+# vim: set sw=2 et:
+#
+# 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
-
class SearchPlugin < Plugin
BotConfig.register BotConfigIntegerValue.new('google.hits',
:default => 3,
"#{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
- end
+ return unless first_pars > 0
+
+ Utils.get_first_pars urls, first_pars, :http_util => @bot.httputil, :message => m
+
end
def wikipedia(m, params)