X-Git-Url: https://git.netwichtig.de/gitweb/?a=blobdiff_plain;f=data%2Frbot%2Fplugins%2Fsearch.rb;h=dc0c2ca9bb58419912ef91ac8bf6d327a232df7c;hb=297c80c7632e76e5c5a55cabad57154706911b57;hp=27f9519c37b4d5e3bec5bbda080380526371bed9;hpb=6fe052181de09dd4a618241e600e2473f6706343;p=user%2Fhenk%2Fcode%2Fruby%2Frbot.git

diff --git a/data/rbot/plugins/search.rb b/data/rbot/plugins/search.rb
index 27f9519c..dc0c2ca9 100644
--- a/data/rbot/plugins/search.rb
+++ b/data/rbot/plugins/search.rb
@@ -1,53 +1,38 @@
-require 'uri'
-
-Net::HTTP.version_1_2
-
+#-- vim:sw=2:et
+#++
+#
+# :title: Google and Wikipedia search plugin for rbot
+#
+# Author:: Tom Gilbert (giblet) <tom@linuxbrit.co.uk>
+# Author:: Giuseppe "Oblomov" Bilotta <giuseppe.bilotta@gmail.com>
+#
+# Copyright:: (C) 2002-2005 Tom Gilbert
+# Copyright:: (C) 2006 Tom Gilbert, Giuseppe Bilotta
+# Copyright:: (C) 2006-2007 Giuseppe Bilotta
+
+# TODO:: use lr=lang_<code> or whatever is most appropriate to let google know
+#        it shouldn't use the bot's location to find the preferred language
+# TODO:: support localized uncyclopedias -- not easy because they have different names
+#        for most languages
+
+GOOGLE_SEARCH = "http://www.google.com/search?oe=UTF-8&q="
+GOOGLE_WAP_SEARCH = "http://www.google.com/wml/search?hl=en&q="
 GOOGLE_WAP_LINK = /<a accesskey="(\d)" href=".*?u=(.*?)">(.*?)<\/a>/im
-
-class ::String
-  def ircify_html
-    txt = self
-
-    # bold and strong -> bold
-    txt.gsub!(/<\/?(?:b|strong)\s*>/, "#{Bold}")
-
-    # italic, emphasis and underline -> underline
-    txt.gsub!(/<\/?(?:i|em|u)\s*>/, "#{Underline}")
-
-    ## This would be a nice addition, but the results are horrible
-    ## Maybe make it configurable?
-    # txt.gsub!(/<\/?a( [^>]*)?>/, "#{Reverse}")
-
-    # Paragraph and br tags are converted to whitespace.
-    txt.gsub!(/<\/?(p|br)\s*\/?\s*>/, ' ')
-    txt.gsub!("\n", ' ')
-
-    # All other tags are just removed
-    txt.gsub!(/<[^>]+>/, '')
-
-    # Remove double formatting options, since they only waste bytes
-    txt.gsub!(/#{Bold}\s*#{Bold}/,"")
-    txt.gsub!(/#{Underline}\s*#{Underline}/,"")
-
-    # And finally whitespace is squeezed
-    txt.gsub!(/\s+/, ' ')
-
-    # Decode entities and strip whitespace
-    return Utils.decode_html_entities(txt).strip!
-  end
-end
+GOOGLE_CALC_RESULT = %r{<img src=/images/calc_img\.gif(?: width=40 height=30 alt="")?></td><td>&nbsp;</td><td nowrap(?: dir=ltr)?>(?:<h2 class=r>)?<font size=\+1><b>(.+)</b>(?:</h2>)?</td>}
+GOOGLE_COUNT_RESULT = %r{<font size=-1>Results <b>1<\/b> - <b>10<\/b> of about <b>(.*)<\/b> for}
+GOOGLE_DEF_RESULT = %r{<p> (Web definitions for .*?)<br/>(.*?)<br/>(.*?)\s-\s+<a href}
 
 class SearchPlugin < Plugin
-  BotConfig.register BotConfigIntegerValue.new('google.hits',
+  Config.register Config::IntegerValue.new('google.hits',
     :default => 3,
     :desc => "Number of hits to return from Google searches")
-  BotConfig.register BotConfigIntegerValue.new('google.first_par',
+  Config.register Config::IntegerValue.new('google.first_par',
     :default => 0,
     :desc => "When set to n > 0, the bot will return the first paragraph from the first n search hits")
-  BotConfig.register BotConfigIntegerValue.new('wikipedia.hits',
+  Config.register Config::IntegerValue.new('wikipedia.hits',
     :default => 3,
     :desc => "Number of hits to return from Wikipedia searches")
-  BotConfig.register BotConfigIntegerValue.new('wikipedia.first_par',
+  Config.register Config::IntegerValue.new('wikipedia.first_par',
     :default => 1,
     :desc => "When set to n > 0, the bot will return the first paragraph from the first n wikipedia search hits")
 
@@ -55,16 +40,22 @@ class SearchPlugin < Plugin
     case topic
     when "search", "google"
       "#{topic} <string> => search google for <string>"
+    when "gcalc"
+      "gcalc <equation> => use the google calculator to find the answer to <equation>"
+    when "gdef"
+      "gdef <term(s)> => use the google define mechanism to find a definition of <term(s)>"
     when "wp"
       "wp [<code>] <string> => search for <string> on Wikipedia. You can select a national <code> to only search the national Wikipedia"
+    when "unpedia"
+      "unpedia <string> => search for <string> on Uncyclopedia"
     else
-      "search <string> (or: google <string>) => search google for <string> | wp <string> => search for <string> on Wikipedia"
+      "search <string> (or: google <string>) => search google for <string> | wp <string> => search for <string> on Wikipedia | unpedia <string> => search for <string> on Uncyclopedia"
     end
   end
 
   def google(m, params)
     what = params[:words].to_s
-    searchfor = URI.escape what
+    searchfor = CGI.escape what
     # This method is also called by other methods to restrict searching to some sites
     if params[:site]
       site = "site:#{params[:site]}+"
@@ -75,12 +66,17 @@ class SearchPlugin < Plugin
     # e.g.: "Wikipedia, the free encyclopedia" when doing Wikipedia searches
     filter = params[:filter] || ""
 
-    url = "http://www.google.com/wml/search?q=#{site}#{searchfor}"
+    url = GOOGLE_WAP_SEARCH + site + searchfor
 
     hits = params[:hits] || @bot.config['google.hits']
 
+    first_pars = params[:firstpar] || @bot.config['google.first_par']
+
+    single = (hits == 1 and first_pars == 1)
+
     begin
-      wml = @bot.httputil.get_cached(url)
+      wml = @bot.httputil.get(url)
+      raise unless wml
     rescue => e
       m.reply "error googling for #{what}"
       return
@@ -90,64 +86,116 @@ class SearchPlugin < Plugin
       m.reply "no results found for #{what}"
       return
     end
+    single ||= (results.length==1)
     urls = Array.new
     results = results[0...hits].map { |res|
       n = res[0]
       t = Utils.decode_html_entities res[2].gsub(filter, '').strip
       u = URI.unescape res[1]
       urls.push(u)
-      "#{n}. #{Bold}#{t}#{Bold}: #{u}"
+      single ? u : "#{n}. #{Bold}#{t}#{Bold}: #{u}"
     }.join(" | ")
 
+    # If we return a single, full result, change the output to a more compact representation
+    if single
+      m.reply "Result for %s: %s -- %s" % [what, results, Utils.get_first_pars(urls, first_pars)], :overlong => :truncate
+      return
+    end
+
     m.reply "Results for #{what}: #{results}", :split_at => /\s+\|\s+/
 
-    first_pars = params[:firstpar] || @bot.config['google.first_par']
+    return unless first_pars > 0
+
+    Utils.get_first_pars urls, first_pars, :message => m
+
+  end
+
+  def gcalc(m, params)
+    what = params[:words].to_s
+    searchfor = CGI.escape(what)
+    
+    debug "Getting gcalc thing: #{searchfor.inspect}"
+    url = GOOGLE_SEARCH + searchfor
+
+    begin
+      html = @bot.httputil.get(url)
+    rescue => e
+      m.reply "error googlecalcing #{what}"
+      return
+    end
 
-    idx = 0
-    while first_pars > 0 and urls.length > 0
-      url.replace(urls.shift)
-      idx += 1
-
-      # FIXME what happens if some big file is returned? We should share
-      # code with the url plugin to only retrieve partial file content!
-      xml = @bot.httputil.get_cached(url)
-      if xml.nil?
-        debug "Unable to retrieve #{url}"
-        next
-      end
-      # We get the first par after the first main heading, if possible
-      header_found = xml.match(/<h1(?:\s+[^>]*)?>(.*?)<\/h1>/im)
-      txt = String.new
-      if header_found
-        debug "Found header: #{header_found[1].inspect}"
-        while txt.empty? 
-          header_found = $'
-          candidate = header_found[/<p(?:\s+[^>]*)?>.*?<\/p>/im]
-          break unless candidate
-          txt.replace candidate.ircify_html
-        end
-      end
-      # If we haven't found a first par yet, try to get it from the whole
-      # document
-      if txt.empty?
-	header_found = xml
-        while txt.empty? 
-          candidate = header_found[/<p(?:\s+[^>]*)?>.*?<\/p>/im]
-          break unless candidate
-          txt.replace candidate.ircify_html
-          header_found = $'
-        end
-      end
-      # Nothing yet, try title
-      if txt.empty?
-        debug "No first par found\n#{xml}"
-	# FIXME only do this if the 'url' plugin is loaded
-	txt.replace @bot.plugins['url'].get_title_from_html(xml)
-        next if txt.empty?
-      end
-      m.reply "[#{idx}] #{txt}", :overlong => :truncate
-      first_pars -=1
+    debug "#{html.size} bytes of html recieved"
+    
+    results = html.scan(GOOGLE_CALC_RESULT)
+    debug "results: #{results.inspect}"
+    
+    if results.length != 1
+      m.reply "couldn't calculate #{what}"
+      return
+    end
+    
+    result = results[0][0].ircify_html
+    debug "replying with: #{result.inspect}"
+    m.reply "#{result}"
+  end
+  
+  def gcount(m, params)
+    what = params[:words].to_s
+    searchfor = CGI.escape(what)
+    
+    debug "Getting gcount thing: #{searchfor.inspect}"
+    url = GOOGLE_SEARCH + searchfor
+
+    begin
+      html = @bot.httputil.get(url)
+    rescue => e
+      m.reply "error googlecounting #{what}"
+      return
+    end
+
+    debug "#{html.size} bytes of html recieved"
+    
+    results = html.scan(GOOGLE_COUNT_RESULT)
+    debug "results: #{results.inspect}"
+    
+    if results.length != 1
+      m.reply "couldn't count #{what}"
+      return
+    end
+    
+    result = results[0][0].ircify_html
+    debug "replying with: #{result.inspect}"
+    m.reply "total results: #{result}"
+
+  end
+
+  def gdef(m, params)
+    what = params[:words].to_s
+    searchfor = CGI.escape("define " + what)
+    
+    debug "Getting gdef thing: #{searchfor.inspect}"
+    url = GOOGLE_WAP_SEARCH + searchfor
+
+    begin
+      html = @bot.httputil.get(url)
+    rescue => e
+      m.reply "error googledefining #{what}"
+      return
+    end
+
+    debug html
+    results = html.scan(GOOGLE_DEF_RESULT)
+    debug "results: #{results.inspect}"
+    
+    if results.length != 1
+      m.reply "couldn't find a definition for #{what} on Google"
+      return
     end
+    
+    head = results[0][0].ircify_html
+    text = results[0][1].ircify_html
+    link = results[0][2]
+    m.reply "#{head} -- #{link}\n#{text}"
   end
 
   def wikipedia(m, params)
@@ -160,12 +208,26 @@ class SearchPlugin < Plugin
     params[:firstpar] = @bot.config['wikipedia.first_par']
     return google(m, params)
   end
+
+  def unpedia(m, params)
+    site = "uncyclopedia.org"
+    debug "Looking up things on #{site}"
+    params[:site] = site
+    params[:filter] = / - Uncyclopedia.*$/
+    params[:hits] = @bot.config['wikipedia.hits']
+    params[:firstpar] = @bot.config['wikipedia.first_par']
+    return google(m, params)
+  end
 end
 
 plugin = SearchPlugin.new
 
-plugin.map "search *words", :action => 'google'
-plugin.map "google *words", :action => 'google'
-plugin.map "wp :lang *words", :action => 'wikipedia', :requirements => { :lang => /^\w\w\w?$/ }
-plugin.map "wp *words", :action => 'wikipedia'
+plugin.map "search *words", :action => 'google', :threaded => true
+plugin.map "google *words", :action => 'google', :threaded => true
+plugin.map "gcount *words", :action => 'gcount', :threaded => true 
+plugin.map "gcalc *words", :action => 'gcalc', :threaded => true
+plugin.map "gdef *words", :action => 'gdef', :threaded => true
+plugin.map "wp :lang *words", :action => 'wikipedia', :requirements => { :lang => /^\w\w\w?$/ }, :threaded => true
+plugin.map "wp *words", :action => 'wikipedia', :threaded => true
+plugin.map "unpedia *words", :action => 'unpedia', :threaded => true