bix 0.0.1 → 0.0.2

data/lib/bix.rb

@@ -1 +1,8 @@
+# Define it here, will get used elsewhere
+module Bix
+end
+
 require "#{File.dirname(__FILE__)}/bix/fastq"
+require "#{File.dirname(__FILE__)}/bix/fasta"
+require "#{File.dirname(__FILE__)}/bix/blast"
+

data/lib/bix/blast.rb

@@ -0,0 +1,87 @@
+#!/usr/bin/env ruby
+
+require 'swak'
+
+module Bix
+  module Blast
+    class Hit
+      Attribs =  [:query, :subject, :perc_id, :align_len, :mismatches, 
+                    :gap_opens, :query_start, :query_end, :subject_start,
+                    :subject_end, :e_value, :bit_score]
+
+      attr_accessor *Attribs
+
+      def initialize(line=nil)
+        if line != nil && line[0] != '#'
+          f = line.chomp.split("\t", -1)
+
+          raise "Illegal number of fields for blast hit" unless f.size == 12
+
+          @query = f[0]
+          @subject = f[1]
+          @perc_id = f[2].to_f_strict
+          @align_len = f[3].to_i_strict
+          @mismatches = f[4].to_i_strict
+          @gap_opens = f[5].to_i_strict
+          @query_start = f[6].to_i_strict
+          @query_end = f[7].to_i_strict
+          @subject_start = f[8].to_i_strict
+          @subject_end = f[9].to_i_strict
+          @e_value = f[10].to_f_strict
+          @bit_score = f[11].to_f_strict
+        end
+      end
+
+      def to_arr
+        return Attribs.map{|a| self.send(a)}
+      end
+    end
+
+    class QueryHits
+      attr_accessor :query, :hits, :best_hit, :pref_min_aln_len
+
+      # Suppose we want the best_hit to point to the best e-value of all hits over a certain aln len
+      # If there are no hits over that length, then we want the best e-value
+      def initialize(hits, pref_min_aln_len=0)
+        @pref_min_aln_len = pref_min_aln_len
+        @hits = hits
+
+        raise "QueryHits error: must have > 0 hits"  unless hits.size > 0
+
+        @query = hits.first.query
+        @best_hit = hits.first
+
+        for h in hits
+          raise "QueryHits error: not all hits have same query!" if h.query != query
+
+          if @best_hit.align_len < pref_min_aln_len && h.align_len >= pref_min_aln_len
+            @best_hit = h
+          elsif h.e_value < @best_hit.e_value
+            @best_hit = h
+          end
+        end
+      end
+    end
+
+    def self.get_all_query_hits(io, pref_min_aln_len=0)
+      hits_by_query = {}
+
+      # Read all blast hits into memory, group them by query name
+      for line in io
+        next if line[0] == '#'
+        h = Hit.new(line)
+        hits_by_query[h.query] ||= []
+        hits_by_query[h.query] << h
+      end
+
+      # Build vector of QueryHits objects
+      query_hits = []
+
+      for query, hits in hits_by_query
+        query_hits << QueryHits.new(hits, pref_min_aln_len)
+      end
+
+      return query_hits
+    end
+  end
+end

data/lib/bix/fasta.rb

@@ -0,0 +1,29 @@
+module Bix
+  class Fasta
+    attr_accessor :header, :seq
+  end
+
+  def self.read_fastas(io)
+    fastas = []
+
+    fa = nil
+
+    for line in io
+      line.chomp!
+      if line[0] == '>'
+        fastas << fa if fa != nil
+
+        fa = Bix::Fasta.new
+        fa.header = line
+        fa.seq = ""
+      else
+        fa.seq << line
+      end
+
+    end
+
+    fastas << fa if fa != nil
+
+    return fastas
+  end
+end

data/lib/bix/fastq.rb

@@ -1,28 +1,29 @@
-class Fastq
-  attr_accessor :header, :seq, :qual, :invalid
+module Bix
+  class Fastq
+    attr_accessor :header, :seq, :qual, :invalid
 
-  def initialize(header="", seq="", qual="")
-    @header = header
-    @seq = seq
-    @qual = qual
-  end
+    def initialize(header="", seq="", qual="")
+      @header = header
+      @seq = seq
+      @qual = qual
+    end
 
-  def from_io(io)
-    @header = io.gets
-    @seq = io.gets
-    io.gets #plus line
-    @qual = io.gets
+    def from_io(io)
+      @header = io.gets
+      @seq = io.gets
+      io.gets #plus line
+      @qual = io.gets
 
-    return false if @header == nil || @seq == nil || @qual == nil
-    return true
-  end
+      return false if @header == nil || @seq == nil || @qual == nil
+      return true
+    end
 
-  def qual_val(i)
-    return @qual[i].ord
-  end
+    def qual_val(i)
+      return @qual[i].ord
+    end
 
-  def to_lines
-    return "#{@header}\n#{@seq}\n+\n#{@qual}"
+    def to_lines
+      return "#{@header}\n#{@seq}\n+\n#{@qual}"
+    end
   end
-
 end

metadata

@@ -2,7 +2,7 @@
 name: bix
 version: !ruby/object:Gem::Version 
   prerelease: 
-  version: 0.0.1
+  version: 0.0.2
 platform: ruby
 authors: 
 - Jesse Rodriguez
@@ -10,9 +10,41 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2011-08-09 00:00:00 Z
-dependencies: []
-
+date: 2011-09-01 00:00:00 Z
+dependencies: 
+- !ruby/object:Gem::Dependency 
+  name: swak
+  prerelease: false
+  requirement: &id001 !ruby/object:Gem::Requirement 
+    none: false
+    requirements: 
+    - - ">="
+      - !ruby/object:Gem::Version 
+        version: "0"
+  type: :runtime
+  version_requirements: *id001
+- !ruby/object:Gem::Dependency 
+  name: vcf
+  prerelease: false
+  requirement: &id002 !ruby/object:Gem::Requirement 
+    none: false
+    requirements: 
+    - - ">="
+      - !ruby/object:Gem::Version 
+        version: "0"
+  type: :runtime
+  version_requirements: *id002
+- !ruby/object:Gem::Dependency 
+  name: sam
+  prerelease: false
+  requirement: &id003 !ruby/object:Gem::Requirement 
+    none: false
+    requirements: 
+    - - ">="
+      - !ruby/object:Gem::Version 
+        version: "0"
+  type: :runtime
+  version_requirements: *id003
 description: Lightweight bioinformatics tools for Ruby, with a focus on next-gen sequencing
 email: jesserod@cs.stanford.edu
 executables: []
@@ -23,8 +55,10 @@ extra_rdoc_files: []
 
 files: 
 - lib/bix.rb
+- lib/bix/fasta.rb
 - lib/bix/fastq.rb
-homepage: http://rubygems.org/gems/bix
+- lib/bix/blast.rb
+homepage: https://github.com/jesserod/bixruby
 licenses: []
 
 post_install_message: 
@@ -47,7 +81,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
 requirements: []
 
 rubyforge_project: 
-rubygems_version: 1.8.5
+rubygems_version: 1.8.10
 signing_key: 
 specification_version: 3
 summary: Lightweight bioinformatics tools for Ruby