parse_fasta 1.9.2 → 2.0.0

data/spec/parse_fasta/seq_file_spec.rb

@@ -0,0 +1,238 @@
+# Copyright 2014 - 2016 Ryan Moore
+# Contact: moorer@udel.edu
+#
+# This file is part of parse_fasta.
+#
+# parse_fasta is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# parse_fasta is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# 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 "spec_helper"
+
+module ParseFasta
+  describe SeqFile do
+    let(:test_dir) {
+      File.join File.dirname(__FILE__), "..", "test_files"
+    }
+
+    let(:fasta) {
+      File.join test_dir, "seqs.fa"
+    }
+    let(:fasta_gz) {
+      File.join test_dir, "seqs.fa.gz"
+    }
+    let(:fastq) {
+      File.join test_dir, "seqs.fq"
+    }
+    let(:fastq_gz) {
+      File.join test_dir, "seqs.fq.gz"
+    }
+
+    let(:fasta_records) {
+      [Record.new(header: "empty seq at beginning",
+                  seq: ""),
+       Record.new(header: "seq1 is fun",
+                  seq: "AACTGGNNN"),
+       Record.new(header: "seq2",
+                  seq: "AATCCTGNNN"),
+       Record.new(header: "empty seq 1",
+                  seq: ""),
+       Record.new(header: "empty seq 2",
+                  seq: ""),
+       Record.new(header: "seq3",
+                  seq: "yyyyyyyyyyyyyyyNNN"),
+       Record.new(header: "seq 4 > has many '>' in header",
+                  seq: "ACTGactg"),
+       Record.new(header: "empty seq at end",
+                  seq: "")]
+    }
+    let(:fastq_records) {
+      [Record.new(header: "seq1",
+                  seq: "AA CC TT GG",
+                  desc: "",
+                  qual: ")# 3g Tq N8"),
+       Record.new(header: "seq2 @pples",
+                  seq:    "ACTG",
+                  desc:   "seq2 +pples",
+                  qual:   "*ujM")]
+    }
+
+    # to test the line endings
+    let(:line_endings_fastq_records) {
+      [Record.new(header: "apple", seq: "ACTG", desc: "", qual: "IIII"),
+       Record.new(header: "pie",   seq: "AACC", desc: "", qual: "BBBB"),]
+    }
+    let(:line_endings_fasta_records) {
+      [Record.new(header: "apple", seq: "ACTG"),
+       Record.new(header: "pie",   seq: "AACC"),]
+    }
+
+
+    describe "::open" do
+      context "when the file doesn't exist" do
+        it "raises FileNotFoundError" do
+          expect { SeqFile.open "arstoien" }.
+              to raise_error ParseFasta::Error::FileNotFoundError
+        end
+      end
+
+      context "when input looks like neither fastA or fastQ" do
+        it "raises a DataFormatError" do
+          fname = File.join test_dir, "not_a_seq_file.txt"
+
+          expect { SeqFile.open(fname) }.
+              to raise_error ParseFasta::Error::DataFormatError
+        end
+      end
+
+      context "when input looks like fastA" do
+        it "sets @type to :fasta" do
+          expect(SeqFile.open(fasta).type).to eq :fasta
+        end
+
+        it "sets @type to :fasta (gzipped)" do
+          expect(SeqFile.open(fasta_gz).type).to eq :fasta
+        end
+      end
+
+      context "when input looks like fastQ" do
+        it "sets @type to :fastq" do
+          expect(SeqFile.open(fastq).type).to eq :fastq
+        end
+
+        it "sets @type to :fastq (gzipped)" do
+          expect(SeqFile.open(fastq_gz).type).to eq :fastq
+        end
+      end
+
+      it "returns a SeqFile" do
+        expect(SeqFile.open fasta).to be_a SeqFile
+      end
+    end
+
+    describe "#each_record" do
+      shared_examples "it yields the records" do
+        it "yields the records" do
+          expect { |b| SeqFile.open(fname).each_record &b  }.
+              to yield_successive_args(*records)
+        end
+      end
+
+      context "input is fastA" do
+        context "with gzipped fastA" do
+          let(:fname) { File.join test_dir, "seqs.fa.gz" }
+          let(:records) { fasta_records }
+
+          include_examples "it yields the records"
+        end
+
+        context "with gzipped fastA with multiple blobs" do
+          # e.g., $ gzip -c a.fa > c.fa.gz; gzip -c b.fa >> c.fa.gz
+          let(:fname) { File.join test_dir, "multi_blob.fa.gz" }
+          let(:records) { fasta_records + fasta_records }
+
+          include_examples "it yields the records"
+        end
+
+        context "with non-gzipped fastA" do
+          let(:fname) { File.join test_dir, "seqs.fa" }
+          let(:records) { fasta_records }
+
+
+          include_examples "it yields the records"
+        end
+      end
+
+      context "input is fastQ" do
+        context "with gzipped fastQ" do
+          let(:fname) { File.join test_dir, "seqs.fq.gz" }
+          let(:records) { fastq_records }
+
+          include_examples "it yields the records"
+        end
+
+        context "with gzipped fastQ with multiple blobs" do
+          # e.g., $ gzip -c a.fq > c.fq.gz; gzip -c b.fq >> c.fq.gz
+          let(:fname) { File.join test_dir, "multi_blob.fq.gz" }
+          let(:records) { fastq_records + fastq_records }
+
+          include_examples "it yields the records"
+        end
+
+        context "with non-gzipped fastQ" do
+          let(:fname) { File.join test_dir, "seqs.fq" }
+          let(:records) { fastq_records }
+
+          include_examples "it yields the records"
+        end
+      end
+
+      context "handles non newline line endings" do
+        context "fastQ, non-gz, carriage return only" do
+          let(:fname) { File.join test_dir, "cr.fq" }
+          let(:records) { line_endings_fastq_records }
+
+          include_examples "it yields the records"
+        end
+
+        context "fastQ, gz, carriage return only" do
+          let(:fname) { File.join test_dir, "cr.fq.gz" }
+          let(:records) { line_endings_fastq_records }
+
+          include_examples "it yields the records"
+        end
+
+        context "fastQ, non-gz, carriage return and newline" do
+          let(:fname) { File.join test_dir, "cr_nl.fq" }
+          let(:records) { line_endings_fastq_records }
+
+          include_examples "it yields the records"
+        end
+
+        context "fastQ, gz, carriage return and newline" do
+          let(:fname) { File.join test_dir, "cr_nl.fq.gz" }
+          let(:records) { line_endings_fastq_records }
+
+          include_examples "it yields the records"
+        end
+
+        context "fastA, non-gz, carriage return only" do
+          let(:fname) { File.join test_dir, "cr.fa" }
+          let(:records) { line_endings_fasta_records }
+
+          include_examples "it yields the records"
+        end
+
+        context "fastA, gz, carriage return only" do
+          let(:fname) { File.join test_dir, "cr.fa.gz" }
+          let(:records) { line_endings_fasta_records }
+
+          include_examples "it yields the records"
+        end
+
+        context "fastA, non-gz, carriage return and newline" do
+          let(:fname) { File.join test_dir, "cr_nl.fa" }
+          let(:records) { line_endings_fasta_records }
+
+          include_examples "it yields the records"
+        end
+
+        context "fastA, gz, carriage return and newline" do
+          let(:fname) { File.join test_dir, "cr_nl.fa.gz" }
+          let(:records) { line_endings_fasta_records }
+
+          include_examples "it yields the records"
+        end
+      end
+    end
+  end
+end

data/spec/parse_fasta_spec.rb

@@ -0,0 +1,25 @@
+# Copyright 2014 - 2016 Ryan Moore
+# Contact: moorer@udel.edu
+#
+# This file is part of parse_fasta.
+#
+# parse_fasta is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# parse_fasta is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# 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 'spec_helper'
+
+describe ParseFasta do
+  it 'has a version number' do
+    expect(ParseFasta::VERSION).not_to be nil
+  end
+end

data/spec/spec_helper.rb

@@ -1,4 +1,4 @@
-# Copyright 2014, 2015 Ryan Moore
+# Copyright 2014 - 2016 Ryan Moore
 # Contact: moorer@udel.edu
 #
 # This file is part of parse_fasta.
@@ -19,47 +19,5 @@
 require 'coveralls'
 Coveralls.wear!
 
+$LOAD_PATH.unshift File.expand_path('../../lib', __FILE__)
 require 'parse_fasta'
-
-module Helpers
-
-  RECORDS = [["empty seq at beginning", ""],
-             ["seq1 is fun", "AACTGGNNN"],
-             ["seq2", "AATCCTGNNN"],
-             ["empty seq 1", ""],
-             ["empty seq 2", ""],
-             ["seq3", "yyyyyyyyyyyyyyyNNN"],
-             ["seq 4 > has many '>' in header", "ACTGactg"],
-             ["empty seq at end", ""]]
-
-  RECORDS_FAST = [["empty seq at beginning", ""],
-                  ["seq1 is fun", "AAC TGG NN N"],
-                  ["seq2", "AATCCTGNNN"],
-                  ["empty seq 1", ""],
-                  ["empty seq 2", ""],
-                  ["seq3", "yyyyyyyyyyyyyyyNNN"],
-                  ["seq 4 > has many '>' in header", "ACTGactg"],
-                  ["empty seq at end", ""]]
-
-  RECORDS_MAP = {
-    "empty seq at beginning" => "",
-    "seq1 is fun" => "AACTGGNNN",
-    "seq2" => "AATCCTGNNN",
-    "empty seq 1" => "",
-    "empty seq 2" => "",
-    "seq3" => "yyyyyyyyyyyyyyyNNN",
-    "seq 4 > has many '>' in header" => "ACTGactg",
-    "empty seq at end" => ""
-  }
-
-  TRUTHY_RECORDS = [["empty seq at beginning", []],
-                    ["seq1 is fun", ["AACTGGNNN"]],
-                    ["seq2", ["AAT", "CCTGNNN"]],
-                    ["empty seq 1", []],
-                    ["empty seq 2", []],
-                    ["seq3", ["yyyyyyyyyy", "yyyyy", "NNN"]],
-                    ["seq 4 > has many '>' in header", ["ACTG" ,
-                                                        "actg"]],
-                    ["empty seq at end", []]]
-
-end

data/spec/test_files/cr.fa

@@ -0,0 +1 @@
+>apple

data/spec/test_files/cr.fq

@@ -0,0 +1 @@
+@apple
+
+

data/spec/test_files/cr_nl.fa

@@ -0,0 +1,4 @@
+>apple
+ACTG
+>pie
+AACC

data/spec/test_files/cr_nl.fq

@@ -0,0 +1,8 @@
+@apple
+ACTG
++
+IIII
+@pie
+AACC
++
+BBBB

data/spec/test_files/not_a_seq_file.txt

@@ -0,0 +1 @@
+this is not a seq file

data/spec/test_files/seqs.fq

@@ -0,0 +1,8 @@
+@seq1
+AA       CC TT GG
++
+)# 3g Tq    N8
+@seq2 @pples
+ACTG
++seq2 +pples
+*ujM

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: parse_fasta
 version: !ruby/object:Gem::Version
-  version: 1.9.2
+  version: 2.0.0
 platform: ruby
 authors:
 - Ryan Moore
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2016-05-20 00:00:00.000000000 Z
+date: 2016-10-21 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
@@ -112,39 +112,50 @@ description: Provides nice, programmatic access to fasta and fastq files, as wel
   as providing Sequence and Quality helper classes. No need for BioRuby ;)
 email:
 - moorer@udel.edu
-executables: []
+executables:
+- console
+- setup
 extensions: []
 extra_rdoc_files: []
 files:
 - .coveralls.yml
 - .gitignore
+- .rspec
 - .travis.yml
 - .yardopts
+- CHANGELOG.md
 - COPYING
 - Gemfile
 - README.md
 - Rakefile
+- bin/console
+- bin/setup
 - lib/parse_fasta.rb
-- lib/parse_fasta/fasta_file.rb
-- lib/parse_fasta/fastq_file.rb
-- lib/parse_fasta/quality.rb
+- lib/parse_fasta/error.rb
+- lib/parse_fasta/record.rb
 - lib/parse_fasta/seq_file.rb
-- lib/parse_fasta/sequence.rb
 - lib/parse_fasta/version.rb
 - parse_fasta.gemspec
-- spec/lib/fasta_file_spec.rb
-- spec/lib/fastq_file_spec.rb
-- spec/lib/quality_spec.rb
-- spec/lib/seq_file_spec.rb
-- spec/lib/sequence_spec.rb
+- spec/parse_fasta/record_spec.rb
+- spec/parse_fasta/seq_file_spec.rb
+- spec/parse_fasta_spec.rb
 - spec/spec_helper.rb
-- test_files/bad.fa
-- test_files/benchmark.rb
-- test_files/bogus.txt
-- test_files/test.fa
-- test_files/test.fa.gz
-- test_files/test.fq
-- test_files/test.fq.gz
+- spec/test_files/cr.fa
+- spec/test_files/cr.fa.gz
+- spec/test_files/cr.fq
+- spec/test_files/cr.fq.gz
+- spec/test_files/cr_nl.fa
+- spec/test_files/cr_nl.fa.gz
+- spec/test_files/cr_nl.fq
+- spec/test_files/cr_nl.fq.gz
+- spec/test_files/multi_blob.fa.gz
+- spec/test_files/multi_blob.fq.gz
+- spec/test_files/not_a_seq_file.txt
+- spec/test_files/poorly_catted.fa
+- spec/test_files/seqs.fa
+- spec/test_files/seqs.fa.gz
+- spec/test_files/seqs.fq
+- spec/test_files/seqs.fq.gz
 homepage: https://github.com/mooreryan/parse_fasta
 licenses:
 - ! 'GPLv3: http://www.gnu.org/licenses/gpl.txt'
@@ -170,10 +181,24 @@ signing_key:
 specification_version: 4
 summary: Easy-peasy parsing of fasta & fastq files!
 test_files:
-- spec/lib/fasta_file_spec.rb
-- spec/lib/fastq_file_spec.rb
-- spec/lib/quality_spec.rb
-- spec/lib/seq_file_spec.rb
-- spec/lib/sequence_spec.rb
+- spec/parse_fasta/record_spec.rb
+- spec/parse_fasta/seq_file_spec.rb
+- spec/parse_fasta_spec.rb
 - spec/spec_helper.rb
+- spec/test_files/cr.fa
+- spec/test_files/cr.fa.gz
+- spec/test_files/cr.fq
+- spec/test_files/cr.fq.gz
+- spec/test_files/cr_nl.fa
+- spec/test_files/cr_nl.fa.gz
+- spec/test_files/cr_nl.fq
+- spec/test_files/cr_nl.fq.gz
+- spec/test_files/multi_blob.fa.gz
+- spec/test_files/multi_blob.fq.gz
+- spec/test_files/not_a_seq_file.txt
+- spec/test_files/poorly_catted.fa
+- spec/test_files/seqs.fa
+- spec/test_files/seqs.fa.gz
+- spec/test_files/seqs.fq
+- spec/test_files/seqs.fq.gz
 has_rdoc: