parse_fasta 1.4.0 → 1.5.0

checksums.yaml

@@ -1,7 +1,15 @@
 ---
-SHA1:
-  metadata.gz: 689bddd57bc87d882ad8e8711acc84f0f1c33825
-  data.tar.gz: 85d056f1581a23cc25c1a60224fd2c177266ab23
+!binary "U0hBMQ==":
+  metadata.gz: !binary |-
+    YTBkNDczMWQyMjI5YzNkMmEwY2M1MWY3ODRlMzI3ZjE2MThhM2FkMw==
+  data.tar.gz: !binary |-
+    YzNlOGU2Y2FiMDlhMjMxMDkxNTE2Y2NjMzdlOWE2MmUzOWRlMzdiMw==
 SHA512:
-  metadata.gz: f68c1541107d1881177606398c63d09b7da2b04d7e0f30b3165ebd51cc7310caddbcba7788f72db71f21c778c5c1f022382a17e4b450cd172946534665f360a1
-  data.tar.gz: eb8a9bf47345d0ee4b164ef053da5ddb4cb6785d0c0a91424b2e4549d9231d2f55ab79f9355e0df8f924d51be7be738a95f5887b6333aa99c4a5a945941f34f3
+  metadata.gz: !binary |-
+    NGE2NGMyOThlODMzODk0M2ZjMTI2YjI3ZWY4MWVmMTgwODQzNjdhNmI1MjM2
+    NGNkNmFmODEzMTgxMTA1ZWIyN2NlZTBmZjg4NDgzZjEyMjE4ZGY1MDM0ZjBj
+    OGE3NTFiNzE2YjEyNDJlN2VmYjdlYWY1YmUwZTI4OGE3MmE5N2I=
+  data.tar.gz: !binary |-
+    ZjYwMzM2OWQ2MjA2MjE1ZTg4YWUyODdkZWMyODY4NjVhNjk1NDEwNGIwZmMz
+    YTJlM2U5YjFkMWI2MTNiZjEzMGY1MWIyZjQ2ZDZhOWFhZDBkMTk5NzhkMzg3
+    OTlhZDYyN2QxMThhNmU0ZDRiY2U4ODllMmVkNzZlY2YzNjJlMDc=

data/README.md

@@ -29,7 +29,7 @@ and over.
 ## Documentation ##
 
 Checkout
-[parse_fasta docs](http://rubydoc.info/gems/parse_fasta/1.4.0/frames)
+[parse_fasta docs](http://rubydoc.info/gems/parse_fasta/1.5.0/frames)
 to see the full documentation.
 
 ## Usage ##
@@ -58,6 +58,10 @@ Now we can parse fastq files as well!
 
 ## Versions ##
 
+### 1.5.0 ###
+
+Now accepts gzipped files. Huzzah!
+
 ### 1.4.0 ###
 
 Added methods:

data/lib/parse_fasta/fasta_file.rb

@@ -1,4 +1,4 @@
-# Copyright 2014 Ryan Moore
+# Copyright 2014, 2015 Ryan Moore
 # Contact: moorer@udel.edu
 #
 # This file is part of parse_fasta.
@@ -16,23 +16,26 @@
 # You should have received a copy of the GNU General Public License
 # along with parse_fasta.  If not, see <http://www.gnu.org/licenses/>.
 
-# Provides simple interface for parsing fasta format files.
+require 'zlib'
+
+# Provides simple interface for parsing fasta format files. Gzipped
+# files are no problem.
 class FastaFile < File
 
-  # Analagous to File#each_line, #each_record is used to go through a
-  # fasta file record by record.
+  # Analagous to IO#each_line, #each_record is used to go through a
+  # fasta file record by record. It will accept gzipped files as well.
   #
   # @param separate_lines [Object] If truthy, separate lines of record
-  #   into an array, but if falsy, yield a Sequence object for the
-  #   sequence instead.
+  #   into an array of Sequences, but if falsy, yield a Sequence
+  #   object for the sequence instead.
   #
-  # @example Parsing a fasta file (default behavior)
-  #   FastaFile.open('reads.fna', 'r').each_record do |header, sequence|
+  # @example Parsing a fasta file (default behavior, gzip files are fine)
+  #   FastaFile.open('reads.fna.gz').each_record do |header, sequence|
   #     puts [header, sequence.gc].join("\t")
   #   end
   # 
   # @example Parsing a fasta file (with truthy value param)
-  #   FastaFile.open('reads.fna','r').each_record(1) do |header, sequence|
+  #   FastaFile.open('reads.fna').each_record(1) do |header, sequence|
   #     # header => 'sequence_1'
   #     # sequence => ['AACTG', 'AGTCGT', ... ]
   #   end
@@ -43,22 +46,31 @@ class FastaFile < File
   # @yieldparam header [String] The header of the fasta record without
   #   the leading '>'
   #
-  # @yieldparam sequence [Sequence, Array<String>] The sequence of the
+  # @yieldparam sequence [Sequence, Array<Sequence>] The sequence of the
   #   fasta record. If `separate_lines` is falsy (the default
   #   behavior), will be Sequence, but if truthy will be
   #   Array<String>.
   def each_record(separate_lines=nil)
+    begin
+      f = Zlib::GzipReader.open(self)
+    rescue Zlib::GzipFile::Error => e
+      f = self
+    end      
+
     if separate_lines
-      self.each("\n>") do |line|
+      f.each("\n>") do |line|
         header, sequence = parse_line_separately(line)
         yield(header.strip, sequence)
       end
     else
-      self.each("\n>") do |line|
+      f.each("\n>") do |line| 
         header, sequence = parse_line(line)
         yield(header.strip, Sequence.new(sequence))
       end
     end
+
+    f.close if f.instance_of?(Zlib::GzipReader)
+    return f
   end
 
   private
@@ -67,10 +79,11 @@ class FastaFile < File
   end
 
   def parse_line_separately(line)
-    #line.chomp.split("\n", 2).map { |s| s.gsub(/>/, '') }
     header, sequence = 
       line.chomp.split("\n", 2).map { |s| s.gsub(/>/, '') }
-    sequences = sequence.split("\n").reject { |s| s.empty? }
+    sequences = sequence.split("\n")
+      .reject { |s| s.empty? }
+      .map { |s| Sequence.new(s) }
 
     [header, sequences]
   end

data/lib/parse_fasta/fastq_file.rb

@@ -1,4 +1,4 @@
-# Copyright 2014 Ryan Moore
+# Copyright 2014, 2015 Ryan Moore
 # Contact: moorer@udel.edu
 #
 # This file is part of parse_fasta.
@@ -16,15 +16,21 @@
 # You should have received a copy of the GNU General Public License
 # along with parse_fasta.  If not, see <http://www.gnu.org/licenses/>.
 
+require 'zlib'
+
 # Provides simple interface for parsing four-line-per-record fastq
-# format files.
+# format files. Gzipped files are no problem.
 class FastqFile < File
 
-  # Analagous to File#each_line, #each_record is used to go through a
-  # fastq file record by record.
+  # Analagous to IO#each_line, #each_record is used to go through a
+  # fastq file record by record. It will accept gzipped files as well.
   #
   # @example Parsing a fastq file
-  #   FastqFile.open('reads.fq', 'r').each_record do |head, seq, desc, qual|
+  #   FastqFile.open('reads.fq').each_record do |head, seq, desc, qual|
+  #     # do some fun stuff here!
+  #   end
+  # @example Use the same syntax for gzipped files!
+  #   FastqFile.open('reads.fq.gz').each_record do |head, seq, desc, qual|
   #     # do some fun stuff here!
   #   end
   # 
@@ -43,8 +49,14 @@ class FastqFile < File
     sequence = ''
     description = ''
     quality = ''
+
+    begin
+      f = Zlib::GzipReader.open(self)
+    rescue Zlib::GzipFile::Error => e
+      f = self
+    end      
     
-    self.each_line do |line|
+    f.each_line do |line|
       line.chomp!
 
       case count % 4
@@ -61,5 +73,8 @@ class FastqFile < File
       
       count += 1
     end
+    
+    f.close if f.instance_of?(Zlib::GzipReader)
+    return f
   end
 end

data/lib/parse_fasta/quality.rb

@@ -1,4 +1,4 @@
-# Copyright 2014 Ryan Moore
+# Copyright 2014, 2015 Ryan Moore
 # Contact: moorer@udel.edu
 #
 # This file is part of parse_fasta.

data/lib/parse_fasta/sequence.rb

@@ -1,4 +1,4 @@
-# Copyright 2014 Ryan Moore
+# Copyright 2014, 2015 Ryan Moore
 # Contact: moorer@udel.edu
 #
 # This file is part of parse_fasta.

data/lib/parse_fasta/version.rb

@@ -1,4 +1,4 @@
-# Copyright 2014 Ryan Moore
+# Copyright 2014, 2015 Ryan Moore
 # Contact: moorer@udel.edu
 #
 # This file is part of parse_fasta.
@@ -17,5 +17,5 @@
 # along with parse_fasta.  If not, see <http://www.gnu.org/licenses/>.
 
 module ParseFasta
-  VERSION = "1.4.0"
+  VERSION = "1.5.0"
 end

data/lib/parse_fasta.rb

@@ -1,4 +1,4 @@
-# Copyright 2014 Ryan Moore
+# Copyright 2014, 2015 Ryan Moore
 # Contact: moorer@udel.edu
 #
 # This file is part of parse_fasta.

data/parse_fasta.gemspec

@@ -8,8 +8,11 @@ Gem::Specification.new do |spec|
   spec.version       = ParseFasta::VERSION
   spec.authors       = ["Ryan Moore"]
   spec.email         = ["moorer@udel.edu"]
-  spec.summary       = %q{Easy-peasy parsing of fasta files}
-  spec.description   = %q{So you want to parse a fasta or fastq file...}
+  spec.summary       = %q{Easy-peasy parsing of fasta & fastq files!}
+  spec.description   =
+    "Provides nice, programmatic access to fasta " +
+    "and fastq files, as well as providing Sequence and Quality " +
+    "helper classes. No need for BioRuby ;)"
   spec.homepage      = "https://github.com/mooreryan/parse_fasta"
   spec.license       = "GPLv3: http://www.gnu.org/licenses/gpl.txt"
 

data/spec/lib/fasta_file_spec.rb

@@ -1,4 +1,4 @@
-# Copyright 2014 Ryan Moore
+# Copyright 2014, 2015 Ryan Moore
 # Contact: moorer@udel.edu
 #
 # This file is part of parse_fasta.
@@ -20,41 +20,82 @@ require 'spec_helper'
 
 describe FastaFile do
   describe "#each_record" do
+    let(:records) {
+      [["seq1 is fun", "AACTGGNNN"],
+       ["seq2", "AATCCTGNNN"],
+       ["seq3", "yyyyyyyyyyyyyyyNNN"]]
+    }
 
-    let(:fname) { "#{File.dirname(__FILE__)}/../../test_files/test.fa" }
+    let(:truthy_records) {
+      [["seq1 is fun", ["AACTGGNNN"]],
+       ["seq2", ["AAT", "CCTGNNN"]],
+       ["seq3", ["yyyyyyyyyy", "yyyyy", "NNN"]]]
+    }
+    let(:f_handle) { FastaFile.open(@fname).each_record { |s| } }
 
-    context "with no arguments" do 
-      it "yields header and sequence for each record in a fasta file" do
-        seqs = []
-        FastaFile.open(fname, 'r').each_record do |header, sequence|
-          seqs << [header, sequence]
+    shared_examples_for "any FastaFile" do
+      context "with no arguments" do
+        it "yields proper header and sequence for each record" do
+          expect { |b|
+            FastaFile.open(@fname).each_record(&b)
+          }.to yield_successive_args(*records)
         end
-        
-        expect(seqs).to eq([["seq1 is fun", "AACTGGNNN"],
-                            ["seq2", "AATCCTGNNN"],
-                            ["seq3", "yyyyyyyyyyyyyyyNNN"]])
 
+        it "yields the sequence as a Sequence class" do
+          FastaFile.open(@fname).each_record do |_, seq|
+            expect(seq).to be_an_instance_of Sequence
+          end
+        end
       end
 
-      it "yields sequence of type Sequence as second parameter" do
-        FastaFile.open(fname, 'r').each_record do |header, sequence|
-          expect(sequence).to be_an_instance_of Sequence
-          break
+      context "with a truthy argument" do
+        it "yields proper header and sequence for each record" do
+          expect { |b|
+            FastaFile.open(@fname).each_record(1, &b)
+          }.to yield_successive_args(*truthy_records)
         end
-      end      
-    end
 
-    context "with a truthy argument" do
-      it "yields header and array of lines for each record" do
-        seqs = []
-        FastaFile.open(fname, 'r').each_record(1) do |header, sequence|
-          seqs << [header, sequence]
+        it "yields the sequence as a Sequence class" do
+          FastaFile.open(@fname).each_record(1) do |_, seq|
+            all_Sequences = seq.map { |s| s.instance_of?(Sequence) }.all?
+            expect(all_Sequences).to be true
+          end
         end
+        
+      end
+    end
+
+    context "with a gzipped file" do
+      before(:each) do
+        @fname = "#{File.dirname(__FILE__)}/../../test_files/test.fa.gz"
+      end
+
+      it_behaves_like "any FastaFile"
+
+      it "closes the GzipReader" do
+        expect(f_handle).to be_closed
+      end
+
+      it "returns GzipReader object" do
+        expect(f_handle).to be_an_instance_of Zlib::GzipReader
+      end
+    end
+
+    context "with a non-gzipped file" do
+      before(:each) do
+        @fname = "#{File.dirname(__FILE__)}/../../test_files/test.fa"
+      end
+
+      it_behaves_like "any FastaFile"
 
-        expect(seqs).to eq([["seq1 is fun", ["AACTGGNNN"]],
-                            ["seq2", ["AAT", "CCTGNNN"]],
-                            ["seq3", ["yyyyyyyyyy", "yyyyy", "NNN"]]])
+      it "doesn't close the FastqFile (approx regular file behavior)" do
+        expect(f_handle).not_to be_closed
+      end
+
+      it "returns FastaFile object" do
+        expect(f_handle).to be_an_instance_of FastaFile
       end
     end
-  end
+  end      
 end
+

data/spec/lib/fastq_file_spec.rb

@@ -1,4 +1,4 @@
-# Copyright 2014 Ryan Moore
+# Copyright 2014, 2015 Ryan Moore
 # Contact: moorer@udel.edu
 #
 # This file is part of parse_fasta.
@@ -19,32 +19,67 @@
 require 'spec_helper'
 
 describe FastqFile do
-  describe "#each_record" do
-    let(:fname) { "#{File.dirname(__FILE__)}/../../test_files/test.fq" }
-
-    context "with a 4 line per record fastq file" do
-      before do
-        @records = []
-        FastqFile.open(fname, 'r').each_record do |head, seq, desc, qual|
-          @records << [head, seq, desc, qual]
-        end
+  let(:records) {
+    [["seq1", "AACCTTGG", "", ")#3gTqN8"],
+     ["seq2 apples", "ACTG", "seq2 apples", "*ujM"]] }
+  let(:f_handle) { FastqFile.open(@fname).each_record { |s| } }
+
+
+  shared_examples_for "any FastqFile" do
+    it "yields proper header, sequence, description, and quality" do
+      expect { |b|
+        FastqFile.open(@fname).each_record(&b)
+      }.to yield_successive_args(records[0], records[1])
+    end
+
+    it "yields the sequence as a Sequence class" do
+      FastqFile.open(@fname).each_record do |_, seq, _, _|
+        expect(seq).to be_an_instance_of Sequence
       end
+    end
 
-      it "yields the header, sequence, desc, and qual" do
-        expect(@records).to eq([["seq1", "AACCTTGG", "", ")#3gTqN8"],
-                               ["seq2 apples", "ACTG", "seq2 apples",
-                                "*ujM"]])
+    it "yields the quality as a Quality class" do
+      FastqFile.open(@fname).each_record do |_, _, _, qual|
+        expect(qual).to be_an_instance_of Quality
       end
-      
-      it "yields the sequence as a Sequence class" do
-        the_sequence = @records[0][1]
-        expect(the_sequence).to be_a(Sequence)
+    end
+
+  end
+  
+  context "with a 4 line per record fastq file" do
+    describe "#each_record" do
+      context "with a gzipped file" do
+        before(:each) do
+          @fname = "#{File.dirname(__FILE__)}/../../test_files/test.fq.gz"
+        end
+
+        it_behaves_like "any FastqFile"
+
+        it "closes the GzipReader" do
+          expect(f_handle).to be_closed
+        end
+
+        it "returns GzipReader object" do
+          expect(f_handle).to be_an_instance_of Zlib::GzipReader
+        end
       end
 
-      it "yields the quality string as a Quality class" do
-        the_quality = @records[0][3]
-        expect(the_quality).to be_a(Quality)
+      context "with a non-gzipped file" do
+        before(:each) do
+          @fname = "#{File.dirname(__FILE__)}/../../test_files/test.fq"
+        end
+
+        it_behaves_like "any FastqFile"
+
+        it "doesn't close the FastqFile (approx regular file behavior)" do
+          expect(f_handle).not_to be_closed
+        end
+
+        it "returns FastqFile object" do
+          expect(f_handle).to be_an_instance_of FastqFile
+        end
       end
     end
-  end
+  end        
 end
+

data/spec/lib/quality_spec.rb

@@ -1,4 +1,4 @@
-# Copyright 2014 Ryan Moore
+# Copyright 2014, 2015 Ryan Moore
 # Contact: moorer@udel.edu
 #
 # This file is part of parse_fasta.

data/spec/lib/sequence_spec.rb

@@ -1,4 +1,4 @@
-# Copyright 2014 Ryan Moore
+# Copyright 2014, 2015 Ryan Moore
 # Contact: moorer@udel.edu
 #
 # This file is part of parse_fasta.

data/spec/spec_helper.rb

@@ -1,4 +1,4 @@
-# Copyright 2014 Ryan Moore
+# Copyright 2014, 2015 Ryan Moore
 # Contact: moorer@udel.edu
 #
 # This file is part of parse_fasta.

data/test_files/benchmark.rb

@@ -1,6 +1,6 @@
 #!/usr/bin/env ruby
 
-# Copyright 2014 Ryan Moore
+# Copyright 2014, 2015 Ryan Moore
 # Contact: moorer@udel.edu
 
 # This file is part of parse_fasta.

metadata

@@ -1,93 +1,94 @@
 --- !ruby/object:Gem::Specification
 name: parse_fasta
 version: !ruby/object:Gem::Version
-  version: 1.4.0
+  version: 1.5.0
 platform: ruby
 authors:
 - Ryan Moore
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2014-08-13 00:00:00.000000000 Z
+date: 2015-01-08 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - "~>"
+    - - ~>
       - !ruby/object:Gem::Version
         version: '1.6'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - "~>"
+    - - ~>
       - !ruby/object:Gem::Version
         version: '1.6'
 - !ruby/object:Gem::Dependency
   name: rake
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - "~>"
+    - - ~>
       - !ruby/object:Gem::Version
         version: '10.3'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - "~>"
+    - - ~>
       - !ruby/object:Gem::Version
         version: '10.3'
 - !ruby/object:Gem::Dependency
   name: rspec
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - "~>"
+    - - ~>
       - !ruby/object:Gem::Version
         version: '2.14'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - "~>"
+    - - ~>
       - !ruby/object:Gem::Version
         version: '2.14'
 - !ruby/object:Gem::Dependency
   name: bio
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - "~>"
+    - - ~>
       - !ruby/object:Gem::Version
         version: '1.4'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - "~>"
+    - - ~>
       - !ruby/object:Gem::Version
         version: '1.4'
 - !ruby/object:Gem::Dependency
   name: yard
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - "~>"
+    - - ~>
       - !ruby/object:Gem::Version
         version: '0.8'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - "~>"
+    - - ~>
       - !ruby/object:Gem::Version
         version: '0.8'
-description: So you want to parse a fasta or fastq file...
+description: Provides nice, programmatic access to fasta and fastq files, as well
+  as providing Sequence and Quality helper classes. No need for BioRuby ;)
 email:
 - moorer@udel.edu
 executables: []
 extensions: []
 extra_rdoc_files: []
 files:
-- ".gitignore"
+- .gitignore
 - COPYING
 - Gemfile
 - README.md
@@ -106,10 +107,12 @@ files:
 - spec/spec_helper.rb
 - test_files/benchmark.rb
 - test_files/test.fa
+- test_files/test.fa.gz
 - test_files/test.fq
+- test_files/test.fq.gz
 homepage: https://github.com/mooreryan/parse_fasta
 licenses:
-- 'GPLv3: http://www.gnu.org/licenses/gpl.txt'
+- ! 'GPLv3: http://www.gnu.org/licenses/gpl.txt'
 metadata: {}
 post_install_message: 
 rdoc_options: []
@@ -117,12 +120,12 @@ require_paths:
 - lib
 required_ruby_version: !ruby/object:Gem::Requirement
   requirements:
-  - - ">="
+  - - ! '>='
     - !ruby/object:Gem::Version
       version: 1.9.3
 required_rubygems_version: !ruby/object:Gem::Requirement
   requirements:
-  - - ">="
+  - - ! '>='
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
@@ -130,7 +133,7 @@ rubyforge_project:
 rubygems_version: 2.2.2
 signing_key: 
 specification_version: 4
-summary: Easy-peasy parsing of fasta files
+summary: Easy-peasy parsing of fasta & fastq files!
 test_files:
 - spec/lib/fasta_file_spec.rb
 - spec/lib/fastq_file_spec.rb