+#-- 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 omissis_after(len)
- if self.length > len
- return self[0...len].sub(/\s+\S*$/,"...")
- else
- return self
- end
- end
-
- def ircify_html
- txt = self
- txt.gsub!(/<\/?b\s*>/, "#{Bold}")
- txt.gsub!(/<\/?i\s*>/, "#{Underline}")
- ## This would be a nice addition, but the results are horrible
- ## Maybe make it configurable?
- # txt.gsub!(/<\/?a( [^>]*)?>/, "#{Reverse}")
- txt.gsub!(/<\/?(p|br)>/, ' ')
- txt.gsub!("\n", ' ')
- txt.gsub!(/<[^>]+>/, '')
- txt.gsub!(/\s+/, ' ')
- 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> </td><td nowrap><font size=\+1><b>(.+)</b></td></tr><tr><td>}
class SearchPlugin < Plugin
BotConfig.register BotConfigIntegerValue.new('google.hits',
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
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
"#{n}. #{Bold}#{t}#{Bold}: #{u}"
}.join(" | ")
- m.reply "Results for #{what}: #{results}"
+ m.reply "Results for #{what}: #{results}", :split_at => /\s+\|\s+/
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
- 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( [^>]*)?>.*?<\/h1>/im)
- txt = nil
- if header_found
- txt = header_found.post_match[/<p( [^>]*)?>.*?<\/p>/im]
- end
- # If we haven't found a first par yet, try to get it from the whole
- # document
- unless txt
- txt = xml[/<p( [^>]*)?>.*?<\/p>/im]
- end
- # Nothing yet, give up
- unless txt
- debug "No first par found\n#{xml}"
- next
- end
- m.reply "[#{idx}] #{txt.ircify_html}".omissis_after(400)
- 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)
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'