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 fd1aefdc8a70d512c9cba29d26bcf34a1e2156b3..6fb1959aeec31b76f87083c78260cbf02269519b 100644 (file)
--- a/data/rbot/plugins/search.rb
+++ b/data/rbot/plugins/search.rb
@@ -5,16 +5,27 @@ Net::HTTP.version_1_2
 GOOGLE_WAP_LINK = /<a accesskey="(\d)" href=".*?u=(.*?)">(.*?)<\/a>/im
 
 class SearchPlugin < Plugin
+  BotConfig.register BotConfigIntegerValue.new('google.hits',
+    :default => 3,
+    :desc => "Number of hits to return from Google searches")
+  BotConfig.register BotConfigIntegerValue.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',
+    :default => 3,
+    :desc => "Number of hits to return from Wikipedia searches")
+  BotConfig.register BotConfigIntegerValue.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")
+
   def help(plugin, topic="")
     case topic
-    when "search"
-    "search <string> => search google for <string>"
-    when "google"
-    "google <string> => search google for <string>"
+    when "search", "google"
+      "#{topic} <string> => search google for <string>"
     when "wp"
       "wp [<code>] <string> => search for <string> on Wikipedia. You can select a national <code> to only search the national Wikipedia"
     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"
     end
   end
 
@@ -33,9 +44,10 @@ class SearchPlugin < Plugin
 
     url = "http://www.google.com/wml/search?q=#{site}#{searchfor}"
 
+    hits = params[:hits] || @bot.config['google.hits']
 
     begin
-      wml = @bot.httputil.get(url)
+      wml = @bot.httputil.get_cached(url)
     rescue => e
       m.reply "error googling for #{what}"
       return
@@ -45,21 +57,74 @@ class SearchPlugin < Plugin
       m.reply "no results found for #{what}"
       return
     end
-    results = results[0...3].map { |res|
+    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}"
     }.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
+
+      # 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
+    end
   end
 
   def wikipedia(m, params)
     lang = params[:lang]
     site = "#{lang.nil? ? '' : lang + '.'}wikipedia.org"
+    debug "Looking up things on #{site}"
     params[:site] = site
     params[:filter] = / - Wikipedia.*$/
+    params[:hits] = @bot.config['wikipedia.hits']
+    params[:firstpar] = @bot.config['wikipedia.first_par']
     return google(m, params)
   end
 end