Move extensions to standard classes into a specific extends.rb util module

[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 a035b8311acb3efbd6500502ad29fe36952ed890..6fb1959aeec31b76f87083c78260cbf02269519b 100644 (file)
--- a/data/rbot/plugins/search.rb
+++ b/data/rbot/plugins/search.rb
@@ -4,30 +4,6 @@ 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,
@@ -90,7 +66,7 @@ class SearchPlugin < Plugin
       "#{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']
 
@@ -98,28 +74,45 @@ class SearchPlugin < Plugin
     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( [^>]*)?>.*?<\/h1>/im)
-      txt = nil
+      header_found = xml.match(/<h1(?:\s+[^>]*)?>(.*?)<\/h1>/im)
+      txt = String.new
       if header_found
-        txt = header_found.post_match[/<p( [^>]*)?>.*?<\/p>/im]
+        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
-      unless txt
-        txt = xml[/<p( [^>]*)?>.*?<\/p>/im]
+      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, give up
-      unless txt
+      # Nothing yet, try title
+      if txt.empty?
         debug "No first par found\n#{xml}"
-        next
+       # 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.ircify_html}".omissis_after(400)
+      m.reply "[#{idx}] #{txt}", :overlong => :truncate
       first_pars -=1
     end
   end