docparser 0.1.2 → 0.1.3

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: edbb546c57582191dd0dbb7acc77a4451bdfd9a0
-  data.tar.gz: 6f5e932a6cbaaafc93d9d7abb35ed647055efd46
+  metadata.gz: 907927469491015367a9b5ba12ff4a8122495428
+  data.tar.gz: 5c842a24a58026c8296d61ca95d921f9ab20ccf9
 SHA512:
-  metadata.gz: 433c99cae004c59432928aceb5d3f75e3fc5a9a33eadbd9fb6c7ae7cb359678eaa938304dfce8d4627e56f159ede28af2671e46d33c5357e5e33e7d51a045157
-  data.tar.gz: 89748f88e7ac14f99f75fdccbfcae36e439bdcd1a92e57baab761b1c27856503077a437ee30ea7844b755d3d4afb3e113ca24dd83915779c97124cd808e03bef
+  metadata.gz: 126b0563186b5f7dc9b94a55ee576d3f07818119056c99bd8dd938f940cb5c19b942cdb380ad9f2dc0367383b4e30cf42b8a2468cb9cad734f5cd716e92ce192
+  data.tar.gz: 7abef08de7561f3e8486141c311655bf8f13e1d4c6a658b9a9919c56f0d23fc48c071b6df14211f19f0d3987018d97b739065c01bb5fed267c38f3e86292071e

data/Gemfile

@@ -6,7 +6,7 @@ source 'https://rubygems.org'
 group :test do
   gem 'coveralls', require: false
   gem 'rake'
-  gem 'rubocop'
+  gem 'rubocop', git: 'git://github.com/bbatsov/rubocop.git', ref: '011ca3536c0a'
   gem 'simplecov', require: false
   gem 'simple_mock'
 end

data/Rakefile

@@ -1,5 +1,8 @@
-require "bundler/gem_tasks"
+require 'bundler/gem_tasks'
 require 'rake/testtask'
+require 'rubocop'
+require 'yard'
+YARD::Rake::YardocTask.new
 
 Rake::TestTask.new do |t|
   t.libs << 'lib/docparser'
@@ -10,7 +13,10 @@ end
 task test: :rubocop
 
 task :rubocop do
-  sh 'rubocop'
+  puts "Running Rubocop #{Rubocop::Version::STRING}"
+  args = FileList['**/*.rb', 'Rakefile', 'docparser.gemspec']
+  cli = Rubocop::CLI.new
+  fail unless cli.run(args) == 0
 end
 
-task :default => :test
+task default: :test

data/docparser.gemspec

@@ -3,27 +3,28 @@ $LOAD_PATH.unshift(lib) unless $LOAD_PATH.include?(lib)
 require 'docparser/version'
 
 Gem::Specification.new do |spec|
-  spec.name          = "docparser"
+  spec.name          = 'docparser'
   spec.version       = DocParser::VERSION
-  spec.authors       = ["Jurriaan Pruis"]
-  spec.email         = ["email@jurriaanpruis.nl"]
-  spec.description   = %q{DocParser is a Ruby Gem for webscraping}
-  spec.summary       = %q{DocParser is a Ruby Gem for webscraping}
-  spec.homepage      = "https://github.com/jurriaan/docparser"
-  spec.license       = "MIT"
+  spec.authors       = ['Jurriaan Pruis']
+  spec.email         = ['email@jurriaanpruis.nl']
+  spec.description   = 'DocParser is a Ruby Gem for webscraping'
+  spec.summary       = 'DocParser is a Ruby Gem for webscraping'
+  spec.homepage      = 'https://github.com/jurriaan/docparser'
+  spec.license       = 'MIT'
   spec.platform      = Gem::Platform::RUBY
 
-  spec.files         = `git ls-files`.split($/)
-  spec.executables   = spec.files.grep(%r{^bin/}) { |f| File.basename(f) }
-  spec.test_files    = spec.files.grep(%r{^(test|spec|features)/})
-  spec.require_paths = ["lib"]
+  spec.files         = `git ls-files`.split($RS)
+  spec.executables   = spec.files.grep(/^bin\//) { |f| File.basename(f) }
+  spec.test_files    = spec.files.grep(/^(test|spec|features)\//)
+  spec.require_paths = ['lib']
+  spec.extra_rdoc_files = ['README.md', 'LICENSE']
 
   spec.add_runtime_dependency 'nokogiri', '~> 1.5.9'
   spec.add_runtime_dependency 'parallel', '~> 0.6.4'
   spec.add_runtime_dependency 'axlsx', '~> 1.3.6'
   spec.add_runtime_dependency 'terminal-table', '~> 1.4.5'
   spec.add_runtime_dependency 'pageme', '~> 0.0.3'
-  spec.add_runtime_dependency 'json', '~> 1.7.7'
+  spec.add_runtime_dependency 'multi_json', '~> 1.7'
   spec.add_runtime_dependency 'log4r', '~> 1.1.10'
 
   spec.add_development_dependency 'yard'

data/example.rb

@@ -4,7 +4,6 @@
 #
 $LOAD_PATH.unshift __dir__
 require File.expand_path('lib/docparser.rb', __dir__)
-require 'tmpdir'
 
 include DocParser
 output = MultiOutput.new(filename: 'hackaday')

data/lib/docparser/document.rb

@@ -5,6 +5,10 @@ module DocParser
   # @see Output
   class Document
     attr_reader :filename, :doc, :encoding, :results
+
+    # @return [String] the source of the document
+    attr_reader :html
+
     def initialize(filename: nil, encoding: 'utf-8', parser: nil)
       if encoding == 'utf-8'
         encodingstring = 'r:utf-8'
@@ -37,11 +41,6 @@ module DocParser
       @title ||= xpath_content('//head/title')
     end
 
-    # @return [String] the source of the document
-    def html
-      @html
-    end
-
     # Executes a xpath query
     def xpath(query)
       res = @doc.search(query)

data/lib/docparser/output.rb

@@ -57,6 +57,6 @@ module DocParser
     end
   end
 
-  class MissingHeaderException < StandardError
-  end
+  # MissingHeaderException gets thrown if a required header is missing.
+  class MissingHeaderException < StandardError; end
 end

data/lib/docparser/output/json_output.rb

@@ -1,4 +1,4 @@
-require 'json'
+require 'multi_json'
 module DocParser
   # The JSONOutput class generates a JSON file containing all rows as seperate
   # Array elements
@@ -25,7 +25,7 @@ module DocParser
           @doc[@header[counter]] = ''
         end
       end
-      @file << JSON.dump(@doc)
+      @file << MultiJson.dump(@doc)
     end
 
     def footer

data/lib/docparser/version.rb

@@ -1,4 +1,4 @@
 module DocParser
   # The current version of DocParser
-  VERSION = '0.1.2'
+  VERSION = '0.1.3'
 end

data/test/.rubocop.yml

@@ -0,0 +1,3 @@
+# Do not introduce global variables.
+AvoidGlobalVars:
+  Enabled: false

data/test/lib/docparser/document_test.rb

@@ -110,7 +110,7 @@ describe DocParser::Document do
   it 'should add the row to the results' do
     @test_doc.add_row ['test']
     @test_doc.add_row 'test', 'test2'
-    @test_doc.results.must_equal [[['test'], ['test', 'test2']]]
+    @test_doc.results.must_equal [[%w(test), %w(test test2)]]
   end
 
   it 'should be possible to not use outputs' do

data/test/lib/docparser/output/csv_output_test.rb

@@ -30,7 +30,7 @@ describe DocParser::CSVOutput do
     Dir.mktmpdir do |dir|
       filename = File.join(dir, 'test.csv')
       output = DocParser::CSVOutput.new(filename: filename)
-      output.add_row ['aap', 'noot', 'mies']
+      output.add_row %w(aap noot mies)
       output.add_row ['aap', 'noot', 'mies;']
       output.close
       open(filename).read.must_equal "aap;noot;mies\naap;noot;\"mies;\"\n"
@@ -43,8 +43,8 @@ describe DocParser::CSVOutput do
       output = DocParser::CSVOutput.new(filename: filename)
       output.header = 'test', 'the', 'header'
       output.rowcount.must_equal 0
-      output.add_row ['aap', 'noot', 'mies']
-      output.add_row ['aap', 'noot', 'mies']
+      output.add_row %w(aap noot mies)
+      output.add_row %w(aap noot mies)
       output.rowcount.must_equal 2
     end
   end

data/test/lib/docparser/output/html_output_test.rb

@@ -22,8 +22,8 @@ describe DocParser::HTMLOutput do
       output = DocParser::HTMLOutput.new(filename: filename)
       output.header = 'test', 'the', 'header'
       output.close
-      open(filename).read.must_include '<thead><tr><th>test</th><th>the</th>'\
-                                       '<th>header</th></tr></thead>'
+      open(filename).read.must_include '<thead><tr><th>test</th><th>the</th>
+      <th>header</th></tr></thead>'.gsub(/\s+/, '')
     end
   end
 
@@ -31,7 +31,7 @@ describe DocParser::HTMLOutput do
     Dir.mktmpdir do |dir|
       filename = File.join(dir, 'test.html')
       output = DocParser::HTMLOutput.new(filename: filename)
-      output.add_row ['aap', 'noot', 'mies']
+      output.add_row %w(aap noot mies)
       output.add_row ['aap', 'noot', 'mies;']
       output.close
       html = open(filename).read
@@ -47,8 +47,8 @@ describe DocParser::HTMLOutput do
       output = DocParser::HTMLOutput.new(filename: filename)
       output.header = 'test', 'the', 'header'
       output.rowcount.must_equal 0
-      output.add_row ['aap', 'noot', 'mies']
-      output.add_row ['aap', 'noot', 'mies']
+      output.add_row %w(aap noot mies)
+      output.add_row %w(aap noot mies)
       output.rowcount.must_equal 2
       output.close
       open(filename).read.must_include('<p>2 rows</p>')

data/test/lib/docparser/output/json_output_test.rb

@@ -31,7 +31,7 @@ describe DocParser::JSONOutput do
       filename = File.join(dir, 'test.json')
       output = DocParser::JSONOutput.new(filename: filename)
       -> do
-        output.add_row ['aap', 'noot', 'mies']
+        output.add_row %w(aap noot mies)
       end.must_raise(DocParser::MissingHeaderException)
     end
   end
@@ -41,13 +41,14 @@ describe DocParser::JSONOutput do
       filename = File.join(dir, 'test.json')
       output = DocParser::JSONOutput.new(filename: filename)
       output.header = 'test', 'the', 'header'
-      output.add_row ['a', 'b', 'c']
-      output.add_row ['aap', 'noot', 'mies"']
-      output.add_row ['aap', 'noot'] # testing empty column
+      output.add_row %w(a b c)
+      output.add_row %w(aap noot mies")
+      output.add_row %w(aap noot) # testing empty column
       output.close
-      open(filename).read.must_equal '[{"test":"a","the":"b","header":"c"}'\
-      ',{"test":"aap","the":"noot","header":"mies\""}'\
-      ',{"test":"aap","the":"noot","header":""}]'
+      expected = '[{"test":"a","the":"b","header":"c"},
+        {"test":"aap","the":"noot","header":"mies\""},
+        {"test":"aap","the":"noot","header":""}]'.gsub(/\s+/, '')
+      open(filename).read.must_equal expected
     end
   end
 
@@ -57,8 +58,8 @@ describe DocParser::JSONOutput do
       output = DocParser::JSONOutput.new(filename: filename)
       output.header = 'test', 'the', 'header'
       output.rowcount.must_equal 0
-      output.add_row ['aap', 'noot', 'mies']
-      output.add_row ['aap', 'noot', 'mies']
+      output.add_row %w(aap noot mies)
+      output.add_row %w(aap noot mies)
       output.rowcount.must_equal 2
     end
   end

data/test/lib/docparser/output/multi_output_test.rb

@@ -36,7 +36,7 @@ describe DocParser::MultiOutput do
       filename = File.join(dir, 'test')
       output = DocParser::MultiOutput.new(filename: filename)
       -> do
-        output.add_row ['aap', 'noot', 'mies']
+        output.add_row %w(aap noot mies)
       end.must_raise(DocParser::MissingHeaderException)
     end
   end
@@ -47,8 +47,8 @@ describe DocParser::MultiOutput do
       output = DocParser::MultiOutput.new(filename: filename)
       output.header = 'test', 'the', 'header'
       output.rowcount.must_equal 0
-      output.add_row ['aap', 'noot', 'mies']
-      output.add_row ['aap', 'noot', 'mies']
+      output.add_row %w(aap noot mies)
+      output.add_row %w(aap noot mies)
       output.rowcount.must_equal 2
     end
   end
@@ -57,7 +57,7 @@ describe DocParser::MultiOutput do
     Dir.mktmpdir do |dir|
       filename = File.join(dir, 'test')
       output = DocParser::MultiOutput.new(filename: filename)
-      methods = [:add_row, :header=, :close]
+      methods = %i(add_row header= close)
       outputs = output.instance_variable_get(:@outputs)
       outputs.map! do |o|
         SimpleMock.new o

data/test/lib/docparser/output/nil_output_test.rb

@@ -20,8 +20,8 @@ describe DocParser::NilOutput do
     output = DocParser::NilOutput.new
     output.header = 'test', 'the', 'header'
     output.rowcount.must_equal 0
-    output.add_row ['aap', 'noot', 'mies']
-    output.add_row ['aap', 'noot', 'mies']
+    output.add_row %w(aap noot mies)
+    output.add_row %w(aap noot mies)
     output.rowcount.must_equal 0
   end
 end

data/test/lib/docparser/output/screen_output_test.rb

@@ -20,19 +20,18 @@ describe DocParser::ScreenOutput do
     output = DocParser::ScreenOutput.new
     output.header = 'test', 'the', 'header'
     output.rowcount.must_equal 0
-    output.add_row ['aap', 'noot', 'mies']
-    output.add_row ['aap', 'noot', 'mies']
+    output.add_row %w(aap noot mies)
+    output.add_row %w(aap noot mies)
     output.rowcount.must_equal 2
   end
 
   it 'must have a header' do
     output = DocParser::ScreenOutput.new
     -> do
-      output.add_row ['aap', 'noot', 'mies']
+      output.add_row %w(aap noot mies)
     end.must_raise(DocParser::MissingHeaderException)
   end
 
-
   it 'must output the data after close' do
     $out = StringIO.new
     output = Class.new DocParser::ScreenOutput do
@@ -43,8 +42,8 @@ describe DocParser::ScreenOutput do
       end
     end.new
     output.header = 'test', 'the', 'header'
-    output.add_row ['aap1', '', 'mies']
-    output.add_row ['aap2', 'mies1']
+    output.add_row ['aap1' , '', 'mies']
+    output.add_row %w(aap2 mies1)
     output.close
     out = $out.string
     out.must_include 'header'

data/test/lib/docparser/output/xlsx_output_test.rb

@@ -31,7 +31,7 @@ describe DocParser::XLSXOutput do
     Dir.mktmpdir do |dir|
       filename = File.join(dir, 'test.xlsx')
       output = DocParser::XLSXOutput.new(filename: filename)
-      output.add_row ['aap', 'noot', 'mies']
+      output.add_row %w(aap noot mies)
       output.add_row ['aap', 'noot', 'mies;']
       output.close
       sheet = output.instance_variable_get(:@sheet)
@@ -45,8 +45,8 @@ describe DocParser::XLSXOutput do
       output = DocParser::XLSXOutput.new(filename: filename)
       output.header = 'test', 'the', 'header'
       output.rowcount.must_equal 0
-      output.add_row ['aap', 'noot', 'mies']
-      output.add_row ['aap', 'noot', 'mies']
+      output.add_row %w(aap noot mies)
+      output.add_row %w(aap noot mies)
       output.rowcount.must_equal 2
     end
   end

data/test/lib/docparser/output/yaml_output_test.rb

@@ -31,7 +31,7 @@ describe DocParser::YAMLOutput do
       filename = File.join(dir, 'test.yml')
       output = DocParser::YAMLOutput.new(filename: filename)
       -> do
-        output.add_row ['aap', 'noot', 'mies']
+        output.add_row %w(aap noot mies)
       end.must_raise(DocParser::MissingHeaderException)
     end
   end
@@ -41,9 +41,9 @@ describe DocParser::YAMLOutput do
       filename = File.join(dir, 'test.csv')
       output = DocParser::YAMLOutput.new(filename: filename)
       output.header = 'test', 'the', 'header'
-      output.add_row ['a', 'b', 'c']
-      output.add_row ['aap', 'noot', 'mies"']
-      output.add_row ['aap', 'noot'] # testing empty column
+      output.add_row %w(a b c)
+      output.add_row %w(aap noot mies")
+      output.add_row %w(aap noot) # testing empty column
       output.close
       open(filename).read.must_equal <<-YAMLEND
 ---
@@ -68,8 +68,8 @@ YAMLEND
       output = DocParser::YAMLOutput.new(filename: filename)
       output.header = 'test', 'the', 'header'
       output.rowcount.must_equal 0
-      output.add_row ['aap', 'noot', 'mies']
-      output.add_row ['aap', 'noot', 'mies']
+      output.add_row %w(aap noot mies)
+      output.add_row %w(aap noot mies)
       output.rowcount.must_equal 2
     end
   end

data/test/lib/docparser/output_test.rb

@@ -32,7 +32,7 @@ describe DocParser::Output do
         output.header = 'test', 'the', 'header'
       end
       header = output.instance_variable_get(:@header)
-      header.must_equal ['test', 'the', 'header']
+      header.must_equal %w(test the header)
       $method_id.must_equal :header
     end
   end
@@ -75,7 +75,6 @@ describe DocParser::Output do
   end
 
   it 'should raise a NotImplementedError on write_row' do
-
     Dir.mktmpdir do |dir|
       filename = File.join(dir, 'test.csv')
       output = DocParser::Output.new(filename: filename)

data/test/lib/docparser/parser_test.rb

@@ -162,7 +162,6 @@ describe DocParser::Parser do
     mock_output2.verify.must_equal true
   end
 
-
   it 'should support parallel processing' do
     mock_output = SimpleMock.new DocParser::NilOutput.new
     mock_output.expect :close, nil
@@ -193,5 +192,4 @@ describe DocParser::Parser do
     $method_id.must_equal :fork
     mock_output.verify.must_equal true
   end
-
 end

data/test/test_helper.rb

@@ -5,7 +5,7 @@ SimpleCov.formatter = SimpleCov::Formatter::MultiFormatter[
   Coveralls::SimpleCov::Formatter
 ]
 SimpleCov.start do
-  #add_filter '/test/'
+  # add_filter '/test/'
 end
 require 'minitest/autorun'
 require 'minitest/pride'

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: docparser
 version: !ruby/object:Gem::Version
-  version: 0.1.2
+  version: 0.1.3
 platform: ruby
 authors:
 - Jurriaan Pruis
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2013-04-28 00:00:00.000000000 Z
+date: 2013-05-12 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: nokogiri
@@ -81,19 +81,19 @@ dependencies:
       - !ruby/object:Gem::Version
         version: 0.0.3
 - !ruby/object:Gem::Dependency
-  name: json
+  name: multi_json
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - ~>
       - !ruby/object:Gem::Version
-        version: 1.7.7
+        version: '1.7'
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - ~>
       - !ruby/object:Gem::Version
-        version: 1.7.7
+        version: '1.7'
 - !ruby/object:Gem::Dependency
   name: log4r
   requirement: !ruby/object:Gem::Requirement
@@ -127,7 +127,9 @@ email:
 - email@jurriaanpruis.nl
 executables: []
 extensions: []
-extra_rdoc_files: []
+extra_rdoc_files:
+- README.md
+- LICENSE
 files:
 - .coveralls.yml
 - .gitignore
@@ -153,6 +155,7 @@ files:
 - lib/docparser/output/yaml_output.rb
 - lib/docparser/parser.rb
 - lib/docparser/version.rb
+- test/.rubocop.yml
 - test/lib/docparser/blackbox_test.rb
 - test/lib/docparser/document_test.rb
 - test/lib/docparser/logging_test.rb
@@ -238,6 +241,7 @@ signing_key:
 specification_version: 4
 summary: DocParser is a Ruby Gem for webscraping
 test_files:
+- test/.rubocop.yml
 - test/lib/docparser/blackbox_test.rb
 - test/lib/docparser/document_test.rb
 - test/lib/docparser/logging_test.rb