demultiplexer 0.0.1 → 0.1.0

data/test/test_data_io.rb

@@ -1,7 +1,163 @@
 #!/usr/bin/env ruby
-$:.unshift File.join(File.dirname(__FILE__), '..')
+$LOAD_PATH.unshift File.join(File.dirname(__FILE__), '..')
 
 require 'test/helper'
 
-class TestDataIO < Test::Unit::TestCase 
+# Class for testing the DataIO class.
+class TestDataIO < Test::Unit::TestCase
+  SAMPLES      = <<-DATA.gsub(/^\s+\|/, '')
+                   |#ID\tIndex1\tIndex2
+                   |sample1\tAT\tCG
+                 DATA
+
+  FASTQ_INDEX1 = <<-DATA.gsub(/^\s+\|/, '')
+                   |@index1
+                   |at
+                   |+
+                   |II
+                 DATA
+
+  FASTQ_INDEX2 = <<-DATA.gsub(/^\s+\|/, '')
+                   |@index2
+                   |ct
+                   |+
+                   |II
+                 DATA
+
+  FASTQ_READ1  = <<-DATA.gsub(/^\s+\|/, '')
+                   |@read1
+                   |A
+                   |+
+                   |I
+                 DATA
+
+  FASTQ_READ2  = <<-DATA.gsub(/^\s+\|/, '')
+                   |@read2
+                   |G
+                   |+
+                   |I
+                 DATA
+
+  def setup
+    @tmp_input  = Dir.mktmpdir('data_io_input')
+    @tmp_output = Dir.mktmpdir('data_io_input')
+
+    file_samples = File.join(@tmp_input, 'samples.txt')
+    file_index1  = File.join(@tmp_input, 'x_S1_L001_I1_001x.fq')
+    file_index2  = File.join(@tmp_input, 'x_S1_L001_I2_001x.fq')
+    file_read1   = File.join(@tmp_input, 'x_S1_L001_R1_001x.fq')
+    file_read2   = File.join(@tmp_input, 'x_S1_L001_R2_001x.fq')
+
+    File.open(file_samples, 'w') { |ios| ios.puts SAMPLES }
+    File.open(file_index1,  'w') { |ios| ios.puts FASTQ_INDEX1 }
+    File.open(file_index2,  'w') { |ios| ios.puts FASTQ_INDEX2 }
+    File.open(file_read1,   'w') { |ios| ios.puts FASTQ_READ1 }
+    File.open(file_read2,   'w') { |ios| ios.puts FASTQ_READ2 }
+
+    @samples     = SampleReader.read(file_samples, false, false)
+    @fastq_files = [file_index1, file_index2, file_read1, file_read2]
+  end
+
+  def teardown
+    FileUtils.rm_rf(@tmp_input)
+    FileUtils.rm_rf(@tmp_output)
+  end
+
+  test 'DataIO#new with FASTQ files not matching _R1_ fails' do
+    fastq_files = %w(x_S1_L001_R2_001x)
+
+    assert_raise(DataIOError) do
+      DataIO.new(@samples, fastq_files, false, @tmp_output)
+    end
+  end
+
+  test 'DataIO#new with FASTQ files not matching _R2_ fails' do
+    fastq_files = %w(x_S1_L001_R1_001x)
+
+    assert_raise(DataIOError) do
+      DataIO.new(@samples, fastq_files, false, @tmp_output)
+    end
+  end
+
+  test 'DataIO#new with multiple FASTQ files matching _R1_ fails' do
+    fastq_files = %w(_R1_ _R1_ x_S1_L001_R2_001x)
+
+    assert_raise(DataIOError) do
+      DataIO.new(@samples, fastq_files, false, @tmp_output)
+    end
+  end
+
+  test 'DataIO#new with multiple FASTQ files matching _R2_ fails' do
+    fastq_files = %w(_R2_ _R2_ x_S1_L001_R1_001x)
+
+    assert_raise(DataIOError) do
+      DataIO.new(@samples, fastq_files, false, @tmp_output)
+    end
+  end
+
+  test 'DataIO#new with badly formatted SLR R1 string fails' do
+    fastq_files = %w(x_S1_XXXX_R1_001x x_S1_L001_R2_001x)
+
+    assert_raise(DataIOError) do
+      DataIO.new(@samples, fastq_files, false, @tmp_output)
+    end
+  end
+
+  test 'DataIO#new with badly formatted SLR R2 string fails' do
+    fastq_files = %w(x_S1_XXXX_R2_001x x_S1_L001_R1_001x)
+
+    assert_raise(DataIOError) do
+      DataIO.new(@samples, fastq_files, false, @tmp_output)
+    end
+  end
+
+  test 'DataIO#new with bad number of input files fails' do
+    fastq_files = %w(x_S1_L001_R1_001x
+                     x_S1_L001_R2_001x)
+
+    assert_raise(DataIOError) do
+      DataIO.new(@samples, fastq_files, false, @tmp_output)
+    end
+  end
+
+  test '#open_input_files returns readable file handles' do
+    data_io = DataIO.new(@samples, @fastq_files, false, @tmp_output)
+
+    data_io.open_input_files do |ios|
+      ios.each do |index1, index2, read1, read2|
+        assert_equal('at', index1.seq)
+        assert_equal('ct', index2.seq)
+        assert_equal('A',  read1.seq)
+        assert_equal('G',  read2.seq)
+      end
+    end
+  end
+
+  test '#open_output_files returns writable file handles' do
+    data_io = DataIO.new(@samples, @fastq_files, false, @tmp_output)
+
+    data_io.open_output_files do |ios|
+      ios[0].first.puts 'foo 0'
+      ios[0].last.puts 'bar 0'
+      ios[1].first.puts 'foo 1'
+      ios[1].last.puts 'bar 1'
+    end
+
+    files = Dir["#{@tmp_output}/*"]
+
+    expected_names = ['sample1_S1_L001_R1_001.fastq',
+                      'sample1_S1_L001_R2_001.fastq',
+                      'Undetermined_S1_L001_R1_001.fastq',
+                      'Undetermined_S1_L001_R2_001.fastq']
+
+    result_names = files.map { |file| File.basename(file) }
+
+    assert_equal(expected_names, result_names)
+
+    expected_content = files.each_with_object([]) { |e, a| a << File.read(e) }
+
+    result_content = ["foo 0\n", "bar 0\n", "foo 1\n", "bar 1\n"]
+
+    assert_equal(expected_content, result_content)
+  end
 end

data/test/test_demultiplexer.rb

@@ -1,7 +1,211 @@
 #!/usr/bin/env ruby
-$:.unshift File.join(File.dirname(__FILE__), '..')
+$LOAD_PATH.unshift File.join(File.dirname(__FILE__), '..')
 
 require 'test/helper'
 
-class TestDemultiplexer < Test::Unit::TestCase 
+# Class for testing Demultiplexer.
+class TestDemultiplexer < Test::Unit::TestCase
+  SAMPLES      = <<-DATA.gsub(/^\s+\|/, '')
+                   |#ID\tIndex1\tIndex2
+                   |sample1\taT\tCg
+                 DATA
+
+  FASTQ_INDEX1 = <<-DATA.gsub(/^\s+\|/, '')
+                   |@index1
+                   |Aa
+                   |+
+                   |II
+                 DATA
+
+  FASTQ_INDEX2 = <<-DATA.gsub(/^\s+\|/, '')
+                   |@index2
+                   |cC
+                   |+
+                   |II
+                 DATA
+
+  FASTQ_READ1  = <<-DATA.gsub(/^\s+\|/, '')
+                   |@read1
+                   |A
+                   |+
+                   |I
+                 DATA
+
+  FASTQ_READ2  = <<-DATA.gsub(/^\s+\|/, '')
+                   |@read2
+                   |G
+                   |+
+                   |I
+                 DATA
+
+  STATUS1      = <<-YAML.gsub(/^\s+\|/, '')
+                 |---
+                 |:count: 2
+                 |:match: 0
+                 |:undetermined: 2
+                 |:undetermined_percent: 100.0
+                 |:index1_bad_mean: 0
+                 |:index2_bad_mean: 0
+                 |:index1_bad_min: 0
+                 |:index2_bad_min: 0
+                 |:sample_ids:
+                 |- sample1
+                 |:index1:
+                 |- AT
+                 |:index2:
+                 |- CG
+                 |:time_elapsed: '00:00:00'
+               YAML
+
+  STATUS2      = <<-YAML.gsub(/^\s+\|/, '')
+                 |---
+                 |:count: 2
+                 |:match: 2
+                 |:undetermined: 0
+                 |:undetermined_percent: 0.0
+                 |:index1_bad_mean: 0
+                 |:index2_bad_mean: 0
+                 |:index1_bad_min: 0
+                 |:index2_bad_min: 0
+                 |:sample_ids:
+                 |- sample1
+                 |:index1:
+                 |- AT
+                 |:index2:
+                 |- CG
+                 |:time_elapsed: '00:00:00'
+               YAML
+
+  def setup
+    @tmp_input  = Dir.mktmpdir('data_io_input')
+    @tmp_output = Dir.mktmpdir('data_io_input')
+
+    @file_samples = File.join(@tmp_input, 'samples.txt')
+    file_index1   = File.join(@tmp_input, 'x_S1_L001_I1_001x.fq')
+    file_index2   = File.join(@tmp_input, 'x_S1_L001_I2_001x.fq')
+    file_read1    = File.join(@tmp_input, 'x_S1_L001_R1_001x.fq')
+    file_read2    = File.join(@tmp_input, 'x_S1_L001_R2_001x.fq')
+
+    File.open(@file_samples, 'w') { |ios| ios.puts SAMPLES }
+    File.open(file_index1,   'w') { |ios| ios.puts FASTQ_INDEX1 }
+    File.open(file_index2,   'w') { |ios| ios.puts FASTQ_INDEX2 }
+    File.open(file_read1,    'w') { |ios| ios.puts FASTQ_READ1 }
+    File.open(file_read2,    'w') { |ios| ios.puts FASTQ_READ2 }
+
+    @fastq_files = [file_index1, file_index2, file_read1, file_read2]
+  end
+
+  def teardown
+    FileUtils.rm_rf(@tmp_input)
+    FileUtils.rm_rf(@tmp_output)
+  end
+
+  test '#run produces the correct output files' do
+    Demultiplexer.run(@fastq_files, samples_file: @file_samples,
+                                    output_dir: @tmp_output)
+
+    result = Dir["#{@tmp_output}/*"].map { |file| File.basename file }
+
+    expected = ['Demultiplex.log',
+                'sample1_S1_L001_R1_001.fastq',
+                'sample1_S1_L001_R2_001.fastq',
+                'Undetermined_S1_L001_R1_001.fastq',
+                'Undetermined_S1_L001_R2_001.fastq']
+
+    assert_equal(expected, result)
+  end
+
+  test '#run produces OK log file with mismatches_max: 0' do
+    Demultiplexer.run(@fastq_files, samples_file: @file_samples,
+                                    mismatches_max: 0, output_dir: @tmp_output)
+
+    result = File.read(File.join(@tmp_output, 'Demultiplex.log'))
+
+    assert_equal(STATUS1, result)
+  end
+
+  test '#run produces OK R1 file with mismatches_max: 0' do
+    Demultiplexer.run(@fastq_files, samples_file: @file_samples,
+                                    mismatches_max: 0, output_dir: @tmp_output)
+
+    result = File.read(File.join(@tmp_output, 'sample1_S1_L001_R1_001.fastq'))
+
+    assert_equal('', result)
+  end
+
+  test '#run produces OK R2 file with mismatches_max: 0' do
+    Demultiplexer.run(@fastq_files, samples_file: @file_samples,
+                                    mismatches_max: 0, output_dir: @tmp_output)
+
+    result = File.read(File.join(@tmp_output, 'sample1_S1_L001_R2_001.fastq'))
+
+    assert_equal('', result)
+  end
+
+  test '#run produces OK Undetermined R1 file with mismatches_max: 0' do
+    Demultiplexer.run(@fastq_files, samples_file: @file_samples,
+                                    mismatches_max: 0, output_dir: @tmp_output)
+
+    file   = File.join(@tmp_output, 'Undetermined_S1_L001_R1_001.fastq')
+    result = File.read(file)
+
+    assert_equal("@read1 Aa\nA\n+\nI\n", result)
+  end
+
+  test '#run produces OK Undetermined R2 file with mismatches_max: 0' do
+    Demultiplexer.run(@fastq_files, samples_file: @file_samples,
+                                    mismatches_max: 0, output_dir: @tmp_output)
+
+    file   = File.join(@tmp_output, 'Undetermined_S1_L001_R2_001.fastq')
+    result = File.read(file)
+
+    assert_equal("@read2 cC\nG\n+\nI\n", result)
+  end
+
+  test '#run produces OK log with mismatches_max: 1' do
+    Demultiplexer.run(@fastq_files, samples_file: @file_samples,
+                                    mismatches_max: 1, output_dir: @tmp_output)
+
+    result = File.read(File.join(@tmp_output, 'Demultiplex.log'))
+
+    assert_equal(STATUS2, result)
+  end
+
+  test '#run produces OK R1 file with mismatches_max: 1' do
+    Demultiplexer.run(@fastq_files, samples_file: @file_samples,
+                                    mismatches_max: 1, output_dir: @tmp_output)
+
+    result = File.read(File.join(@tmp_output, 'sample1_S1_L001_R1_001.fastq'))
+
+    assert_equal("@read1\nA\n+\nI\n", result)
+  end
+
+  test '#run produces OK R2 file with mismatches_max: 1' do
+    Demultiplexer.run(@fastq_files, samples_file: @file_samples,
+                                    mismatches_max: 1, output_dir: @tmp_output)
+
+    result = File.read(File.join(@tmp_output, 'sample1_S1_L001_R2_001.fastq'))
+
+    assert_equal("@read2\nG\n+\nI\n", result)
+  end
+
+  test '#run produces OK Undetermined R1 file with mismatches_max: 1' do
+    Demultiplexer.run(@fastq_files, samples_file: @file_samples,
+                                    mismatches_max: 1, output_dir: @tmp_output)
+
+    file   = File.join(@tmp_output, 'Undetermined_S1_L001_R1_001.fastq')
+    result = File.read(file)
+
+    assert_equal('', result)
+  end
+
+  test '#run produces OK Undetermined R2 file with mismatches_max: 1' do
+    Demultiplexer.run(@fastq_files, samples_file: @file_samples,
+                                    mismatches_max: 1, output_dir: @tmp_output)
+
+    file   = File.join(@tmp_output, 'Undetermined_S1_L001_R2_001.fastq')
+    result = File.read(file)
+
+    assert_equal('', result)
+  end
 end

data/test/test_index_builder.rb

@@ -1,7 +1,62 @@
 #!/usr/bin/env ruby
-$:.unshift File.join(File.dirname(__FILE__), '..')
+$LOAD_PATH.unshift File.join(File.dirname(__FILE__), '..')
 
 require 'test/helper'
 
-class TestIndexBuilder < Test::Unit::TestCase 
+# Class for testing IndexBuilder.
+class TestIndexBuilder < Test::Unit::TestCase
+  DATA1 = <<-DATA.gsub(/^\s+\|/, '')
+            |#ID\tIndex1\tIndex2
+            |A\tAT\tCG
+          DATA
+
+  DATA2 = <<-DATA.gsub(/^\s+\|/, '')
+            |#ID\tIndex1\tIndex2
+            |A\tAT\tCG
+            |B\tAA\tCC
+          DATA
+
+  def setup
+    @file1 = Tempfile.new('sample_reader')
+    @file2 = Tempfile.new('sample_reader')
+
+    File.open(@file1, 'w') { |ios| ios.puts DATA1 }
+    File.open(@file2, 'w') { |ios| ios.puts DATA2 }
+
+    @samples1 = SampleReader.read(@file1, false, false)
+    @samples2 = SampleReader.read(@file2, false, false)
+  end
+
+  def teardown
+    @file1.close
+    @file1.unlink
+    @file2.close
+    @file2.unlink
+  end
+
+  test 'IndexBuilder#build with mismatches_max 0 returns correctly' do
+    ghash = IndexBuilder.build(@samples1, 0)
+    size  = 0
+
+    ghash.each { |_| size += 1 }
+
+    assert_equal(1, size)
+    assert_equal(0, ghash['ATCG'.hash])
+  end
+
+  test 'IndexBuilder#build with mismatches_max 1 returns correctly' do
+    ghash = IndexBuilder.build(@samples1, 1)
+    size  = 0
+    keys  = %w(AT AA AC AG AT TT CT GT).product(%w(CA CT CC CG AG TG CG GG)).
+            each_with_object([]) { |(a, b), m| m << a + b }.uniq
+
+    ghash.each { |_| size += 1 }
+
+    assert_equal(keys.size, size)
+    keys.each { |key| assert_equal(0, ghash[key.hash]) }
+  end
+
+  test 'IndexBuilder#build with non-unique index combo fails' do
+    assert_raise(IndexBuilderError) { IndexBuilder.build(@samples2, 1) }
+  end
 end

data/test/test_sample_reader.rb

@@ -1,7 +1,80 @@
 #!/usr/bin/env ruby
-$:.unshift File.join(File.dirname(__FILE__), '..')
+$LOAD_PATH.unshift File.join(File.dirname(__FILE__), '..')
 
 require 'test/helper'
 
-class TestSampleReader < Test::Unit::TestCase 
+# Class for testing SampleReader.
+class TestSampleReader < Test::Unit::TestCase
+  DATA1 = <<-DATA.gsub(/^\s+\|/, '')
+            |#ID\tIndex1\tIndex2
+            |A\tATCG\tTCCG
+            |B\tATCG\tTCCC
+          DATA
+
+  DATA2 = <<-DATA.gsub(/^\s+\|/, '')
+            |#ID\tIndex1\tIndex2
+            |A\tATCG\tTCCG
+            |A\tATCG\tTCCC
+          DATA
+
+  DATA3 = <<-DATA.gsub(/^\s+\|/, '')
+            |#ID\tIndex1\tIndex2
+            |A\tATCG\tTCCG
+            |B\tATCG\tTCCG
+          DATA
+
+  def setup
+    @file1 = Tempfile.new('sample_reader')
+    @file2 = Tempfile.new('sample_reader')
+    @file3 = Tempfile.new('sample_reader')
+
+    File.open(@file1, 'w') { |ios| ios.puts DATA1 }
+    File.open(@file2, 'w') { |ios| ios.puts DATA2 }
+    File.open(@file3, 'w') { |ios| ios.puts DATA3 }
+  end
+
+  def teardown
+    @file1.close
+    @file1.unlink
+    @file2.close
+    @file2.unlink
+    @file3.close
+    @file3.unlink
+  end
+
+  test 'SampleReader#read returns correctly' do
+    data = SampleReader.read(@file1, false, false)
+
+    assert_equal(["A\tATCG\tTCCG", "B\tATCG\tTCCC"], data.map(&:to_s))
+  end
+
+  test 'SampleReader#read with index1 reverse_complemented returns correctly' do
+    data = SampleReader.read(@file1, true, false)
+
+    assert_equal(["A\tCGAT\tTCCG", "B\tCGAT\tTCCC"], data.map(&:to_s))
+  end
+
+  test 'SampleReader#read with index2 reverse_complemented returns correctly' do
+    data = SampleReader.read(@file1, false, true)
+
+    assert_equal(["A\tATCG\tCGGA", "B\tATCG\tGGGA"], data.map(&:to_s))
+  end
+
+  test 'SampleReader#read with both reverse_complemented returns correctly' do
+    data = SampleReader.read(@file1, true, true)
+
+    assert_equal(["A\tCGAT\tCGGA", "B\tCGAT\tGGGA"], data.map(&:to_s))
+  end
+
+  test 'SampleReader#read with non-unique sample IDs fails' do
+    assert_raise(SampleReaderError) do
+      capture_stderr { SampleReader.read(@file2, true, true) }
+    end
+  end
+
+  test 'SampleReader#read with non-unique index combination fails' do
+    assert_raise(SampleReaderError) do
+      capture_stderr { SampleReader.read(@file3, true, true) }
+    end
+  end
 end

data/test/test_status.rb

@@ -1,7 +1,61 @@
 #!/usr/bin/env ruby
-$:.unshift File.join(File.dirname(__FILE__), '..')
+$LOAD_PATH.unshift File.join(File.dirname(__FILE__), '..')
 
 require 'test/helper'
 
-class TestStatus < Test::Unit::TestCase 
+# Class for testing the Status class.
+class TestStatus < Test::Unit::TestCase
+  SAMPLES = <<-DATA.gsub(/^\s+\|/, '')
+              |#ID\tIndex1\tIndex2
+              |A\tATCG\tTCCG
+              |B\tATCG\tTCCC
+            DATA
+
+  EXPECTED = <<-YAML.gsub(/^\s+\|/, '')
+               |---
+               |:count: 0
+               |:match: 0
+               |:undetermined: 0
+               |:undetermined_percent: 0.0
+               |:index1_bad_mean: 0
+               |:index2_bad_mean: 0
+               |:index1_bad_min: 0
+               |:index2_bad_min: 0
+               |:sample_ids:
+               |- A
+               |- B
+               |:index1:
+               |- ATCG
+               |:index2:
+               |- TCCC
+               |- TCCG
+               |:time_elapsed: '00:00:00'
+             YAML
+
+  def setup
+    @file_samples = Tempfile.new('samples')
+    @file_status  = Tempfile.new('status')
+
+    File.open(@file_samples, 'w') { |ios| ios.puts SAMPLES }
+
+    @samples = SampleReader.read(@file_samples, false, false)
+    @status  = Status.new(@samples)
+  end
+
+  def teardown
+    @file_samples.close
+    @file_samples.unlink
+    @file_status.close
+    @file_status.unlink
+  end
+
+  test 'Status#to_s returns correctly' do
+    assert_equal(EXPECTED, @status.to_s)
+  end
+
+  test 'Status#save does so correctly' do
+    @status.save(@file_status)
+
+    assert_equal(EXPECTED, File.read(@file_status))
+  end
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: demultiplexer
 version: !ruby/object:Gem::Version
-  version: 0.0.1
+  version: 0.1.0
 platform: ruby
 authors:
 - Martin A. Hansen
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2015-03-22 00:00:00.000000000 Z
+date: 2015-03-24 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: biopieces
@@ -90,7 +90,6 @@ files:
 - test/test_demultiplexer.rb
 - test/test_index_builder.rb
 - test/test_sample_reader.rb
-- test/test_screen.rb
 - test/test_status.rb
 homepage: http://github.com/maasha/demultiplexer
 licenses:
@@ -112,7 +111,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
 requirements: []
 rubyforge_project: demultiplexer
-rubygems_version: 2.4.5
+rubygems_version: 2.2.2
 signing_key: 
 specification_version: 4
 summary: Demultiplexer
@@ -122,6 +121,4 @@ test_files:
 - test/test_demultiplexer.rb
 - test/test_index_builder.rb
 - test/test_sample_reader.rb
-- test/test_screen.rb
 - test/test_status.rb
-has_rdoc: 

data/test/test_screen.rb

@@ -1,7 +0,0 @@
-#!/usr/bin/env ruby
-$:.unshift File.join(File.dirname(__FILE__), '..')
-
-require 'test/helper'
-
-class TestScreen < Test::Unit::TestCase 
-end