bio-samtools 0.3.2 → 0.4.0

data/Gemfile

@@ -14,4 +14,5 @@ group :development do
   gem "rcov", ">= 0"
   gem "bio", ">= 1.4.2"
   gem "ffi"
+  gem 'rdoc'
 end

data/Gemfile.lock

@@ -9,9 +9,12 @@ GEM
       bundler (~> 1.0.0)
       git (>= 1.2.5)
       rake
-    rake (0.9.2)
+    json (1.6.1)
+    rake (0.9.2.2)
     rcov (0.9.9)
     rcov (0.9.9-java)
+    rdoc (3.11)
+      json (~> 1.4)
     shoulda (2.11.3)
 
 PLATFORMS
@@ -24,4 +27,5 @@ DEPENDENCIES
   ffi
   jeweler
   rcov
+  rdoc
   shoulda

data/README.rdoc

@@ -16,37 +16,44 @@ the gee_fu genome browser (http://github.com/danmaclean/gee_fu).
 'gem install bio-samtools'
 
 == Usage
-The easiest way to see bio-samtools in action is to try the tests
-rake basic_test 
-rake pileup_test
+See doc/tutorial.html or doc/tutorial.pdf for a walkthrough tutorial. The documentation in doc/index.html provides information on individual objects
+and methods.
 
-See doc/tutorial.html or doc/tutorial.pdf for a walkthrough tutorial
+The easiest way to see bio-samtools in action is to try the tests:
+rake test 
+is all you should need to get the tests to run.
 
+However on some systems (at least the one belonging to one of our developers on Mac OS X Lion, with the Apple provided Ruby) then you can get
+a LoadError: no such file to load 'ffi', despite it being installed. This is something to do with RubyGems and rake test that we
+haven't got to the bottom of yet. The workaround is to run the tests manually. 
+1) change into the directory bio-samtools installed into (you can find this with the command 'gem which bio-samtools')
+2) run the tests as if they were any other ruby script eg ruby test/test_pileup.rb
 
 == Dependencies:
 -FFI (http://github.com/ffi/ffi)
--libbam
-	This can be obtained from samtools.  (http://samtools.sourceforge.net/ )
+
 	
 == FAQ.
 I´m getting a segmentation Fault, what did I do wrong?
-There are two known segmentation faults at the moment
+-There are two known segmentation faults at the moment
 	-When you try to load a text file as binary file
 	-When you try to lad a binary file as a text file
 
-If this is not the problem, or you have any other question,  don´t hesitate on 
-dropping a line to
-Ricardo dot Ramirez-Gonzalez at bbsrc dot ac dot uk
+I keep seeing this "Invalid gemspec in [some ruby gem path...]", what is wrong?
+- This appears to be a bug in RubyGems that doesn't affect the running of the tools. It will keep happening until someone updates RubyGems. If it really bugs you, downgrade RubyGems.
+
+To whom do I complain?
+-about bio-samtools? Try
+Ricardo dot Ramirez-Gonzalez at tgac dot ac dot uk
 dan.maclean@tsl.ac.uk
 
 == TODO
 -Filter to the fetching algorithm (give a condition that has to be satisfied to add the alignment to the list)
--Examples of how to use it, besides the test folder
--Operating system independent, DONE ( test needed)
+
 
 == IMPORTANT NOTE
--Libraries are downloaded, compiled and installed inside the gem at install time on the host system
--Library .a, I didn't check if it works on linux (test needed) 
+-Libraries (libbam) are downloaded, compiled and installed inside the gem at install time on the host system
+ 
 $ openssl dgst libbam.so.1
 MD5(libbam.so.1)= c45cfccfb41ffeb2730ee4b227d244c4
 

data/Rakefile

@@ -32,17 +32,6 @@ Jeweler::Tasks.new do |gem|
 end
 Jeweler::RubygemsDotOrgTasks.new
 
-desc "Basic Functionality Test"
-task :basic_test do
-  ruby "test/basictest.rb"
-end
-
-
-desc "Test the Pileup class"
-task :pileup_test do 
-  ruby "test/pileup.rb"
-end
-
 require 'rake/testtask'
 Rake::TestTask.new(:test) do |test|
   test.libs << 'lib' << 'test'
@@ -59,8 +48,8 @@ end
 
 task :default => :test
 
-require 'rake/rdoctask'
-Rake::RDocTask.new do |rdoc|
+require 'rdoc/task'
+RDoc::Task.new do |rdoc|
   version = File.exist?('VERSION') ? File.read('VERSION') : ""
 
   rdoc.rdoc_dir = 'rdoc'

data/VERSION

@@ -1 +1 @@
-0.3.2
+0.4.0

data/bio-samtools.gemspec

@@ -5,11 +5,11 @@
 
 Gem::Specification.new do |s|
   s.name = %q{bio-samtools}
-  s.version = "0.3.2"
+  s.version = "0.4.0"
 
   s.required_rubygems_version = Gem::Requirement.new(">= 0") if s.respond_to? :required_rubygems_version=
   s.authors = [%q{Ricardo Ramirez-Gonzalez}, %q{Dan MacLean}, %q{Raoul J.P. Bonnal}]
-  s.date = %q{2011-10-27}
+  s.date = %q{2011-11-09}
   s.description = %q{Binder of samtools for ruby, on the top of FFI. 
 
   This project was born from the need to add support of BAM files to 
@@ -94,6 +94,7 @@ Gem::Specification.new do |s|
     "doc/tutorial.html",
     "doc/tutorial.pdf",
     "ext/Makefile-bioruby.patch",
+    "ext/Rakefile",
     "ext/mkrf_conf.rb",
     "lib/bio-samtools.rb",
     "lib/bio/.DS_Store",
@@ -101,10 +102,13 @@ Gem::Specification.new do |s|
     "lib/bio/db/sam/bam.rb",
     "lib/bio/db/sam/external/COPYING",
     "lib/bio/db/sam/external/VERSION",
+    "lib/bio/db/sam/external/libbam.1.dylib",
+    "lib/bio/db/sam/external/libbam.a",
     "lib/bio/db/sam/faidx.rb",
     "lib/bio/db/sam/library.rb",
     "lib/bio/db/sam/pileup.rb",
     "lib/bio/db/sam/sam.rb",
+    "lib/bio/db/sam/vcf.rb",
     "test/basictest.rb",
     "test/coverage.rb",
     "test/coverage_plot.rb",
@@ -123,7 +127,6 @@ Gem::Specification.new do |s|
     "test/samples/small/test_chr.fasta.amb",
     "test/samples/small/test_chr.fasta.ann",
     "test/samples/small/test_chr.fasta.bwt",
-    "test/samples/small/test_chr.fasta.fai",
     "test/samples/small/test_chr.fasta.pac",
     "test/samples/small/test_chr.fasta.rbwt",
     "test/samples/small/test_chr.fasta.rpac",
@@ -131,7 +134,10 @@ Gem::Specification.new do |s|
     "test/samples/small/test_chr.fasta.sa",
     "test/samples/small/testu.bam",
     "test/samples/small/testu.bam.bai",
-    "test/test_bio-samtools.rb"
+    "test/test_basic.rb",
+    "test/test_bio-samtools.rb",
+    "test/test_pileup.rb",
+    "test/test_vcf.rb"
   ]
   s.homepage = %q{http://github.com/helios/bioruby-samtools}
   s.licenses = [%q{MIT}]
@@ -151,6 +157,7 @@ Gem::Specification.new do |s|
       s.add_development_dependency(%q<rcov>, [">= 0"])
       s.add_development_dependency(%q<bio>, [">= 1.4.2"])
       s.add_development_dependency(%q<ffi>, [">= 0"])
+      s.add_development_dependency(%q<rdoc>, [">= 0"])
     else
       s.add_dependency(%q<ffi>, [">= 0"])
       s.add_dependency(%q<bio>, [">= 1.4.2"])
@@ -160,6 +167,7 @@ Gem::Specification.new do |s|
       s.add_dependency(%q<rcov>, [">= 0"])
       s.add_dependency(%q<bio>, [">= 1.4.2"])
       s.add_dependency(%q<ffi>, [">= 0"])
+      s.add_dependency(%q<rdoc>, [">= 0"])
     end
   else
     s.add_dependency(%q<ffi>, [">= 0"])
@@ -170,6 +178,7 @@ Gem::Specification.new do |s|
     s.add_dependency(%q<rcov>, [">= 0"])
     s.add_dependency(%q<bio>, [">= 1.4.2"])
     s.add_dependency(%q<ffi>, [">= 0"])
+    s.add_dependency(%q<rdoc>, [">= 0"])
   end
 end
 

data/ext/Rakefile

@@ -0,0 +1,48 @@
+require 'rbconfig'
+require 'open-uri'
+require 'fileutils'
+include FileUtils::Verbose
+require 'rake/clean'
+
+URL = "http://sourceforge.net/projects/samtools/files/samtools/0.1.18/samtools-0.1.18.tar.bz2/download"
+
+task :download do
+  open(URL) do |uri|
+    File.open("samtools-0.1.18.tar.bz2",'wb') do |fout|
+      fout.write(uri.read)
+    end #fout 
+  end #uri
+end
+    
+task :compile do
+  sh "tar xvfj samtools-0.1.18.tar.bz2"
+  cd("samtools-0.1.18") do
+    sh "patch < ../Makefile-bioruby.patch"
+    case Config::CONFIG['host_os']
+      when /linux/
+        #sh "CFLAGS='-g -Wall -O2 -fPIC' make -e"
+        sh "make"
+        cp("libbam.a","/Users/macleand/bioruby-samtools/ext/../lib/bio/db/sam/external")
+        #sh "CFLAGS='-g -Wall -O2 -fPIC' make -e libbam.so.1-local"
+        sh "make libbam.so.1-local"
+        cp("libbam.so.1","/Users/macleand/bioruby-samtools/ext/../lib/bio/db/sam/external")
+      when /darwin/
+        sh "make"
+        cp("libbam.a","/Users/macleand/bioruby-samtools/ext/../lib/bio/db/sam/external")
+        sh "make libbam.1.dylib-local"
+        cp("libbam.1.dylib","/Users/macleand/bioruby-samtools/ext/../lib/bio/db/sam/external")      
+      when /mswin|mingw/ then raise NotImplementedError, "BWA library is not available for Windows platform"  
+    end #case
+  end #cd
+end
+  
+task :clean do
+  cd("samtools-0.1.18") do
+    sh "make clean"
+  end
+  rm("samtools-0.1.18.tar.bz2")
+  rm_rf("samtools-0.1.18")
+end
+
+task :default => [:download, :compile, :clean]
+  

data/lib/bio/db/sam.rb

@@ -2,6 +2,8 @@ require 'bio/db/sam/library'
 require 'bio/db/sam/bam'
 require 'bio/db/sam/faidx'
 require 'bio/db/sam/sam'
+require 'bio/db/sam/pileup'
+require 'bio/db/sam/vcf'
 
 module LibC
   extend FFI::Library
@@ -277,76 +279,76 @@ module Bio
       #call the option as a symbol of the flag, eg -r for region is called :r => "some SAM compatible region"
       #eg bam.mpileup(:r => "chr1:1000-2000", :q => 50) gets the bases with quality > 50 on chr1 between 1000-5000 
       def mpileup( opts )
-        
-        raise SAMException.new(), "No BAMFile provided" unless @sam and @binary
-        raise SAMException.new(), "No FastA provided" unless @fasta_path
-        #long option form to short samtools form..
-        long_opts = {
-        :region => :r,
-        :illumina_quals => :six,
-        :count_anomalous => :A,
-        :no_baq => :B,
-        :adjust_mapq => :C,
-        :max_per_bam_depth => :d,
-        :extended_baq => :E,
-        :exclude_reads_file => :G,
-        :list_of_positions => :l,
-        :mapping_quality_cap => :M,
-        :ignore_rg => :R,
-        :min_mapping_quality => :q,
-        :min_base_quality => :Q
-        }
-        
-        ##convert any long_opts to short opts 
-        opts.each_pair do |k,v|
-          if long_opts[k]
-            opts[long_opts[k]] = v 
-            opts.delete(k)
-          end
-        end
-        
-        ##remove any calls to -g or -u for mpileup, bcf output is not yet supported
-        ##and also associated output options
-        [:g, :u, :e, :h, :I, :L, :o, :p].each {|x| opts.delete(x) }
-        
-        strptrs = []
-        strptrs << FFI::MemoryPointer.from_string("mpileup")
-        opts.each do |k,v|
-          next unless opts[k] ##dont bother unless the values provided are true.. 
-          k = '6' if k == :six
-          k = '-' + k.to_s
-          strptrs << FFI::MemoryPointer.from_string(k)
-          strptrs << FFI::MemoryPointer.from_string(v.to_s) unless ["-R", "-B", "-E", "-6", "-A"].include?(k) #these are just flags so don't pass a value...
-        end
-        strptrs << FFI::MemoryPointer.from_string('-f')
-        strptrs << FFI::MemoryPointer.from_string(@fasta_path)
-        strptrs << FFI::MemoryPointer.from_string(@sam)
-        strptrs << nil
 
-        # Now load all the pointers into a native memory block
-        argv = FFI::MemoryPointer.new(:pointer, strptrs.length)
-        strptrs.each_with_index do |p, i|
-           argv[i].put_pointer(0,  p)
-        end
-        
-        old_stdout = STDOUT.clone
-        read_pipe, write_pipe = IO.pipe()
-        STDOUT.reopen(write_pipe)
-          #int bam_mpileup(int argc, char *argv[])
-          Bio::DB::SAM::Tools.bam_mpileup(strptrs.length - 1,argv)
-          if fork
-            write_pipe.close
-            STDOUT.reopen(old_stdout) #beware .. stdout from other processes eg tests calling this method can get mixed in...
-            begin
-              while line = read_pipe.readline
-                yield Pileup.new(line)
+              raise SAMException.new(), "No BAMFile provided" unless @sam and @binary
+              raise SAMException.new(), "No FastA provided" unless @fasta_path
+              #long option form to short samtools form..
+              long_opts = {
+              :region => :r,
+              :illumina_quals => :six,
+              :count_anomalous => :A,
+              :no_baq => :B,
+              :adjust_mapq => :C,
+              :max_per_bam_depth => :d,
+              :extended_baq => :E,
+              :exclude_reads_file => :G,
+              :list_of_positions => :l,
+              :mapping_quality_cap => :M,
+              :ignore_rg => :R,
+              :min_mapping_quality => :q,
+              :min_base_quality => :Q
+              }
+
+              ##convert any long_opts to short opts 
+              opts.each_pair do |k,v|
+                if long_opts[k]
+                  opts[long_opts[k]] = v 
+                  opts.delete(k)
+                end
               end
-            rescue EOFError
-              read_pipe.close
-              Process.wait
+
+              ##remove any calls to -g or -u for mpileup, bcf output is not yet supported
+              ##and also associated output options
+              [:g, :u, :e, :h, :I, :L, :o, :p].each {|x| opts.delete(x) }
+
+              strptrs = []
+              strptrs << FFI::MemoryPointer.from_string("mpileup")
+              opts.each do |k,v|
+                next unless opts[k] ##dont bother unless the values provided are true.. 
+                k = '6' if k == :six
+                k = '-' + k.to_s
+                strptrs << FFI::MemoryPointer.from_string(k)
+                strptrs << FFI::MemoryPointer.from_string(v.to_s) unless ["-R", "-B", "-E", "-6", "-A"].include?(k) #these are just flags so don't pass a value...
+              end
+              strptrs << FFI::MemoryPointer.from_string('-f')
+              strptrs << FFI::MemoryPointer.from_string(@fasta_path)
+              strptrs << FFI::MemoryPointer.from_string(@sam)
+              strptrs << nil
+
+              # Now load all the pointers into a native memory block
+              argv = FFI::MemoryPointer.new(:pointer, strptrs.length)
+              strptrs.each_with_index do |p, i|
+                 argv[i].put_pointer(0,  p)
+              end
+
+              old_stdout = STDOUT.clone
+              read_pipe, write_pipe = IO.pipe()
+              STDOUT.reopen(write_pipe)
+                #int bam_mpileup(int argc, char *argv[])
+                Bio::DB::SAM::Tools.bam_mpileup(strptrs.length - 1,argv)
+                if fork
+                  write_pipe.close
+                  STDOUT.reopen(old_stdout) #beware .. stdout from other processes eg tests calling this method can get mixed in...
+                  begin
+                    while line = read_pipe.readline
+                      yield Pileup.new(line)
+                    end
+                  rescue EOFError
+                    read_pipe.close
+                    Process.wait
+                  end
+                end
             end
-          end
-      end
       
       # utility method that does not use the samtools API, it calls samtools directly as if on the command line and catches the output,
       # to use this method you must have a version of samtools that supports the pileup command (< 0.1.17)
@@ -373,6 +375,49 @@ module Bio
         end
         pipe.close
       end
+      
+      
+      def index_stats 
+        raise SAMException.new(), "No BAMFile provided" unless @sam and @binary
+        raise SAMException.new(), "No FastA provided" unless @fasta_path
+        strptrs = []
+        strptrs << FFI::MemoryPointer.from_string("idxstats")
+        strptrs << FFI::MemoryPointer.from_string(@sam)
+        strptrs << nil
+
+        # Now load all the pointers into a native memory block
+        argv = FFI::MemoryPointer.new(:pointer, strptrs.length)
+        strptrs.each_with_index do |p, i|
+           argv[i].put_pointer(0,  p)
+        end
+        
+        index_stats = {}
+        
+        old_stdout = STDOUT.clone
+        read_pipe, write_pipe = IO.pipe()
+        STDOUT.reopen(write_pipe)
+        
+        #int bam_idxstats(int argc, char *argv[])
+        Bio::DB::SAM::Tools.bam_idxstats(strptrs.length - 1,argv)
+        if fork
+          write_pipe.close
+          STDOUT.reopen(old_stdout) #beware .. stdout from other processes eg tests calling this method can get mixed in...
+          begin
+            
+            while line = read_pipe.readline #TAB delimited with each line consisting of reference sequence name, sequence length, # mapped reads and # unmapped reads.
+                info = line.split(/\t/)
+                next unless info.length == 4
+                index_stats[ info[0] ] = {:length => info[1].to_i, :mapped_reads => info[2].to_i, :unmapped_reads => info[3].to_i } 
+            end
+            rescue EOFError
+              read_pipe.close
+              Process.wait
+            end
+        end #fork
+        index_stats
+      end
+      
+      
 
     end
 

data/lib/bio/db/sam/bam.rb

@@ -1,5 +1,5 @@
-# require 'rubygems'
-# require'ffi'
+#require 'rubygems'
+#require 'ffi'
 # require 'bio/db/sam/faidx'
 # require 'bio/db/sam/sam'
 module Bio

data/lib/bio/db/sam/sam.rb

@@ -79,6 +79,7 @@ module Bio
         attach_function :sampileup, [ :pointer, :int, :bam_pileup_f, :pointer ], :int
         attach_function :samfaipath, [ :string ], :string
         attach_function :bam_mpileup, [ :int, :pointer ], :int
+        attach_function :bam_idxstats, [:int, :pointer ], :int
       end
     end
   end

data/lib/bio/db/sam/vcf.rb

@@ -0,0 +1,78 @@
+# :title: Vcf
+# = Vcf 
+# A class representing information in Variant Call Format
+# Forked from vcfruby at https://github.com/jesserod/vcfruby 
+# Modified and tests written by Dan MacLean (dan.maclean@tsl.ac.uk)
+# VCF is described at http://www.1000genomes.org/node/101
+
+class Vcf
+public
+  attr_accessor :chrom, :pos, :id, :ref, :alt, :qual, :filter, :info, :format, :samples
+
+  #create the vcf object, use the ordered list of sample names to label samples if provided ['A', 'B', 'C'], otherwise uses, 1,2,3 etc
+  def initialize(line=nil, sample_names=nil)
+    next if line[0,1] == '#' #skip any header or meta-lines... inadvertantly passed.. 
+    @info = {}
+    @samples = {}
+    parse_line(line, sample_names) if line != nil
+  end
+
+  def int_or_raw(x)
+    Integer.new(x) rescue x
+  end
+
+  #gets the info in the Vcf lines and parses the
+  def parse_line(line, sample_names=nil)
+    return false if line[0,1] == '#'
+
+    f = line.chomp.split("\t", -1)
+    raise "VCF lines must have at least 8 fields" if f.size < 8
+    @chrom = f[0]
+    @pos = f[1].to_i
+    @id =  '.' == f[2] ? nil : f[2]
+    @ref = f[3]
+    @alt = f[4]
+    @qual = f[5].to_f
+    @filter = '.' == f[6] ? nil : f[6]
+
+    @info = '.' == f[7] ? nil : {}
+    if @info
+      info_vec = f[7].split(";")
+      info_vec.each do |x|
+        keyval = x.split("=", -1)
+        if keyval.size == 2 # If it's key=value
+          @info[keyval[0]] = keyval[1]
+        else # Otherwise, it's just a flag
+          @info[x] = nil
+        end
+      end
+    end
+
+    @samples = {}
+    return true if f.size == 8  # Has just upto info
+    raise "Can't have format with no samples" if f.size == 9
+    
+    @format = f[8]
+
+    sample_keys = @format.split(":")
+
+    num_samples = f.size - 9 # How many fields are past the format
+
+    if sample_names == nil # Make the sample names just ["1", "2", ... , "num_samples}"
+      sample_names = (1..num_samples).to_a.map{|i| i.to_s} 
+    elsif sample_names.size != num_samples
+      raise "Unexpected number of samples (#{num_samples}) based on the provided sample names (#{sample_names.inspect})"
+    end
+
+    sample_names.each_with_index do |sample_name, sample_index|
+      i = sample_index + 9 # index into columns (f)
+      sample_values = f[i].split(":")
+      raise "Expected number of sample values to be <= number of sample keys in FORMAT column Format=#{@format} but sample=#{f[i]}" if sample_values.size > sample_keys.size
+      @samples[sample_name] = {}
+      sample_keys.each_with_index {|key, value_index| @samples[sample_name][key] = sample_values[value_index] || ""}
+    end
+    
+    return true;
+  end
+
+end

data/test/helper.rb

@@ -1,5 +1,5 @@
-require 'rubygems'
 require 'bundler'
+
 begin
   Bundler.setup(:default, :development)
 rescue Bundler::BundlerError => e

data/test/test_basic.rb

@@ -0,0 +1,342 @@
+$: << File.expand_path(File.dirname(__FILE__) + '/../lib')
+$: << File.expand_path('.')
+require 'rubygems'
+require 'ffi'
+require "test/unit"
+require "bio/db/sam"
+require "bio/db/sam/sam"
+
+
+class TestBioDbSam < Test::Unit::TestCase
+
+  #Set up the paths
+  def setup
+    @test_folder                = "test/samples/small"
+    @testTAMFile                = @test_folder + "/test.tam"
+    @testBAMFile                = @test_folder + "/testu.bam"
+    @testReference              = @test_folder + "/test_chr.fasta"
+    
+  end
+  
+  #Removing the index files
+  def teardown
+    begin
+      File.delete(@testReference + ".fai")
+      p "deleted: " + @testReference + ".fai "
+    rescue
+    end
+    begin
+      File.delete(@testBAMFile + ".fai")
+      p "deleted: " + @testBAMFile + ".bai "
+    rescue
+    end
+  end
+=begin
+  def default_test
+    puts $LOAD_PATH
+    assert(true, "Unit test test")
+  end
+
+  def test_openSAMFile
+    bamfile                      = Bio::DB::SAM::Tools.samopen(@testTAMFile,"r",nil)
+    Bio::DB::SAM::Tools.samclose(bamfile)
+    assert(true, "file open and closed")
+  end
+
+  def test_new_class_empty
+    begin
+      bam                        = Bio::DB::Sam.new({})
+      assert(false, "Should fail while opening without parameters")
+    rescue Bio::DB::SAMException => e
+      puts e.message
+      assert(true, e.message)
+    end
+  end
+
+  def test_new_class_empty_invalid_path
+    begin
+      sam                        = Bio::DB::Sam.new({:bam=>"INVALID"})
+      sam.open
+      sam.close
+      assert(false, "Should fail with an invalid path")
+    rescue Bio::DB::SAMException => e
+      puts e.message
+      assert(true, e.message)
+    end
+  end
+
+  def test_class_text_read_no_faidx
+    sam                          = Bio::DB::Sam.new({:tam=>@testTAMFile})
+    sam.open
+    sam.close
+    assert(true, "file open and closed with the class")
+  end
+
+  def test_class_text_read_no_close
+
+    fam                          = Bio::DB::Sam.new({:tam=>@testTAMFile})
+    fam.open
+    fam                          = nil   
+    ObjectSpace.garbage_collect
+
+    assert(true, "file openend but not closed")
+  end
+
+  def test_class_binary_read_no_close
+
+    Bio::DB::Sam.new({:bam=>@testBAMFile}).open
+    ObjectSpace.garbage_collect
+    assert(true, "BINARY file openend but not closed")  
+  end
+
+  def test_read_coverage
+     sam       = Bio::DB::Sam.new({:bam=>@testBAMFile, :fasta=>@testReference})
+     sam.open
+  	File.open( @test_folder +"/ids2.txt", "r") do |file|
+  	  puts "file opened"
+  	  file.each_line{|line|
+        fetching = line.split(' ')[0]
+        puts "fetching: " + fetching
+  	    sam.load_reference
+    	  seq = sam.fetch_reference(fetching, 0, 16000)
+  #	  puts seq 
+  #	  puts seq.length
+  	  als = sam.fetch(fetching, 0, seq.length) 
+     #      p als
+  	  if als.length() > 0 then
+  	       p fetching
+                p als
+  	    end
+           }
+
+  	end
+    sam.close
+   assert(true, "Finish")
+  end
+#  def test_read_TAM_as_BAM
+#    begin
+#      sam                          = Bio::DB::Sam.new({:bam=>@testTAMFile})
+#      sam.open
+#      sam.close
+#      assert(false, "Should raise an exception for reading a BAM as TAM") 
+#    rescue Bio::DB::SAMException => e
+#      assert(true, "Properly handled")
+#    end 
+#  end
+
+# def test_read_BAM_as_TAM
+#    begin
+#      sam                          = Bio::DB::Sam.new({:tam=>@testBAMFile})
+#      sam.open
+#      sam.close
+#      assert(false, "Should raise an exception for reading a BAM as TAM") 
+#    rescue Bio::DB::SAMException => e
+#      assert(true, "Properly handled")
+#    end 
+#  end
+
+  def test_bam_load_index
+    sam       = Bio::DB::Sam.new({:bam=>@testBAMFile})
+    sam.open
+    index = sam.load_index
+    sam.close
+    assert(true, "BAM index loaded")
+    #  attach_function :bam_index_build, [ :string ], :int
+    #  attach_function :bam_index_load, [ :string ], :pointer
+    #  attach_function :bam_index_destroy, [ :pointer ], :void
+  end
+
+  def test_tam_load_index
+    begin
+      sam       = Bio::DB::Sam.new({:tam=>@testTAMFile})
+      sam.open
+      sam.load_index
+      sam.close
+      assert(false, "TAM index loaded")
+    rescue Bio::DB::SAMException => e
+      assert(true, "Unable to load an index for a TAM file")
+    end
+  end
+
+  def test_read_segment
+    sam       = Bio::DB::Sam.new({:bam=>@testBAMFile})
+    sam.open
+    als = sam.fetch("chr_1", 0, 500)
+    p als 
+    sam.close
+    assert(true, "Seems it ran the query")
+    #node_7263       238     60 has 550+, query from 0 to 500, something shall come.... 
+  end
+
+  def test_read_invalid_reference
+    sam       = Bio::DB::Sam.new({:bam=>@testBAMFile})
+    sam.open
+    begin
+      als = sam.fetch("Chr1", 0, 500)
+      p als 
+      sam.close
+      assert(false, "Seems it ran the query")
+    rescue Bio::DB::SAMException => e
+      p e
+      assert(true, "Exception generated and catched")
+    end
+  end
+
+  def test_read_invalid_reference_start_coordinate
+    sam       = Bio::DB::Sam.new({:bam=>@testBAMFile})
+    sam.open
+    begin
+      als = sam.fetch("chr", -1, 500)
+      p als 
+      sam.close
+      assert(false, "Seems it ran the query")
+    rescue Bio::DB::SAMException => e
+      p e
+      assert(true, "Exception generated and catched")
+    end
+  end
+
+  def test_read_invalid_reference_end_coordinate
+    sam       = Bio::DB::Sam.new({:bam=>@testBAMFile})
+    sam.open
+    begin
+      als = sam.fetch("chr", 0, 50000)
+      p als 
+      sam.close
+      assert(false, "Seems it ran the query")
+    rescue  Bio::DB::SAMException => e
+      p e
+      assert(true, "Exception generated and catched")
+    end
+  end
+  
+  def test_read_invalid_reference_swaped_coordinates
+    sam       = Bio::DB::Sam.new({:bam=>@testBAMFile})
+    sam.open
+    begin
+      als = sam.fetch("chr", 500, 0)
+      p als 
+      sam.close
+      assert(false, "Seems it ran the query")
+    rescue  Bio::DB::SAMException => e
+      p e
+      assert(true, "Exception generated and catched")
+    end
+  end
+
+  def test_fasta_load_index
+    sam = Bio::DB::Sam.new({:fasta=>@testReference})
+    sam.load_reference
+    seq = sam.fetch_reference("chr_1", 0, 500)
+    p seq 
+    sam.close
+    assert(true, "The reference was loaded")
+  end
+  
+  def test_fasta_load_index
+    sam = Bio::DB::Sam.new({:fasta=>@testReference})
+    sam.load_reference
+    begin
+      seq = sam.fetch_reference("chr1", 0, 500)
+      p "Error seq:"+ seq 
+      sam.close
+      assert(false, "The reference was loaded")
+    rescue Bio::DB::SAMException => e
+      p e
+      assert(true,  "The references was not loaded")
+    end
+  end
+  
+  def test_load_feature
+    
+    fs = Feature.find_by_bam("chr_1", 0, 500,@testBAMFile)
+    
+    p fs
+    assert(true, "Loaded as features")
+  end
+  
+  def test_avg_coverage
+    sam = Bio::DB::Sam.new({:fasta=>@testReference, :bam=>@testBAMFile })
+    sam.open
+    cov = sam.average_coverage("chr_1", 60, 30)
+    p "Coverage: " + cov.to_s
+    sam.close
+    assert(true, "Average coverage ran")
+    assert(3 == cov, "The coverage is 3")
+  end
+  
+
+  def test_chromosome_coverage
+    sam = Bio::DB::Sam.new({:fasta=>@testReference, :bam=>@testBAMFile })
+    sam.open
+    covs = sam.chromosome_coverage("chr_1", 0, 60)
+    p "Coverage: "
+    p covs
+    puts "POS\tCOV"
+    covs.each_with_index{ |cov, i| puts "#{i}\t#{cov}" }
+    sam.close
+    assert(true, "Average coverage ran")
+    #assert(3 == cov, "The coverage is 3")
+  end
+=end
+  #test whether the call to mpileup works and returns 10 objects of class pileup
+  def test_pileup
+    sam = Bio::DB::Sam.new(:fasta=>@testReference, :bam=>@testBAMFile )
+    pileup_list = []
+    sam.mpileup(:region => "chr_1:100-109") do |pile|
+      next unless pile.ref_name == 'chr_1' ##required because in the test environment stdout gets mixed in with the captured stdout in the function and non pileup lines are passed...
+      pileup_list << pile
+    end
+    assert_equal(10,pileup_list.length)
+    pileup_list.each  do |p|
+      assert_kind_of(Pileup, p)
+    end
+  end
+
+
+  #test whether the call to mpileup returns a vcf object if :g => true is used on the command-line
+#  def test_vcf
+
+#    sam = Bio::DB::Sam.new(:fasta=>@testReference, :bam=>@testBAMFile )
+#        sam.mpileup(:region => "chr_1:100-109", :u => true ) do |p|
+#        $stderr.puts "p is #{p}"
+#        assert_kind_of(String,p)
+#    end
+#  end
+  def test_indexstats
+
+    sam = Bio::DB::Sam.new(:bam => @testBAMFile, :fasta => @testReference)
+    assert_equal({"chr_1"=>{:length=>69930, :mapped_reads=>0, :unmapped_reads=>0},
+                  "*"=>{:length=>0, :mapped_reads=>0, :unmapped_reads=>0}
+                 }, sam.index_stats)
+  end
+
+end
+
+class Feature 
+attr_reader :start, :end, :strand, :sequence, :quality
+
+def initialize(a={})
+  p a
+  @start = a[:start]
+  @end = a[:enf]
+  @strand = a[:strand]
+  @sequence = a[:sequence]
+  @quality = a[:quality]
+end
+
+def self.find_by_bam(reference,start,stop,bam_file_path)
+  
+  sam = Bio::DB::Sam.new({:bam=>bam_file_path})
+  features = []
+  sam.open
+  
+  fetchAlignment = Proc.new do |a|
+    a.query_strand ? strand = '+'  : strand = '-'
+    features << Feature.new({:start=>a.pos,:end=>a.calend,:strand=>strand,:sequence=>a.seq,:quality=>a.qual})
+  end
+  sam.fetch_with_function(reference, start, stop, fetchAlignment)
+  sam.close
+  features
+end
+end

data/test/test_bio-samtools.rb

@@ -1,7 +1 @@
 require 'helper'
-
-class TestBioSamtools < Test::Unit::TestCase
-  should "probably rename this file and start testing for real" do
-    flunk "hey buddy, you should probably rename this file and start testing for real"
-  end
-end

data/test/test_pileup.rb

@@ -0,0 +1,68 @@
+$: << File.expand_path(File.dirname(__FILE__) + '/../lib')
+$: << File.expand_path('.')
+
+require "bio/db/sam/pileup"
+require "test/unit"
+
+class TestPileup < Test::Unit::TestCase
+  
+  def setup
+    @six_col = Pileup.new("seq1\t272\tT\t24\t,.$.....,,.,.,...,,,.,..^+.\t<<<+;<<<<<<<<<<<=<;<;7<&")
+    @ten_col = Pileup.new("seq2\t151\tG\tG\t36\t0\t99\t12\t...........A\t:9<;;7=<<<<<")
+    @snp = Pileup.new("seq1\t272\tT\t24\t,.$.....,,.gGgGgGgGgGgGg^+.\t<<<+;<<<<<<<<<<<=<;<;7<&") 
+    @snp_2 = Pileup.new("seq1\t272\tT\t24\t......aaaaaaggggggcccccc$^+\t<<<+;<<<<<<<<<<<=<;<;7<&")
+    @snp_3 = Pileup.new("seq1\t272\tT\t24\t......aaaaaaaagggggccccc$^+\t<<<+;<<<<<<<<<<<=<;<;7<&")
+  end
+  
+  def test_new_from_6_column
+    assert_equal("seq1", @six_col.ref_name)
+    assert_equal(272, @six_col.pos)
+    assert_equal("T", @six_col.ref_base)
+    assert_equal(24, @six_col.coverage)
+    assert_equal(",.$.....,,.,.,...,,,.,..^+.", @six_col.read_bases)
+    assert_equal("<<<+;<<<<<<<<<<<=<;<;7<&", @six_col.read_quals)
+  end
+  
+  def test_new_from_10_column
+    assert_equal("seq2", @ten_col.ref_name)
+    assert_equal(151, @ten_col.pos)
+    assert_equal("G", @ten_col.ref_base)
+    assert_equal("G", @ten_col.consensus)
+    assert_equal(36, @ten_col.consensus_quality)
+    assert_equal(0, @ten_col.snp_quality)
+    assert_equal(99, @ten_col.rms_mapq)
+    assert_equal(12, @ten_col.coverage)
+    assert_equal("...........A", @ten_col.read_bases)
+    assert_equal(":9<;;7=<<<<<", @ten_col.read_quals)
+  end
+  
+  def test_non_refs
+    assert_equal({:A => 1, :C => 0, :T => 0, :G => 0}, @ten_col.non_refs)
+    assert_equal({:A => 0, :C => 0, :T => 0, :G => 0}, @six_col.non_refs)
+  end
+  
+  def test_consensus
+    assert_equal("G", @snp.consensus)
+    assert_equal("ACGT", @snp_2.consensus)
+    assert_equal("A", @snp_3.consensus)
+  end
+  
+  def test_non_ref_count
+    assert_equal(13,@snp.non_ref_count)
+    assert_equal(18,@snp_2.non_ref_count)
+    assert_equal(18,@snp_3.non_ref_count)
+  end
+  
+  def test_ref_count
+    assert_equal(11,@snp.ref_count)
+    assert_equal(6,@snp_2.ref_count)
+    assert_equal(6,@snp_3.ref_count)
+  end
+  
+  def test_ref_plus_non_ref_equal_to_coverage
+    assert_equal(@snp.coverage,@snp.ref_count + @snp.non_ref_count)
+   assert_equal(@snp_2.coverage,@snp_2.ref_count + @snp_2.non_ref_count)
+   assert_equal(@snp_3.coverage,@snp_3.ref_count + @snp_3.non_ref_count)
+  end
+ 
+end

data/test/test_vcf.rb

@@ -0,0 +1,68 @@
+$: << File.expand_path(File.dirname(__FILE__) + '/../lib')
+$: << File.expand_path('.')
+require 'rubygems'
+require 'bio/db/sam/vcf'
+require "test/unit"
+
+
+class TestVcf < Test::Unit::TestCase
+  
+  def setup
+    @vcf1 = Vcf.new("19	111	.	A	C	9.6	.	.	GT:HQ	0|0:10,10	0|0:10,10	0/1:3,3",["a","b","c"]) #from a 3.3 vcf file
+    @vcf2 = Vcf.new("20	14370	rs6054257	G	A	29	0	NS=3;DP=14;AF=0.5;DB;H2	GT:GQ:DP:HQ	0|0:48:1:51,51	1|0:48:8:51,51	1/1:43:5:-1,-1") #from a 3.3 vcf file
+    @vcf3 = Vcf.new("19	111	.	A	C	9.6	.	.	GT:HQ	0|0:10,10	0|0:10,10	0/1:3,3") #from a 4.0 vcf file
+    @vcf4 = Vcf.new("20	14370	rs6054257	G	A	29	PASS	NS=3;DP=14;AF=0.5;DB;H2	GT:GQ:DP:HQ	0|0:48:1:51,51	1|0:48:8:51,51	1/1:43:5:.,") #from a 4.0 vcf file
+  end
+  
+  
+  def test_parse
+    assert_equal("19", @vcf1.chrom)
+    assert_equal(111, @vcf1.pos)
+    assert_equal(nil, @vcf1.id)
+    assert_equal("A", @vcf1.ref)
+    assert_equal("C",@vcf1.alt)
+    assert_equal(9.6,@vcf1.qual)
+    assert_equal(nil, @vcf1.filter)
+    assert_equal(nil, @vcf1.info)
+    assert_equal({"a"=>{"GT"=>"0|0", "HQ"=>"10,10"},
+                 "b"=>{"GT"=>"0|0", "HQ"=>"10,10"},
+                 "c"=>{"GT"=>"0/1", "HQ"=>"3,3"}}, @vcf1.samples)
+
+    assert_equal("20", @vcf2.chrom)
+    assert_equal(14370, @vcf2.pos)
+    assert_equal('rs6054257', @vcf2.id)
+    assert_equal("G", @vcf2.ref)
+    assert_equal("A",@vcf2.alt)
+    assert_equal(29,@vcf2.qual)
+    assert_equal("0", @vcf2.filter)
+    assert_equal({"DP"=>"14", "AF"=>"0.5", "NS"=>"3", "DB"=>nil, "H2"=>nil}, @vcf2.info)
+    assert_equal({"1"=>{"DP"=>"1", "GT"=>"0|0", "HQ"=>"51,51", "GQ"=>"48"},
+     "2"=>{"DP"=>"8", "GT"=>"1|0", "HQ"=>"51,51", "GQ"=>"48"},
+     "3"=>{"DP"=>"5", "GT"=>"1/1", "HQ"=>"-1,-1", "GQ"=>"43"}}, @vcf2.samples)
+
+     assert_equal("19", @vcf3.chrom)
+     assert_equal(111, @vcf3.pos)
+     assert_equal(nil, @vcf3.id)
+     assert_equal("A", @vcf3.ref)
+     assert_equal("C",@vcf3.alt)
+     assert_equal(9.6,@vcf3.qual)
+     assert_equal(nil, @vcf3.filter)
+     assert_equal(nil, @vcf3.info)
+     assert_equal({"1"=>{"GT"=>"0|0", "HQ"=>"10,10"},
+                  "2"=>{"GT"=>"0|0", "HQ"=>"10,10"},
+                  "3"=>{"GT"=>"0/1", "HQ"=>"3,3"}}, @vcf3.samples)
+
+     assert_equal("20", @vcf4.chrom)
+     assert_equal(14370, @vcf4.pos)
+     assert_equal('rs6054257', @vcf4.id)
+     assert_equal("G", @vcf4.ref)
+     assert_equal("A",@vcf4.alt)
+     assert_equal(29,@vcf4.qual)
+     assert_equal("PASS", @vcf4.filter)
+     assert_equal({"DP"=>"14", "AF"=>"0.5", "NS"=>"3", "DB"=>nil, "H2"=>nil}, @vcf4.info)
+     assert_equal({"1"=>{"DP"=>"1", "GT"=>"0|0", "HQ"=>"51,51", "GQ"=>"48"},
+                   "2"=>{"DP"=>"8", "GT"=>"1|0", "HQ"=>"51,51", "GQ"=>"48"},
+                   "3"=>{"DP"=>"5", "GT"=>"1/1", "HQ"=>".,", "GQ"=>"43"}}, @vcf4.samples)
+  end
+  
+end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: bio-samtools
 version: !ruby/object:Gem::Version
-  version: 0.3.2
+  version: 0.4.0
   prerelease: 
 platform: ruby
 authors:
@@ -11,11 +11,11 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2011-10-27 00:00:00.000000000Z
+date: 2011-11-09 00:00:00.000000000Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: ffi
-  requirement: &2152165380 !ruby/object:Gem::Requirement
+  requirement: &2151840320 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -23,10 +23,10 @@ dependencies:
         version: '0'
   type: :runtime
   prerelease: false
-  version_requirements: *2152165380
+  version_requirements: *2151840320
 - !ruby/object:Gem::Dependency
   name: bio
-  requirement: &2152164200 !ruby/object:Gem::Requirement
+  requirement: &2151830060 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -34,10 +34,10 @@ dependencies:
         version: 1.4.2
   type: :runtime
   prerelease: false
-  version_requirements: *2152164200
+  version_requirements: *2151830060
 - !ruby/object:Gem::Dependency
   name: shoulda
-  requirement: &2152162920 !ruby/object:Gem::Requirement
+  requirement: &2151828360 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -45,10 +45,10 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *2152162920
+  version_requirements: *2151828360
 - !ruby/object:Gem::Dependency
   name: bundler
-  requirement: &2152161080 !ruby/object:Gem::Requirement
+  requirement: &2151826260 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ~>
@@ -56,10 +56,10 @@ dependencies:
         version: 1.0.0
   type: :development
   prerelease: false
-  version_requirements: *2152161080
+  version_requirements: *2151826260
 - !ruby/object:Gem::Dependency
   name: jeweler
-  requirement: &2152160280 !ruby/object:Gem::Requirement
+  requirement: &2151823300 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -67,10 +67,10 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *2152160280
+  version_requirements: *2151823300
 - !ruby/object:Gem::Dependency
   name: rcov
-  requirement: &2152158920 !ruby/object:Gem::Requirement
+  requirement: &2151821740 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -78,10 +78,10 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *2152158920
+  version_requirements: *2151821740
 - !ruby/object:Gem::Dependency
   name: bio
-  requirement: &2152158080 !ruby/object:Gem::Requirement
+  requirement: &2151820520 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -89,10 +89,10 @@ dependencies:
         version: 1.4.2
   type: :development
   prerelease: false
-  version_requirements: *2152158080
+  version_requirements: *2151820520
 - !ruby/object:Gem::Dependency
   name: ffi
-  requirement: &2152156660 !ruby/object:Gem::Requirement
+  requirement: &2151819540 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -100,7 +100,18 @@ dependencies:
         version: '0'
   type: :development
   prerelease: false
-  version_requirements: *2152156660
+  version_requirements: *2151819540
+- !ruby/object:Gem::Dependency
+  name: rdoc
+  requirement: &2151818000 !ruby/object:Gem::Requirement
+    none: false
+    requirements:
+    - - ! '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :development
+  prerelease: false
+  version_requirements: *2151818000
 description: ! "Binder of samtools for ruby, on the top of FFI. \n\n  This project
   was born from the need to add support of BAM files to \n  the gee_fu genome browser
   (http://github.com/danmaclean/gee_fu)."
@@ -185,6 +196,7 @@ files:
 - doc/tutorial.html
 - doc/tutorial.pdf
 - ext/Makefile-bioruby.patch
+- ext/Rakefile
 - ext/mkrf_conf.rb
 - lib/bio-samtools.rb
 - lib/bio/.DS_Store
@@ -192,10 +204,13 @@ files:
 - lib/bio/db/sam/bam.rb
 - lib/bio/db/sam/external/COPYING
 - lib/bio/db/sam/external/VERSION
+- lib/bio/db/sam/external/libbam.1.dylib
+- lib/bio/db/sam/external/libbam.a
 - lib/bio/db/sam/faidx.rb
 - lib/bio/db/sam/library.rb
 - lib/bio/db/sam/pileup.rb
 - lib/bio/db/sam/sam.rb
+- lib/bio/db/sam/vcf.rb
 - test/basictest.rb
 - test/coverage.rb
 - test/coverage_plot.rb
@@ -214,7 +229,6 @@ files:
 - test/samples/small/test_chr.fasta.amb
 - test/samples/small/test_chr.fasta.ann
 - test/samples/small/test_chr.fasta.bwt
-- test/samples/small/test_chr.fasta.fai
 - test/samples/small/test_chr.fasta.pac
 - test/samples/small/test_chr.fasta.rbwt
 - test/samples/small/test_chr.fasta.rpac
@@ -222,7 +236,10 @@ files:
 - test/samples/small/test_chr.fasta.sa
 - test/samples/small/testu.bam
 - test/samples/small/testu.bam.bai
+- test/test_basic.rb
 - test/test_bio-samtools.rb
+- test/test_pileup.rb
+- test/test_vcf.rb
 homepage: http://github.com/helios/bioruby-samtools
 licenses:
 - MIT
@@ -238,7 +255,7 @@ required_ruby_version: !ruby/object:Gem::Requirement
       version: '0'
       segments:
       - 0
-      hash: 1935979892005829555
+      hash: -1183520990124056077
 required_rubygems_version: !ruby/object:Gem::Requirement
   none: false
   requirements:

data/test/samples/small/test_chr.fasta.fai

@@ -1 +0,0 @@
-chr_1	69930	7	70	71