From ec67cbca6543d4eb7a95e83fb10f340383bb47e7 Mon Sep 17 00:00:00 2001
From: Giuseppe Bilotta <giuseppe.bilotta@gmail.com>
Date: Wed, 5 Jul 2006 16:19:40 +0000
Subject: Add plugin for dictionary lookup using the (Italian) De Mauro/Paravia
 online dictionary

---
 data/rbot/plugins/demauro.rb | 95 ++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 95 insertions(+)
 create mode 100644 data/rbot/plugins/demauro.rb

diff --git a/data/rbot/plugins/demauro.rb b/data/rbot/plugins/demauro.rb
new file mode 100644
index 00000000..7e5705a1
--- /dev/null
+++ b/data/rbot/plugins/demauro.rb
@@ -0,0 +1,95 @@
+# TODO: cache definitions
+
+require 'erb'
+
+class DeMauroPlugin < Plugin
+  include ERB::Util
+
+  def initialize
+    super
+    @dmurl = "http://www.demauroparavia.it/"
+  end
+
+
+  def help(plugin, topic="")
+    return "demauro <parola> => fornisce il link della definizione della parola dal dizionario 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)
+	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}, while trying to get #{uri}"
+      m.reply "Errore"
+      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
+    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)
+	  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}' while trying to get and parse #{uri}"
+	m.reply "Errore"
+	return
+      end
+    }
+    pre = lemmas.length > 1 ? "Lemmi trovati" : "Lemma trovato"
+    m.reply "#{pre}: #{lemmas.join(' ; ')}"
+  end
+end
+
+plugin = DeMauroPlugin.new
+plugin.map 'demauro :parola', :action => 'demauro'
+
-- 
cgit v1.2.3