sosowa 0.0.2 → 0.1

data/CHANGELOG.md

@@ -0,0 +1,6 @@
+## 0.1
+* Sosowa#searchを追加
+* Sosowa::Novel#plainを使って<br>タグや改行コードが取り除かれたテキストを得ることが出来ます。
+
+## 0.0.2
+* 最初のリリースです！

data/README.md

@@ -2,6 +2,11 @@
 
 Sosowa Parser for Ruby 1.9.x
 
+## Requirements
+
+* Ruby 1.9.x
+* mechanize gem
+
 ## Installation
 
 	gem install sosowa
@@ -17,6 +22,9 @@ Sosowa Parser for Ruby 1.9.x
 	# 作品集番号156の1320873807を持ってくる
 	novel = Sosowa.get(:log => 156, :key => 1320873807)
 	puts novel.text
+	
+	# "ナズーリン"がタイトルに含まれているSSの数を出力する
+	puts Sosowa.search("ナズーリン", :type => :title).size
 
 ## Contributing
 

data/lib/sosowa/parser.rb

@@ -6,8 +6,18 @@ module Sosowa
       @agent.user_agent = "Sosowa Ruby #{Sosowa::VERSION}"
     end
     
+    def search(query, args={})
+      params = Sosowa.serialize_parameter({:mode => :search, :type => (args[:type] ? args[:type] : :insubject), :query => query.tosjis})
+      parse_index(URI.join(Sosowa::BASE_URL, params))
+    end
+    
     def fetch_index(log)
-      page = @agent.get("#{Sosowa::BASE_URL}/?log=#{log}")
+      params = Sosowa.serialize_parameter({:log => log})
+      parse_index(URI.join(Sosowa::BASE_URL, params))
+    end
+    
+    def parse_index(url)
+      page = @agent.get(url)
       indexes = []
       tr = page.search("tr")
       tr = tr[1, tr.size-1]
@@ -19,6 +29,7 @@ module Sosowa
         else
           title = tr.search(%{td[@class="title cell_title"] > a}).inner_html.to_s.toutf8.strip
           tags = tr.search(%{td[@class="title cell_title"] > a})[0].attributes["title"].value.split(" / ")
+          log = tr.search(%{td[@class="title cell_title"] > a})[0].attributes["href"].value.gsub(/log=(\d+)$/, '\1').to_i
           key = tr.search(%{td[@class="title cell_title"] > a})[0].attributes["href"].value.gsub(/^.+key=(.+?)&.+$/, '\1').to_i
           author = tr.search(%{td[@class="cell_author"]}).inner_html.to_s.toutf8.strip
           created_at = Time.parse(tr.search(%{td[@class="cell_created"]}).inner_html.to_s.toutf8.strip)

data/lib/sosowa/scheme.rb

@@ -22,7 +22,8 @@ module Sosowa
     end
     
     def fetch(log, key)
-      @page =  @agent.get("#{Sosowa::BASE_URL}/?mode=read&log=#{log}&key=#{key}")
+      params = Sosowa.serialize_parameter({:mode => :read, :log => log, :key => key})
+      @page = @agent.get(URI.join(Sosowa::BASE_URL, params))
       tags = (@page/%{dl[@class="info"][1] > dd > a}).map{|t| t.inner_html.to_s.toutf8 }
       text = (@page/%{div[@class="contents ss"]})[0].inner_html.to_s.toutf8
       ps = (@page/%{div[@class="aft"]})[0].inner_html.to_s.toutf8
@@ -78,6 +79,10 @@ module Sosowa
       form.field_with(:name => "point").option_with(:value => (params[:point].to_s || "0")).select
       form.click_button
     end
+    
+    def plain
+      return @element[:text].gsub(/(<br>|\r?\n)/, "")
+    end
   end
   
   class Comment < Scheme

data/lib/sosowa/version.rb

@@ -1,3 +1,3 @@
 module Sosowa
-  VERSION = "0.0.2"
-end
+  VERSION = "0.1"
+end

data/lib/sosowa.rb

@@ -1,14 +1,27 @@
-$LOAD_PATH.unshift(File.expand_path("../", __FILE__))
 require "kconv"
 require "mechanize"
 require "time"
-require "cgi"
+require "uri"
+
+$LOAD_PATH.unshift(File.expand_path("../", __FILE__))
 require "sosowa/version"
 require "sosowa/scheme"
 require "sosowa/parser"
 
 module Sosowa
-  BASE_URL = "http://coolier.sytes.net:8080/sosowa/ssw_l"
+  BASE_URL = "http://coolier.sytes.net:8080/sosowa/ssw_l/"
+  
+  # @param [Hash] parameter
+  # @return [String] URL Serialized parameters
+  def self.serialize_parameter parameter
+    return "" unless parameter.class == Hash
+    ant = Hash.new
+    parameter.each do |key, value|
+      ant[key.to_sym] = value.to_s
+    end
+    param = ant.inject(""){|k,v|k+"&#{v[0]}=#{URI.escape(v[1])}"}.sub!(/^&/,"?")
+    return param ? param : ""
+  end
   
   def self.get(args={})
     args[:log] ||= 0
@@ -19,4 +32,9 @@ module Sosowa
       parser.fetch_index(args[:log])
     end
   end
+  
+  def self.search(query, args={})
+    parser = Parser.new
+    parser.search(query, args)
+  end
 end

data/samples/token_segment.rb

@@ -0,0 +1,35 @@
+#!/usr/bin/env ruby
+# coding: utf-8
+# 創想話の最新版から適当なSSを取得してMeCab(+ 東方MeCab辞書)を用いてトークナイズします。
+
+require "MeCab"
+require "sosowa"
+
+module MeCab
+  class Tagger
+    alias_method :parseToNode_org, :parseToNode
+    private :parseToNode_org
+
+    def parseToNode(*args)
+      node = parseToNode_org(*args)
+      nodes = []
+      while node
+        nodes.push(node)
+        node = node.next
+      end
+      return nodes[1, nodes.size - 2]
+    end
+  end
+end
+
+puts "Fetching thdic-mecab..."
+system("curl -L https://github.com/oame/thdic-mecab/raw/master/pkg/thdic-mecab.dic > thdic-mecab.dic")
+
+puts "Done. Initialize MeCab::Tagger"
+mecab = MeCab::Tagger.new("-u thdic-mecab.dic")
+
+text = Sosowa.get.sample.fetch.text.gsub(/(<br>|\r?\n)/, "")
+tokens = mecab.parseToNode(text)
+tokens.each do |token|
+  puts token.feature
+end

data/sosowa.gemspec

@@ -5,7 +5,7 @@ Gem::Specification.new do |gem|
   gem.authors       = ["Oame"]
   gem.email         = ["oame@oameya.com"]
   gem.description   = %q{Sosowa Parser for Ruby}
-  gem.summary       = %q{Sosowa Parser for Ruby.}
+  gem.summary       = %q{Sosowa Parser for Ruby 1.9.x.}
   gem.homepage      = ""
 
   gem.files         = `git ls-files`.split($\)

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: sosowa
 version: !ruby/object:Gem::Version
-  version: 0.0.2
+  version: '0.1'
   prerelease: 
 platform: ruby
 authors:
@@ -9,11 +9,11 @@ authors:
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2012-07-09 00:00:00.000000000Z
+date: 2012-07-10 00:00:00.000000000Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: mechanize
-  requirement: &70129465130960 !ruby/object:Gem::Requirement
+  requirement: &70307433904340 !ruby/object:Gem::Requirement
     none: false
     requirements:
     - - ! '>='
@@ -21,7 +21,7 @@ dependencies:
         version: '0'
   type: :runtime
   prerelease: false
-  version_requirements: *70129465130960
+  version_requirements: *70307433904340
 description: Sosowa Parser for Ruby
 email:
 - oame@oameya.com
@@ -30,6 +30,7 @@ extensions: []
 extra_rdoc_files: []
 files:
 - .gitignore
+- CHANGELOG.md
 - Gemfile
 - LICENSE
 - README.md
@@ -38,6 +39,7 @@ files:
 - lib/sosowa/parser.rb
 - lib/sosowa/scheme.rb
 - lib/sosowa/version.rb
+- samples/token_segment.rb
 - sosowa.gemspec
 homepage: ''
 licenses: []
@@ -62,6 +64,6 @@ rubyforge_project:
 rubygems_version: 1.8.10
 signing_key: 
 specification_version: 3
-summary: Sosowa Parser for Ruby.
+summary: Sosowa Parser for Ruby 1.9.x.
 test_files: []
 has_rdoc: