.seq, @s_entry.seq, space, kmer)
+      align_recurse(@q_entry.seq, @s_entry.seq, space, KMER)
       matches_upcase
       gaps_insert
     end
@@ -69,6 +69,7 @@ module PairAlign
 
       while (matches.size == 0 and kmer > 0)
         matches = Matches.find(q_seq, s_seq, space.q_min, space.s_min, space.q_max, space.s_max, kmer)
+        #matches = Mem.find(q_seq, s_seq, kmer, space.q_min, space.s_min, space.q_max, space.s_max)
 
         if @matches.empty?
           matches.sort_by! { |m| m.length }
@@ -76,7 +77,6 @@ module PairAlign
           matches = matches_select_by_score(matches, space)
         end
 
-#      $stderr.puts "#{kmer}   #{matches.size}   #{space.q_dim} x #{space.s_dim}"
         kmer /= 2
       end
 
@@ -116,10 +116,6 @@ module PairAlign
         score_diag   = match_score_diag(match, space)
 
         match.score = score_length + score_diag
-
-
-#        $stderr.puts "score_length: #{score_length}   score_diag: #{score_diag}   score: #{match.score}"   # DEBUG
-#        $stderr.puts match
       end
 
       matches.sort_by! { |match| match.score }
@@ -165,104 +161,6 @@ module PairAlign
       dist_min * FACTOR_SCORE_DIAG
     end
 
-    # Method that finds all maximally expanded non-redundant matches shared
-    # between two sequences inside a given search space.
-    def matches_find(q_seq, s_seq, q_min, s_min, q_max, s_max, kmer)
-      matches   = []
-      redundant = Hash.new { |h, k| h[k] = [] }
-
-      s_index = index_seq(s_seq, s_min, s_max, kmer)
-
-      q_pos = q_min
-
-      while q_pos <= q_max - kmer + 1
-        q_oligo = q_seq[q_pos ... q_pos + kmer]
-
-        s_index[q_oligo].each do |s_pos|
-          match = Match.new(q_pos, s_pos, kmer)
-
-          unless match_redundant?(redundant, match)
-            match_expand(match, q_seq, s_seq, q_min, s_min, q_max, s_max)
-            matches << match
-
-            match_redundant_add(redundant, match)
-          end
-        end
-
-        q_pos += 1
-      end
-
-      matches
-    end
-
-    # Method that indexes a sequence within a given interval such that the
-    # index contains all oligos of a given kmer size and the positions where
-    # this oligo was located.
-    def index_seq(seq, min, max, kmer)
-      index_hash = Hash.new { |h, k| h[k] = [] }
-
-      pos = min
-
-      while pos <= max - kmer + 1
-        oligo = seq[pos ... pos + kmer]
-        index_hash[oligo] << pos
-
-        pos += 1
-      end
-
-      index_hash
-    end
-
-    # Method to check if a match is redundant.
-    def match_redundant?(redundant, match)
-      redundant[match.q_beg].each do |s_interval|
-        if s_interval.include? match.s_beg and s_interval.include? match.s_end
-          return true
-        end
-      end
-
-      false
-    end
-
-    # Method that adds a match to the redundancy index.
-    def match_redundant_add(redundant, match)
-      (match.q_beg .. match.q_end).each do |q|
-        redundant[q] << (match.s_beg .. match.s_end)
-      end
-    end
-
-    # Method that expands a match as far as possible to the left and right.
-    def match_expand(match, q_seq, s_seq, q_min, s_min, q_max, s_max)
-      match_expand_left(match, q_seq, s_seq, q_min, s_min)
-      match_expand_right(match, q_seq, s_seq, q_max, s_max)
-
-      match
-    end
-
-    # Method that expands a match as far as possible to the left.
-    def match_expand_left(match, q_seq, s_seq, q_min, s_min)
-      while match.q_beg > q_min and
-            match.s_beg > s_min and 
-            q_seq[match.q_beg - 1] == s_seq[match.s_beg - 1]
-        match.q_beg  -= 1
-        match.s_beg  -= 1
-        match.length += 1
-      end
-
-      match
-    end
-
-    # Method that expands a match as far as possible to the right.
-    def match_expand_right(match, q_seq, s_seq, q_max, s_max)
-      while match.q_end < q_max and
-            match.s_end < s_max and
-            q_seq[match.q_end + 1] == s_seq[match.s_end + 1]
-        match.length += 1
-      end
-
-      match
-    end
-
     # Method for debugging purposes that upcase matching sequence while non-matches
     # sequence is kept in lower case.
     def matches_upcase
@@ -340,31 +238,5 @@ module PairAlign
       true
     end
   end
-
-  # Class for containing a match between two sequences q and s.
-  class Match
-    attr_accessor :q_beg, :s_beg, :length, :score
-
-    def initialize(q_beg, s_beg, length, score = 0.0)
-      @q_beg  = q_beg
-      @s_beg  = s_beg
-      @length = length
-      @score  = score
-    end
-
-    def q_end
-      @q_beg + @length - 1
-    end
-
-    def s_end
-      @s_beg + @length - 1
-    end
-
-    def to_s(seq = nil)
-      s = "q: #{@q_beg} #{q_end} s: #{@s_beg} #{s_end} l: #{@length} s: #{@score}"
-      s << " seq: #{seq[@q_beg .. q_end]}" if seq
-      s
-    end
-  end
 end