maiha-css_parser 0.1

data/README

@@ -0,0 +1,19 @@
+css_parser
+==========
+
+  hpricot helper that scrapes html easily by parser class defined css selector
+
+
+Example
+=======
+
+  class UserParser < CssParser
+    css :name, "div#contents span.name"
+    css :age , "div#contents span.age"
+  end
+
+  parser = UserParser.file('user.html')
+  User.new parser.attributes
+
+
+Copyright (c) 2008 maiha@wota.jp, released under the MIT license

data/Rakefile

@@ -0,0 +1,52 @@
+require 'rubygems'
+require 'rake/gempackagetask'
+
+GEM_NAME = "css_parser"
+AUTHOR = "maiha"
+EMAIL = "maiha@wota.jp"
+HOMEPAGE = "http://github.com/maiha/css_parser"
+SUMMARY = "hpricot helper that scrapes html easily by parser class defined css selector"
+GEM_VERSION = "0.1"
+
+spec = Gem::Specification.new do |s|
+#  s.rubyforge_project = 'merb'
+  s.name = GEM_NAME
+  s.version = GEM_VERSION
+  s.platform = Gem::Platform::RUBY
+  s.has_rdoc = true
+  s.extra_rdoc_files = ["README", "LICENSE", 'TODO']
+  s.summary = SUMMARY
+  s.description = s.summary
+  s.author = AUTHOR
+  s.email = EMAIL
+  s.homepage = HOMEPAGE
+  s.add_dependency('hpricot', '>= 0.1')
+  s.add_dependency('dsl_accessor', '>= 0.1')
+  s.require_path = 'lib'
+  s.files = %w(LICENSE README Rakefile TODO) + Dir.glob("{lib,spec,app,public,stubs}/**/*")
+end
+
+Rake::GemPackageTask.new(spec) do |pkg|
+  pkg.gem_spec = spec
+end
+
+desc "Install the gem"
+task :install do
+  Merb::RakeHelper.install(GEM_NAME, :version => GEM_VERSION)
+end
+
+desc "Uninstall the gem"
+task :uninstall do
+  Merb::RakeHelper.uninstall(GEM_NAME, :version => GEM_VERSION)
+end
+
+desc "Create a gemspec file"
+task :gemspec do
+  File.open("#{GEM_NAME}.gemspec", "w") do |file|
+    file.puts spec.to_ruby
+  end
+end
+
+require 'spec/rake/spectask'
+desc 'Default: run spec examples'
+task :default => 'spec'

data/lib/css_parser.rb

@@ -0,0 +1,79 @@
+require 'rubygems'
+require 'dsl_accessor'
+require 'hpricot'
+
+class CssParser
+  dsl_accessor :stored_css, proc{{}}
+
+  ######################################################################
+  ### Exceptions
+
+  class ReservedCss < StandardError; end
+
+  ######################################################################
+  ### InstanceMethods
+
+  def initialize(html = nil, filename = nil)
+    @html     = html.to_s
+    @filename = filename
+  end
+
+  def parser
+    @parser ||= Hpricot(@html)
+  end
+
+  def attributes(keys = nil)
+    keys ||= self.class.my_stored_css.keys
+    keys.inject({}){|h,key| h[key] = send(key); h}
+  end
+
+  ######################################################################
+  ### Class Methods
+
+  def self.file(file)
+    html = NKF.nkf('-w', Pathname(file).read)
+    new(html, file)
+  end
+
+  def self.css(key, pattern)
+    key = key.to_s.intern
+    guard_from_overridden(key)
+    define_css(key, pattern)
+  end
+
+  private
+    def self.css_module
+      @css_module ||= (include (mod = Module.new); mod)
+    end
+
+    # stored_css object for this class
+    def self.my_stored_css
+      @my_stored_css ||= (stored_css.dup rescue stored_css)
+    end
+
+    def self.define_css(key, pattern)
+      # not defined yet
+      unless instance_methods.include?(key.to_s)
+        css_module.module_eval do
+          define_method(key) do
+            pattern = self.class.my_stored_css[key]
+            element = parser.search(pattern).first
+            element ? element.inner_html : nil
+          end
+        end
+      end
+
+      my_stored_css[key] = pattern
+    end
+
+    def self.guard_from_overridden(key)
+      return if my_stored_css.has_key?(key)
+
+      if instance_methods(true).include?(key.to_s)
+        raise ReservedCss, "#{key} is reserved for #{self.to_s.classify}##{key}"
+      end
+      if %w( attributes parser ).include?(key.to_s)
+        raise ReservedCss, "#{key} is reserved for CssParser module"
+      end
+    end
+end

data/spec/css_parser_spec.rb

@@ -0,0 +1,96 @@
+require File.join( File.dirname(__FILE__), "spec_helper" )
+
+describe CssParser do
+  it "should provide .file" do
+    CssParser.should respond_to(:file)
+  end
+
+  describe ".file" do
+    it "should return a CssParser" do
+      CssParser.file(__FILE__).class.should == CssParser
+    end
+  end
+
+  it "should provide .css" do
+    CssParser.should respond_to(:css)
+  end
+
+  describe ".css" do
+    it "should create an instance-level accessor to the argument" do
+      lambda {CssParser.foo2    }.should raise_error(NoMethodError)
+      lambda {CssParser.new.foo2}.should raise_error(NoMethodError)
+      CssParser.css :foo2, "pattern"
+      lambda {CssParser.foo2    }.should raise_error(NoMethodError)
+      lambda {CssParser.new.foo2}.should_not raise_error(NoMethodError)
+    end
+
+    describe " should raise" do
+      it "when an existing instance method is specified" do
+        lambda {
+          CssParser.css :send, "pattern"
+        }.should raise_error(CssParser::ReservedCss)
+      end
+
+      it "when reserved methods are specified" do
+        lambda {
+          CssParser.css :attributes, "pattern"
+        }.should raise_error(CssParser::ReservedCss)
+        lambda {
+          CssParser.css :parser, "pattern"
+        }.should raise_error(CssParser::ReservedCss)
+      end
+
+    end
+
+    it "should parse" do
+      class CssParser
+        css :foo, "div"
+      end
+
+      foo = CssParser.new('<div>maiha</div>')
+      foo.foo.should == "maiha"
+    end
+
+    it "should respect css selector" do
+      class Foo < CssParser
+        css :name, "div.name"
+      end
+
+      foo = Foo.new('<div>xxx</div><div class=name>maiha</div>')
+      foo.name.should == "maiha"
+    end
+
+    it "should define instance method as module" do
+      class CssParser
+        css :foo, "div"
+
+        def foo
+          "[#{super}]"
+        end
+      end
+
+      foo = CssParser.new('<div>a</div>')
+      foo.foo.should == "[a]"
+    end
+  end
+
+  it "should provide #parser" do
+    CssParser.new.should respond_to(:parser)
+  end
+
+  it "should provide #attributes" do
+    CssParser.new.should respond_to(:attributes)
+  end
+
+  describe "#attributes" do
+    it "should return composed hash" do
+      class Foo < CssParser
+        css :name, "#name"
+        css :age , "#age"
+      end
+
+      Foo.new('').attributes.should == {:age=>nil, :name=>nil}
+    end
+  end
+
+end

data/spec/spec_helper.rb

@@ -0,0 +1,6 @@
+require "rubygems"
+require "hpricot"
+require "spec"
+
+require File.dirname(__FILE__) + "/../init"
+

metadata

@@ -0,0 +1,78 @@
+--- !ruby/object:Gem::Specification 
+name: maiha-css_parser
+version: !ruby/object:Gem::Version 
+  version: "0.1"
+platform: ruby
+authors: 
+- maiha
+autorequire: 
+bindir: bin
+cert_chain: []
+
+date: 2009-01-23 00:00:00 -08:00
+default_executable: 
+dependencies: 
+- !ruby/object:Gem::Dependency 
+  name: hpricot
+  version_requirement: 
+  version_requirements: !ruby/object:Gem::Requirement 
+    requirements: 
+    - - ">="
+      - !ruby/object:Gem::Version 
+        version: "0.1"
+    version: 
+- !ruby/object:Gem::Dependency 
+  name: dsl_accessor
+  version_requirement: 
+  version_requirements: !ruby/object:Gem::Requirement 
+    requirements: 
+    - - ">="
+      - !ruby/object:Gem::Version 
+        version: "0.1"
+    version: 
+description: hpricot helper that scrapes html easily by parser class defined css selector
+email: maiha@wota.jp
+executables: []
+
+extensions: []
+
+extra_rdoc_files: 
+- README
+- LICENSE
+- TODO
+files: 
+- LICENSE
+- README
+- Rakefile
+- TODO
+- lib/css_parser.rb
+- spec/spec_helper.rb
+- spec/css_parser_spec.rb
+has_rdoc: true
+homepage: http://github.com/maiha/css_parser
+post_install_message: 
+rdoc_options: []
+
+require_paths: 
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement 
+  requirements: 
+  - - ">="
+    - !ruby/object:Gem::Version 
+      version: "0"
+  version: 
+required_rubygems_version: !ruby/object:Gem::Requirement 
+  requirements: 
+  - - ">="
+    - !ruby/object:Gem::Version 
+      version: "0"
+  version: 
+requirements: []
+
+rubyforge_project: 
+rubygems_version: 1.2.0
+signing_key: 
+specification_version: 2
+summary: hpricot helper that scrapes html easily by parser class defined css selector
+test_files: []
+