lastfm plugin: perfectionism run

[user/henk/code/ruby/rbot.git] / data / rbot / plugins / search.rb

diff --git a/data/rbot/plugins/search.rb b/data/rbot/plugins/search.rb
index 27f9519c37b4d5e3bec5bbda080380526371bed9..aea6dc4cc4a5f9061261eeb2b88471118b1dc22f 100644 (file)
--- a/data/rbot/plugins/search.rb
+++ b/data/rbot/plugins/search.rb
@@ -1,41 +1,24 @@
+#-- 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
+
 require 'uri'
 
 Net::HTTP.version_1_2
 
 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{<p><table><tr><td><img src=/images/calc_img\.gif></td><td>&nbsp;</td><td nowrap><font size=\+1><b>(.+)</b></td></tr><tr><td>}
 
 class SearchPlugin < Plugin
   BotConfig.register BotConfigIntegerValue.new('google.hits',
@@ -55,6 +38,8 @@ 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 "wp"
       "wp [<code>] <string> => search for <string> on Wikipedia. You can select a national <code> to only search the national Wikipedia"
     else
@@ -80,7 +65,8 @@ class SearchPlugin < Plugin
     hits = params[:hits] || @bot.config['google.hits']
 
     begin
-      wml = @bot.httputil.get_cached(url)
+      wml = @bot.httputil.get(url)
+      raise unless wml
     rescue => e
       m.reply "error googling for #{what}"
       return
@@ -103,51 +89,39 @@ class SearchPlugin < Plugin
 
     first_pars = params[:firstpar] || @bot.config['google.first_par']
 
-    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
+    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 = URI.escape(what).sub('+','%2B')
+    
+    debug "Getting gcalc thing: #{searchfor.inspect}"
+    url = "http://www.google.com/search?q=#{searchfor}"
+
+    begin
+      html = @bot.httputil.get(url)
+    rescue => e
+      m.reply "error googlecalcing #{what}"
+      return
+    end
+
+    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 wikipedia(m, params)
@@ -166,6 +140,7 @@ plugin = SearchPlugin.new
 
 plugin.map "search *words", :action => 'google'
 plugin.map "google *words", :action => 'google'
+plugin.map "gcalc *words", :action => 'gcalc'
 plugin.map "wp :lang *words", :action => 'wikipedia', :requirements => { :lang => /^\w\w\w?$/ }
 plugin.map "wp *words", :action => 'wikipedia'