searchyj 0.5.2 → 0.6.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: bb887209fc4783aa50c14aa5632f5bc76b15051e
-  data.tar.gz: 20334940910579439b6e68d58c635bce524d95c2
+  metadata.gz: 7dd43ee6cf14a41b52906a501cbcb0035e32cb5c
+  data.tar.gz: 407ba21943c025b50d2882b4f709aaa14fbba083
 SHA512:
-  metadata.gz: 8fe93ca7c69a7e7df5ad02799933043f971b3ed677196d251b06bd6538aacb566898011ece9c7463a3c5a3298c4a16049ce3c0b71b8969e64c2daf12a7fa71af
-  data.tar.gz: 4101ef48646eacd94c740d16bad68d031db19796950a053013b0816d91f142213dc25fb9947673a5c1f0ae7fab69cbde34ad1a20fc3f5c95f2637aba36727162
+  metadata.gz: 5ac674566cb5abbffb71ba270a7cb75c260b14136d881219eaa1a11a6edc0b4b96b12b2e629fec105d517f4158b31817fee8b273782ed59ed8fb4a421011596f
+  data.tar.gz: 0240303f9607f2b3c2697fa988761e8a5d41cec2290fcab0532a3995c441e48a1440e0848fc1f36db971df879bfe660d3b61d836f63d7d1412d98663eb17056f

data/README.ja.md

@@ -78,11 +78,11 @@ gem 'searchyj'
 
 未設定時は 'title' となります。
 
-### at_rank
+### rank
 
 指定順位の検索結果を出力します。
 
-    $ searchyj at_rank [options] <SearchTerm>
+    $ searchyj rank [options] <SearchTerm>
 
 検索にひとつも引っかからない場合は文字列 null が出力されます。
 

data/README.md

@@ -83,11 +83,11 @@ The key name for comparing values. You can pass any of 'title' or 'uri'.
 
 The default value is 'title'.
 
-### at_rank
+### rank
 
 Print a record of the search result at a particular rank order in the search ranking.
 
-    $ searchyj at_rank [options] <SearchTerm>
+    $ searchyj rank [options] <SearchTerm>
 
 If the search result was nothing or did not find, print a string of null.
 

data/TODO.md

@@ -0,0 +1,8 @@
+TODO
+====
+
+## Feature
+
+- Use cookie.
+  - Increase the fetching records at once.
+  

data/lib/searchyj/cli.rb

@@ -4,22 +4,6 @@ require 'json'
 
 module SearchYJ
   class CLI < Thor
-    desc 'at_rank',
-         "Get a record in the search result\n" \
-         'at a particular rank order in the search ranking.'
-    option :rank,
-           type:     :numeric,
-           required: true,
-           aliases:  '-r',
-           desc:     'The rank order in the search ranking'
-    def at_rank(term)
-      rank = options[:rank]
-
-      puts JSON.dump(
-          result = SearchYJ.at_rank(term, rank)
-      )
-    end
-
     desc 'detect',
          "Detect a first record that\n" \
          'meet the conditions of a regexp and a key.'
@@ -65,5 +49,21 @@ module SearchYJ
           SearchYJ.list(term, size, from)
       )
     end
+
+    desc 'rank',
+         "Get a record in the search result\n" \
+         'at a particular rank order in the search ranking.'
+    option :rank,
+           type:     :numeric,
+           required: true,
+           aliases:  '-r',
+           desc:     'The rank order in the search ranking'
+    def rank(term)
+      rank = options[:rank]
+
+      puts JSON.dump(
+          SearchYJ.rank(term, rank)
+      )
+    end
   end
 end

data/lib/searchyj/main.rb

@@ -2,19 +2,6 @@ require 'searchyj/searcher'
 
 module SearchYJ
   class Main
-    # Get a record in the search result
-    # at a particular rank order in the search ranking.
-    # @param term [String]  Search term
-    # @param rank [Integer] The rank order in the search ranking
-    #
-    # @return [Hash]
-    #   A result record if matched the arguments
-    #   Else nil
-    def at_rank(term, rank)
-      result = list(term, 1, rank)
-      (result.size > 0) ? result[0] : nil
-    end
-
     # Detect a first record that
     # meet the conditions of a regexp and a key.
     # @param term   [String] Search term
@@ -29,6 +16,7 @@ module SearchYJ
 
       searcher = Searcher.new
       searcher.uri.search_term = term
+      searcher.pager.size      = 100
 
       searcher.run do |record|
         if regexp.match(record[key])
@@ -51,7 +39,7 @@ module SearchYJ
       searcher = Searcher.new
       searcher.uri.search_term = term
       searcher.uri.index       = start_index
-      searcher.page_size       = 10
+      searcher.pager.size      = size
       list = []
 
       searcher.run do |record|
@@ -61,5 +49,18 @@ module SearchYJ
 
       list
     end
+
+    # Get a record in the search result
+    # at a particular rank order in the search ranking.
+    # @param term [String]  Search term
+    # @param rank [Integer] The rank order in the search ranking
+    #
+    # @return [Hash]
+    #   A result record if matched the arguments
+    #   Else nil
+    def rank(term, rank)
+      result = list(term, 1, rank)
+      (result.size > 0) ? result[0] : nil
+    end
   end
 end

data/lib/searchyj/page_size_adjuster.rb

@@ -0,0 +1,67 @@
+
+module SearchYJ
+  #
+  # Manage cookie for page
+  #
+  # @author [indeep-xyz]
+  #
+  class PageSizeAdjuster
+    attr_reader :size
+
+    SIZE_PATTERN = [10, 15, 20, 30, 40, 100]
+    SIZE_DEFAULT = SIZE_PATTERN[0]
+    COOKIE_BASE  = \
+        'sB="n=<<n>>&nw=-1&fp_ipod=0&fp_pl=0"; ' \
+        'path=/; ' \
+        'expire=<<expire>>'
+    EXPIRE_DELAY = 60 * 60 * 24
+
+    # Initialize myself.
+    def initialize
+      @size   = SIZE_DEFAULT
+      @expire = Time.now.to_i + EXPIRE_DELAY
+    end
+
+    def size=(size)
+      @size = optimize_page_size(size)
+    end
+
+    # Optimize the number of the page size for searching.
+    # @param size [Number] The number of the page size
+    #
+    # @return [Number] The optimized number
+    def optimize_page_size(size)
+      SIZE_PATTERN.reverse_each do |n|
+        return n if size >= n
+      end
+
+      SIZE_DEFAULT
+    end
+
+    # Attach the cookie string to the argument
+    # if @size has differed from the default value.
+    # @param hash [Hash]
+    #
+    # @return [Hash]
+    def attach_cookie(hash)
+      if @size > SIZE_DEFAULT
+        hash['Cookie'] = create_cookie
+      end
+
+      hash
+    end
+
+    private
+
+    # Create the cookie string to adjust the page size.
+    #
+    # @return [String] The cookie string
+    def create_cookie
+      COOKIE_BASE.gsub(
+          /<<[^>]+>>/,
+          '<<n>>' => @size.to_s,
+          '<<expire>>' => @expire.to_s
+      )
+    end
+  end
+end

data/lib/searchyj/searcher.rb

@@ -2,6 +2,7 @@ require 'nokogiri'
 require 'open-uri'
 require 'searchyj/uri_manager'
 require 'searchyj/record_sorter'
+require 'searchyj/page_size_adjuster'
 
 module SearchYJ
   #
@@ -13,7 +14,13 @@ module SearchYJ
   #
   class Searcher
     attr_reader :results
-    attr_accessor :limit_loop, :user_agent, :sleep_time, :page_size, :uri
+    attr_accessor \
+        :pager, :uri, \
+        :limit_loop, :user_agent, :sleep_time
+
+    ENCODING   = 'UTF-8'
+    LIMIT_LOOP = 50
+    SLEEP_TIME = 1
     USER_AGENT = \
         'Mozilla/5.0 (Windows NT 6.1; WOW64; rv:38.0)' \
         'Gecko/20100101 Firefox/38.0'
@@ -22,16 +29,17 @@ module SearchYJ
 
     # Initialize myself.
     def initialize
+      @pager      = PageSizeAdjuster.new
       @uri        = UriManager.new
-      @limit_loop = 10
+      @encoding   = ENCODING
+      @limit_loop = LIMIT_LOOP
+      @sleep_time = SLEEP_TIME
       @user_agent = USER_AGENT
-      @sleep_time = 1
-      @page_size  = 10
     end
 
     def run(&block)
       loop_count = 0
-      sorter = RecordSorter.new(@uri.index, @page_size)
+      sorter = RecordSorter.new(@uri.index, @pager.size)
 
       while loop_count < @limit_loop
         fetch_html
@@ -79,6 +87,8 @@ module SearchYJ
           'User-Agent' => @user_agent
       }
 
+      params = @pager.attach_cookie(params)
+
       open(uri, params) do |f|
         fail OpenUriError unless f.status[0] == '200'
         f.read
@@ -89,7 +99,7 @@ module SearchYJ
     # and set the parsed HTML data to my own instance.
     def fetch_html
       raw_html = download_raw_html
-      @html = Nokogiri::HTML.parse(raw_html, nil, 'UTF-8')
+      @html = Nokogiri::HTML.parse(raw_html, nil, @encoding)
     end
 
     # Check whether or not the next page is exist.

data/lib/searchyj/version.rb

@@ -1,3 +1,3 @@
 module SearchYJ
-  VERSION = "0.5.2"
+  VERSION = "0.6.0"
 end

data/lib/searchyj.rb

@@ -10,9 +10,9 @@ module SearchYJ
     mgr.list(term, size, start_index)
   end
 
-  def at_rank(term, rank)
+  def rank(term, rank)
     mgr = SearchYJ::Main.new
-    mgr.at_rank(term, rank)
+    mgr.rank(term, rank)
   end
 
   def detect(term, regexp, key = :title)

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: searchyj
 version: !ruby/object:Gem::Version
-  version: 0.5.2
+  version: 0.6.0
 platform: ruby
 authors:
 - indeep-xyz
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2015-06-22 00:00:00.000000000 Z
+date: 2015-06-23 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: nokogiri
@@ -96,12 +96,14 @@ files:
 - README.ja.md
 - README.md
 - Rakefile
+- TODO.md
 - bin/console
 - bin/searchyj
 - bin/setup
 - lib/searchyj.rb
 - lib/searchyj/cli.rb
 - lib/searchyj/main.rb
+- lib/searchyj/page_size_adjuster.rb
 - lib/searchyj/record_sorter.rb
 - lib/searchyj/searcher.rb
 - lib/searchyj/unique_logger.rb