scbi_fastq 0.0.14 → 0.0.15

data/History.txt

@@ -1,3 +1,7 @@
+=== 0.0.15 2012-05-24
+
+Uses default qual when no qual is present
+
 === 0.0.14 2011-05-31
 
 release to rubygems

data/Rakefile

@@ -23,4 +23,4 @@ Dir['tasks/**/*.rake'].each { |t| load t }
 
 # TODO - want other tests/tasks run by default? Add them to the list
 # remove_task :default
-# task :default => [:spec, :features]
+task :default => [:spec, :features, :redocs]

data/lib/scbi_fastq/fastq_file.rb

@@ -1,15 +1,15 @@
 
 # add ord method to ruby 1.8
 if !String.instance_methods.include?(:ord)
-   class String 
-      
-     def ord
-       return self[0]
-     end
-     
-   end
+  class String
+
+    def ord
+      return self[0]
+    end
+
+  end
 end
-      
+
 
 
 class FastqFile
@@ -21,31 +21,31 @@ class FastqFile
   #------------------------------------
   def initialize(fasta_file_name, mode='r', fastq_type = :sanger, qual_to_array=true, qual_to_phred=true)
 
-  	
+
     if mode.upcase.index('W')
       @fastq_file = File.open(fasta_file_name,'w')
     elsif mode.upcase.index('A')
       if !File.exist?(fasta_file_name)
-      	raise "File #{fasta_file_name} doesn't exists" 
+        raise "File #{fasta_file_name} doesn't exists"
       end
-    	
+
       @fastq_file = File.open(fasta_file_name,'a')
     else #read only
       if !File.exist?(fasta_file_name)
-      	raise "File #{fasta_file_name} doesn't exists" 
+        raise "File #{fasta_file_name} doesn't exists"
       end
-    	
+
       if fasta_file_name.is_a?(IO)
-    	  @fastq_file = fasta_file_name
+        @fastq_file = fasta_file_name
       else
         @fastq_file = File.open(fasta_file_name,'r')
       end
     end
-    
+
     @mode = mode
     @num_seqs = 0
     @fastq_type=fastq_type
-    
+
     #  S - Sanger        Phred+33,  raw reads typically (0, 40)
     #  X - Solexa        Solexa+64, raw reads typically (-5, 40)
     #  I - Illumina 1.3+ Phred+64,  raw reads typically (0, 40)
@@ -63,7 +63,7 @@ class FastqFile
     # > -5.8682532438011537
     # > >>> solexa_quality_from_phred(0.1)
     # > -16.32774717238372
-    # > 
+    # >
     # > >>> def phred_quality_from_solexa(solexa_quality) :
     # > ...     return 10*log(10**(solexa_quality/10.0) + 1, 10)
     # > ...
@@ -75,62 +75,62 @@ class FastqFile
     # > 3.0102999566398116
     # > >>> phred_quality_from_solexa(-20)
     # > 0.043213737826425784
-    
-    
+
+
     #sanger by default
     @to_phred = lambda{|q| q - 33}
     @from_phred = lambda{|q| (q+33).chr}
-    
+
     if @fastq_type == :ilumina
-        @to_phred = lambda{|q| q - 64}
-        # @from_phred = lambda{|q| (q+64).chr}
-        
+      @to_phred = lambda{|q| q - 64}
+      # @from_phred = lambda{|q| (q+64).chr}
+
     elsif @fastq_type == :solexa
-       # 
-       # solexa to phred quals
-       
-       @to_phred = lambda{|q| (10*Math.log(10**(q/10.0)+1,10)).round}
-       # @from_phred = lambda{|q| (10*Math.log(10**(q/10.0)-1,10)).round.chr}
-       
-       #phred to solexa quals
-       
+      #
+      # solexa to phred quals
+
+      @to_phred = lambda{|q| (10*Math.log(10**(q/10.0)+1,10)).round}
+      # @from_phred = lambda{|q| (10*Math.log(10**(q/10.0)-1,10)).round.chr}
+
+      #phred to solexa quals
+
     end
-    
+
     @qual_to_array = qual_to_array
-    
+
     @qual_to_phred = qual_to_phred
-    
+
   end
-  
+
   def close
-      @fastq_file.close
+    @fastq_file.close
   end
- 
-  
+
+
   #------------------------------------
   # Iterate over all sequences
   #------------------------------------
   def each
-        
+
     rewind
 
-	  n,f,q,c=next_seq
-	  
+    n,f,q,c=next_seq
+
     while (!n.nil?)
-			yield(n,f,q,c)
-			n,f,q,c=next_seq
+      yield(n,f,q,c)
+      n,f,q,c=next_seq
     end
 
-  	rewind
-  	
+    rewind
+
   end
 
   # goto first position in file
   def rewind
-     
-     @num_seqs = 0 ;
-     @fastq_file.pos=0
-    
+
+    @num_seqs = 0 ;
+    @fastq_file.pos=0
+
   end
 
   #------------------------------------
@@ -139,114 +139,129 @@ class FastqFile
   def next_seq
     #init variables
     res = read_fastq
-  	return res
+    return res
   end
-  
+
   # write sequence to file in sanger format
   def write_seq(seq_name,seq_fasta,seq_qual,comments='')
-	  name = ""
-	  
-		@fastq_file.puts("@#{seq_name} #{comments}")
-		@fastq_file.puts(seq_fasta)
-		@fastq_file.puts("+#{seq_name} #{comments}")
-		
-		if seq_qual.is_a?(Array)
-		  @fastq_file.puts(seq_qual.map{|e| @from_phred.call(e)}.join)
-	  else
-		  @fastq_file.puts(seq_qual.split(/\s+/).map{|e| @from_phred.call(e.to_i)}.join)
-		end
-    
+    name = ""
+
+    @fastq_file.puts("@#{seq_name} #{comments}")
+    @fastq_file.puts(seq_fasta)
+    @fastq_file.puts("+#{seq_name} #{comments}")
+
+    if seq_qual.is_a?(Array)
+      @fastq_file.puts(seq_qual.map{|e| @from_phred.call(e)}.join)
+    else
+      @fastq_file.puts(seq_qual.split(/\s+/).map{|e| @from_phred.call(e.to_i)}.join)
+    end
+
   end
 
-  
+
   # creates fastq otuput in sanger format
   def self.to_fastq(seq_name,seq_fasta,seq_qual,comments='')
-    
+
     res=[]
-    
-	  name = ""
-	  
-		res << ("@#{seq_name} #{comments}")
-		res << (seq_fasta)
-		res << ("+#{seq_name} #{comments}")
-		
-		if seq_qual.is_a?(Array)
-		  res<<(seq_qual.map{|e| (e+33).chr}.join)
-	  else
-		  res<<(seq_qual.split(/\s+/).map{|e| (e.to_i+33).chr}.join)
-		end
-		
+
+    name = ""
+
+    res << ("@#{seq_name} #{comments}")
+    res << (seq_fasta)
+    res << ("+#{seq_name} #{comments}")
+
+    if !seq_qual.empty?
+      if @qual_to_phred
+        if seq_qual.is_a?(Array)
+          res<<(seq_qual.map{|e| (e+33).chr}.join)
+        else
+          res<<(seq_qual.split(/\s+/).map{|e| (e.to_i+33).chr}.join)
+        end
+      else
+        res << seq_qual
+      end
+    else # no qual provided, use a default value
+      q='D'*seq_fasta.length;
+      res << q
+    end
+
     return res
   end
-  
+
   def with_qual?
     true
   end
-  
-  
-  private 
-  
-  #------------------------------------
-  #  Read one sequence in fastq
-  #------------------------------------
-  # @GEM-108-D02
-  # AAAAGCTGG
-  # +
-  # :::::::::
-
-  def read_fastq
-
-    seq_name = nil
-    seq_fasta = nil
-    seq_qual = nil
-    comments = nil
-    
-    reading = :fasta
-    
-    if !@fastq_file.eof
-      
-      begin
-        #read four lines
-        name_line = @fastq_file.readline.chomp
-        seq_fasta = @fastq_file.readline.chomp
-        name2_line = @fastq_file.readline.chomp
-        seq_qual = @fastq_file.readline.chomp
-
-      
-        # parse name
-        if name_line =~ /^@\s*([^\s]+)\s*(.*)$/
-          # remove comments
-          seq_name = $1
-          comments=$2
-        else
-          raise "Invalid sequence name in #{name_line}"
-        end
-      
-        # parse fasta
-        seq_fasta.strip! if !seq_fasta.empty?
-
-        # parse qual_name
-    
-        if !seq_name.nil? && !seq_qual.empty?
-
-           @num_seqs += 1
-       
-           if @qual_to_phred
-             seq_qual=seq_qual.each_char.map{|e| (@to_phred.call(e.ord))}
-
-             if !@qual_to_array
-                 seq_qual=seq_qual.join(' ')
-             end
-           end
-       
+
+
+  private
+
+    #------------------------------------
+    #  Read one sequence in fastq
+    #------------------------------------
+    # @GEM-108-D02
+    # AAAAGCTGG
+    # +
+    # :::::::::
+
+    def read_fastq
+
+      seq_name = nil
+      seq_fasta = nil
+      seq_qual = nil
+      comments = nil
+
+      reading = :fasta
+
+      if !@fastq_file.eof
+
+        begin
+          #read four lines
+          name_line = @fastq_file.readline.chomp
+          seq_fasta = @fastq_file.readline.chomp
+          name2_line = @fastq_file.readline.chomp
+          seq_qual = @fastq_file.readline.chomp
+          
+          
+          # if there is no qual, but there is a fasta
+          if seq_qual.empty? && !seq_fasta.empty?
+            seq_qual = 'D'*seq_fasta.length
+          end
+
+
+          # parse name
+          if name_line =~ /^@\s*([^\s]+)\s*(.*)$/
+            # remove comments
+            seq_name = $1
+            comments=$2
+          else
+            raise "Invalid sequence name in #{name_line}"
+          end
+
+          # parse fasta
+          seq_fasta.strip! if !seq_fasta.empty?
+
+          # parse qual_name
+
+          if !seq_name.nil? && !seq_qual.empty?
+
+            @num_seqs += 1
+
+            if @qual_to_phred
+              seq_qual=seq_qual.each_char.map{|e| (@to_phred.call(e.ord))}
+
+              if !@qual_to_array
+                seq_qual=seq_qual.join(' ')
+              end
+            end
+
+          end
+        rescue EOFError
+          raise "Bad format in FastQ file"
         end
-      rescue EOFError
-        raise "Bad format in FastQ file"
       end
+
+      return [seq_name,seq_fasta,seq_qual,comments]
     end
-    
-    return [seq_name,seq_fasta,seq_qual,comments]
-  end
-  
-  
+
+
 end

data/lib/scbi_fastq.rb

@@ -3,5 +3,5 @@ $:.unshift(File.dirname(__FILE__)) unless
 
 require 'scbi_fastq/fastq_file'
 module ScbiFastq
-   VERSION = '0.0.14'
+   VERSION = '0.0.15'
 end

data/test/test_scbi_fastq.rb

@@ -29,6 +29,24 @@ class TestScbiFastq < Test::Unit::TestCase
   	f.close
   end
 
+  def fill_file_no_qual(n,offset=33)
+  	f=FastqFile.new(@test_file,'w')
+  	
+  	n.times do |c|
+  	  i = c+1
+  	  
+  	  name = "#{@seq_name+i.to_s}"
+  	  f.write_seq(name,@seq_fasta*i,'','comments')
+      # f.puts('@'+name)
+      # f.puts(@seq_fasta*i)
+      # f.puts('+'+name)
+      # f.puts((@seq_qual*i*@seq_fasta.length).map{|e| (e+offset).chr}.join)
+  	end
+  	
+  	f.close
+  end
+
+
   def test_each
 
     	 # make new file and fill with data
@@ -77,7 +95,7 @@ class TestScbiFastq < Test::Unit::TestCase
   def test_next_seq_comments
 
     	 # make new file and fill with data
-		  fill_file(100) 	
+		  fill_file(100)
 		
 
 			fqr=FastqFile.new(@test_file)
@@ -97,8 +115,36 @@ class TestScbiFastq < Test::Unit::TestCase
 			  end
 			end until n.nil?
 			 
+		  fqr.close
+  end
+  
+  def test_to_fastq
+    puts FastqFile.to_fastq(@seq_name,@seq_fasta*10,'','')
+    
+  end
+  
+  def test_each_no_qual
+
+    	 # make new file and fill with data
+		  fill_file_no_qual(100)
+		
+
+			fqr=FastqFile.new(@test_file,'r',:sanger, false,false)
+
+			i=1
+			
+			fqr.each do |n,s,q|
+        puts n,s,q
+        assert_equal(@seq_name+i.to_s,n)
+        assert_equal(@seq_fasta*i,s)
+        # assert_equal((@seq_qual*i*@seq_fasta.length),q)
+
+				i+=1
+			end
+			 
 		  fqr.close			
   end
+  
 
 
   

metadata

@@ -2,7 +2,7 @@
 name: scbi_fastq
 version: !ruby/object:Gem::Version 
   prerelease: 
-  version: 0.0.14
+  version: 0.0.15
 platform: ruby
 authors: 
 - Dario Guerrero
@@ -10,7 +10,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2011-05-31 00:00:00 Z
+date: 2012-05-24 00:00:00 Z
 dependencies: 
 - !ruby/object:Gem::Dependency 
   name: hoe