From 6cbe66cdd40ef0bc0d25ba37c22bb7e08404a09f Mon Sep 17 00:00:00 2001
From: Giuseppe Bilotta <giuseppe.bilotta@gmail.com>
Date: Sat, 14 Oct 2006 10:02:43 +0000
Subject: Update demauro plugin to use the WAP interface: much better, much
 faster

---
 data/rbot/plugins/demauro.rb | 96 +++++++++++---------------------------------
 1 file changed, 23 insertions(+), 73 deletions(-)

(limited to 'data')

diff --git a/data/rbot/plugins/demauro.rb b/data/rbot/plugins/demauro.rb
index 12a784ff..b502238e 100644
--- a/data/rbot/plugins/demauro.rb
+++ b/data/rbot/plugins/demauro.rb
@@ -1,92 +1,42 @@
 require 'erb'
 
+DEMAURO_LEMMA = /<anchor>(.*?)(?: - (.*?))<go href="lemma.php\?ID=(\d+)"\/><\/anchor>/
 class DeMauroPlugin < Plugin
-  include ERB::Util
-
   def initialize
     super
     @dmurl = "http://www.demauroparavia.it/"
+    @wapurl = "http://wap.demauroparavia.it/"
   end
 
 
   def help(plugin, topic="")
-    return "demauro <parola> => fornisce il link della definizione della parola dal dizionario De Mauro/Paravia"
+    return "demauro <word> => provides a link to the definition of the word from the Italian dictionary De Mauro/Paravia"
   end
 
   def demauro(m, params)
-    parola = params[:parola]
-    url = @dmurl + "cerca?stringa=#{url_encode(parola)}"
-    uri = URI.parse(url)
-    http = @bot.httputil.get_proxy(uri)
-    xml = nil
-    defurls = Array.new
-    begin
-      http.start() { |http|
-	resp = http.get(uri.request_uri())
-	case resp.code
-	when "200"
-	  xml = resp.body
-	when "302"
-	  loc = resp['location']
-	  if loc =~ /#{@dmurl}\d+/
-	    defurls << loc
-	  end
-	else
-	  debug resp.to_a
-	end
-      }
-    rescue => e
-      debug "HttpUtil.get exception: #{e.inspect}, while trying to get #{uri}"
-      debug e.backtrace.join("\n")
-      m.reply "C'è stato un errore nella ricerca"
+    parola = params[:parola].downcase
+    url = @wapurl + "index.php?lemma=#{ERB::Util.url_encode(parola)}"
+    xml = @bot.httputil.get(url)
+    if xml.nil?
+      info = @bot.httputil.last_response
+      info = info ? "(#{info.code} - #{info.message})" : ""
+      m.reply "An error occurred while looking for #{parola}#{info}"
+      return
+    end
+    if xml=~ /Non ho trovato occorrenze per/
+      m.reply "Nothing found for #{parola}"
       return
     end
-    if xml
-      if xml=~ /Non ho trovato occorrenze per/
-	m.reply "Parola non trovata"
-	return
-      else
-	xml.gsub(/href="(\d+)"/) { |match|
-	  debug match.to_a.join(" || ")
-	  defurls << "#{@dmurl}#{$1}"
-	}
-      end
+    entries = xml.scan(DEMAURO_LEMMA)
+    text = parola
+    if !entries.assoc(parola) and !entries.assoc(parola.upcase)
+      text += " not found. Similar words"
     end
-    lemmas = Array.new
-    defurls.each { |url|
-      uri = URI.parse(url)
-      http = @bot.httputil.get_proxy(uri)
-      begin
-	debug "Scanning #{url}"
-	http.start() { |http|
-	  resp = http.get(uri.request_uri())
-	  case resp.code
-	  when "200"
-	    debug "Got data"
-	    matched = /<span class="lemma">(.*)<\/span><br\/><span class="qualifica".*?>(.*?)<\/span><br\/>/.match(resp.body)
-	    dirtylemma = matched[1]
-	    qual = matched[2]
-	    lemma = dirtylemma.gsub(/<\/?span(?: class="pipelemma")?>/,"")
-	    debug lemma
-	    lemma = lemma.gsub(/<sup>1<\/sup>/,'¹').gsub(/<sup>2<\/sup>/,'²').gsub(/<sup>3<\/sup>/,'³')
-	    lemma = lemma.gsub(/<sup>4<\/sup>/,'⁴').gsub(/<sup>5<\/sup>/,'⁵').gsub(/<sup>6<\/sup>/,'⁶')
-	    lemma = lemma.gsub(/<sup>7<\/sup>/,'⁷').gsub(/<sup>8<\/sup>/,'⁸').gsub(/<sup>9<\/sup>/,'⁹')
-	    debug lemma
-	    lemma += " #{qual} (#{uri})"
-	    lemmas << lemma
-	  else
-	    debug resp.to_a.join("\r")
-	  end
-	}
-      rescue => e
-	debug "Exception '#{e.inspect}' while trying to get and parse #{uri}"
-	debug e.backtrace.join("\n")
-	m.reply "C'è stato un errore nell'elaborazione del risultato"
-	return
-      end
-    }
-    pre = lemmas.length > 1 ? "Lemmi trovati" : "Lemma trovato"
-    m.reply "#{pre}: #{lemmas.join(' ; ')}"
+    text += ": "
+    text += entries[0..5].map { |ar|
+      "#{ar[0]} - #{ar[1].gsub(/<\/?em>/,'')}: #{@dmurl}#{ar[2]}"
+    }.join(" | ")
+    m.reply text
   end
 end
 
-- 
cgit v1.2.3