pseudoxml 0.1.0

data/RAKEFILE

@@ -0,0 +1,66 @@
+#!/usr/bin/env ruby
+
+require 'rubygems'
+require 'rake'
+require 'rake/rdoctask'
+require 'rake/gempackagetask'
+require 'rake/contrib/rubyforgepublisher'
+require 'pscp'
+
+PACKAGE_VERSION = '0.1.0'
+
+PACKAGE_FILES = FileList[
+  'README',
+  'CHANGELOG',
+  'RAKEFILE',
+  'lib/**/*.rb'
+].to_a
+
+PROJECT = 'pseudoxml'
+
+ENV['RUBYFORGE_USER'] = "ssmoot@rubyforge.org"
+ENV['RUBYFORGE_PROJECT'] = "/var/www/gforge-projects/#{PROJECT}"
+
+task :default => [:rdoc]
+
+desc 'Generate Documentation'
+rd = Rake::RDocTask.new do |rdoc|
+  rdoc.rdoc_dir = 'doc'
+  rdoc.title = "PseudoXml -- A not-quite Xml parser in pure-ruby"
+  rdoc.options << '--line-numbers' << '--inline-source' << '--main' << 'README'
+  rdoc.rdoc_files.include(PACKAGE_FILES)
+end
+
+gem_spec = Gem::Specification.new do |s|
+  s.platform = Gem::Platform::RUBY
+  s.name = PROJECT
+  s.summary = "PseudoXml -- A not-quite Xml parser in pure-ruby"
+  s.description = "Fast, low memory requirements for the streaming version, but no CDATA or Attribute support"
+  s.version = PACKAGE_VERSION
+  
+  s.authors = 'Sam Smoot', 'Scott Bauer'
+  s.email = 'ssmoot@gmail.com; bauer.mail@gmail.com'
+  s.rubyforge_project = PROJECT
+  s.homepage = 'http://substantiality.net'
+  
+  s.files = PACKAGE_FILES
+  
+  s.require_path = 'lib'
+  s.requirements << 'none'
+  s.autorequire = 'rexchange'
+  
+  s.has_rdoc = true
+  s.rdoc_options << '--line-numbers' << '--inline-source' << '--main' << 'README'
+  s.extra_rdoc_files = rd.rdoc_files.reject { |fn| fn =~ /\.rb$/ }.to_a
+end
+
+Rake::GemPackageTask.new(gem_spec) do |p|
+  p.gem_spec = gem_spec
+  p.need_tar = true
+  p.need_zip = true
+end
+
+desc "Publish RDOC to RubyForge"
+task :rubyforge => [:rdoc, :gem] do
+  Rake::SshDirPublisher.new(ENV['RUBYFORGE_USER'], ENV['RUBYFORGE_PROJECT'], 'doc').upload
+end

data/lib/pseudo_xml.rb

@@ -0,0 +1,5 @@
+#!/usr/bin/env ruby
+
+require 'pseudo_xml/document'
+require 'pseudo_xml/simple_query'
+require 'pseudo_xml/xpath_query'

data/lib/pseudo_xml/collector.rb

@@ -0,0 +1,30 @@
+module PseudoXml
+
+  class Collector
+    
+    def initialize
+      @parsed_elements = []
+    end
+    
+    def start_element(element_text)      
+      @parsed_elements.push Element.new(element_text[/\w+/])
+    end
+    
+    def end_element
+      if @parsed_elements.size > 1
+        node = @parsed_elements.pop
+        @parsed_elements.last.elements << node
+      end
+    end
+    
+    def add_text(text)
+      if text.size > 0 && @parsed_elements.size > 0
+        @parsed_elements.last.text_nodes << text
+      end
+    end
+    
+    def nodes
+      @parsed_elements
+    end
+  end
+end

data/lib/pseudo_xml/document.rb

@@ -0,0 +1,26 @@
+require 'pseudo_xml/xml_string'
+require 'pseudo_xml/collector'
+require 'pseudo_xml/element'
+
+module PseudoXml
+  class Document
+
+    def initialize(xml)
+      parser = XmlString.new(xml)
+      collector = Collector.new
+      
+      parser.parse(collector)
+      
+      @root = collector.nodes.first
+    end
+
+    def root
+      @root
+    end
+    
+    def count_elements
+    	return 0 unless @root
+      @root.count
+    end
+  end
+end

data/lib/pseudo_xml/element.rb

@@ -0,0 +1,37 @@
+module PseudoXml
+  
+  class Element
+    
+    attr_reader :name
+    
+    def initialize(name)
+      @name = name.strip
+    end
+    
+    def text_nodes
+      @text_nodes ||= []
+    end
+    
+    def text
+      self.text_nodes.join($/)
+    end
+    
+    def elements
+      @elements ||= []
+    end
+    
+    include Enumerable
+    
+    def each
+      self.elements.each { |node| yield node }
+    end
+    
+    def to_s
+      @name
+    end
+    
+    def count
+      self.elements.inject(1) { |sum, node| sum += node.count }
+    end
+  end
+end

data/lib/pseudo_xml/simple_query.rb

@@ -0,0 +1,39 @@
+require 'pseudo_xml/xml_string'
+require 'pseudo_xml/collector'
+
+module PseudoXml
+  class SimpleQuery < Collector
+    
+    protected :initialize
+    
+    def self.find(xml, element_name, &b)
+      parser = XmlString.new(xml)
+      collector = self.new(element_name, &b)
+      
+      parser.parse(collector)
+    end
+    
+    def initialize(element_name, &b)
+      @parsed_elements = []
+      @find_element_name = element_name
+      @block = b
+    end
+    
+    def start_element(element_text)
+      element_name = element_text[/\w+/]
+      
+      if @parsed_elements.size > 1 || element_name == @find_element_name
+        @parsed_elements.push Element.new(element_text[/\w+/])
+      end
+    end
+
+    def end_element
+      if @parsed_elements.size > 1
+        node = @parsed_elements.pop
+        @parsed_elements.last.elements << node
+      elsif @parsed_elements.size == 1
+        @block.call(@parsed_elements.pop)
+      end      
+    end
+  end
+end

data/lib/pseudo_xml/xml_string.rb

@@ -0,0 +1,51 @@
+module PseudoXml
+  
+  class XmlString
+    def initialize(xml)
+      @xml = xml
+      reset!
+    end
+    
+    def reset!
+      @offset_start = @offset_end = skip_prolog
+    end
+    
+    def skip_prolog
+      if @xml =~ /^(\<\?.*?\?\>(\r?\n)*)/
+        $1.size
+      else
+      	0
+      end
+    end
+    
+    def parse(collector)
+      until(eof?) do
+        @offset_start = @xml.index('<', @offset_end)
+      
+        return nil unless @offset_start
+      
+        if @offset_start > @offset_end + 1
+          collector.add_text @xml[@offset_end + 1...@offset_start]
+        end
+      
+        @offset_end = @xml.index('>', @offset_start)
+      
+        return nil unless @offset_end
+      
+        element_text = @xml[@offset_start..@offset_end]
+      
+        if element_text[1,1] == '/'
+          collector.end_element
+        else
+          collector.start_element element_text
+        end
+      end
+    end
+    
+    def eof?
+      @offset_start.nil? || @offset_end.nil?
+    end
+    
+  end
+  
+end

data/lib/pseudo_xml/xpath_query.rb

@@ -0,0 +1,55 @@
+require 'pseudo_xml/xml_string'
+require 'pseudo_xml/collector'
+require 'pseudo_xml/simple_query'
+
+module PseudoXml
+  class XPathQuery < Collector
+    
+    protected :initialize
+    
+    def self.find(xml, xpath, &b)
+      parser = XmlString.new(xml)
+      collector = self.new(xpath, &b)
+      
+      parser.parse(collector)
+    end
+    
+    def initialize(xpath, &b)
+      @parsed_elements = []
+      @current_path = []
+      @xpath = xpath.split('/')
+      @xpath.shift if @xpath.first == ''
+      @block = b
+    end
+    
+    def start_element(element_text)
+      element_name = element_text[/\w+/]
+      
+      @current_path.push element_name
+      
+      if @parsed_elements.size > 1 || xpath_match?
+        @parsed_elements.push Element.new(element_text[/\w+/])
+      end
+    end
+    
+    def xpath_match?
+      @xpath.size.times do |i|
+        return false unless @xpath[i] == '*' || @xpath[i] == @current_path[i]
+      end
+      
+      return true
+    end
+    
+    def end_element
+      
+      @current_path.pop
+      
+      if @parsed_elements.size > 1
+        node = @parsed_elements.pop
+        @parsed_elements.last.elements << node
+      elsif @parsed_elements.size == 1
+        @block.call(@parsed_elements.pop)
+      end      
+    end
+  end
+end

metadata

@@ -0,0 +1,56 @@
+--- !ruby/object:Gem::Specification 
+rubygems_version: 0.8.10
+specification_version: 1
+name: pseudoxml
+version: !ruby/object:Gem::Version 
+  version: 0.1.0
+date: 2006-02-01
+summary: "PseudoXml -- A not-quite Xml parser in pure-ruby"
+require_paths: 
+  - lib
+email: ssmoot@gmail.com; bauer.mail@gmail.com
+homepage: http://substantiality.net
+rubyforge_project: pseudoxml
+description: "Fast, low memory requirements for the streaming version, but no CDATA or
+  Attribute support"
+autorequire: rexchange
+default_executable: 
+bindir: bin
+has_rdoc: true
+required_ruby_version: !ruby/object:Gem::Version::Requirement 
+  requirements: 
+    - 
+      - ">"
+      - !ruby/object:Gem::Version 
+        version: 0.0.0
+  version: 
+platform: ruby
+authors: 
+  - Sam Smoot
+  - Scott Bauer
+files: 
+  - README
+  - CHANGELOG
+  - RAKEFILE
+  - lib/pseudo_xml.rb
+  - lib/pseudo_xml/collector.rb
+  - lib/pseudo_xml/document.rb
+  - lib/pseudo_xml/element.rb
+  - lib/pseudo_xml/simple_query.rb
+  - lib/pseudo_xml/xml_string.rb
+  - lib/pseudo_xml/xpath_query.rb
+test_files: []
+rdoc_options: 
+  - "--line-numbers"
+  - "--inline-source"
+  - "--main"
+  - README
+extra_rdoc_files: 
+  - README
+  - CHANGELOG
+  - RAKEFILE
+executables: []
+extensions: []
+requirements: 
+  - none
+dependencies: []