X-Git-Url: https://git.netwichtig.de/gitweb/?a=blobdiff_plain;f=data%2Frbot%2Fplugins%2Furl.rb;h=269227a3c468e21d643d83e8f66296be5e6128fb;hb=69db4133c5ccfa41c35b43c67fce1d5ff640bfd5;hp=ca1e6ed86225e6423e865e1602108626771cf343;hpb=4ee4a465fd25bd0d7e353f0f9e5b69fe165b3b30;p=user%2Fhenk%2Fcode%2Fruby%2Frbot.git

diff --git a/data/rbot/plugins/url.rb b/data/rbot/plugins/url.rb
index ca1e6ed8..269227a3 100644
--- a/data/rbot/plugins/url.rb
+++ b/data/rbot/plugins/url.rb
@@ -1,15 +1,28 @@
-require 'uri'
+define_structure :Url, :channel, :nick, :time, :url, :info
 
-Url = Struct.new("Url", :channel, :nick, :time, :url)
-TITLE_RE = /<\s*?title\s*?>(.+?)<\s*?\/title\s*?>/im
+class ::UrlLinkError < RuntimeError
+end
 
 class UrlPlugin < Plugin
+  TITLE_RE = /<\s*?title\s*?>(.+?)<\s*?\/title\s*?>/im
+  LINK_INFO = "[Link Info]"
+
   BotConfig.register BotConfigIntegerValue.new('url.max_urls',
     :default => 100, :validate => Proc.new{|v| v > 0},
     :desc => "Maximum number of urls to store. New urls replace oldest ones.")
   BotConfig.register BotConfigBooleanValue.new('url.display_link_info',
     :default => false,
     :desc => "Get the title of any links pasted to the channel and display it (also tells if the link is broken or the site is down)")
+  BotConfig.register BotConfigBooleanValue.new('url.titles_only',
+    :default => false,
+    :desc => "Only show info for links that have <title> tags (in other words, don't display info for jpegs, mpegs, etc.)")
+  BotConfig.register BotConfigBooleanValue.new('url.first_par',
+    :default => false,
+    :desc => "Also try to get the first paragraph of a web page")
+  BotConfig.register BotConfigBooleanValue.new('url.info_on_list',
+    :default => false,
+    :desc => "Show link info when listing/searching for urls")
+
 
   def initialize
     super
@@ -22,89 +35,93 @@ class UrlPlugin < Plugin
 
   def get_title_from_html(pagedata)
     return unless TITLE_RE.match(pagedata)
-    title = $1.strip.gsub(/\s*\n+\s*/, " ")
-    title = Utils.decode_html_entities title
-    title = title[0..255] if title.length > 255
-    "[Link Info] title: #{title}"
-  end
-
-  def read_data_from_response(response, amount)
-
-    amount_read = 0
-    chunks = []
-
-    response.read_body do |chunk|   # read body now
-
-      amount_read += chunk.length
-
-      if amount_read > amount
-        amount_of_overflow = amount_read - amount
-        chunk = chunk[0...-amount_of_overflow]
-      end
-
-      chunks << chunk
-
-      break if amount_read >= amount
-
-    end
-
-    chunks.join('')
-
+    $1.ircify_html
   end
 
-  def get_title_for_url(uri_str, depth=@bot.config['http.max_redir'])
-    # This god-awful mess is what the ruby http library has reduced me to.
-    # Python's HTTP lib is so much nicer. :~(
+  def get_title_for_url(uri_str, nick = nil, channel = nil)
 
-    if depth == 0
-        raise "Error: Maximum redirects hit."
-    end
-
-    debug "+ Getting #{uri_str.to_s}"
     url = uri_str.kind_of?(URI) ? uri_str : URI.parse(uri_str)
     return if url.scheme !~ /https?/
 
-    title = nil
+    logopts = Hash.new
+    logopts[:nick] = nick if nick
+    logopts[:channel] = channel if channel
 
-    debug "+ connecting to #{url.host}:#{url.port}"
-    http = @bot.httputil.get_proxy(url)
-    http.start { |http|
-
-      http.request_get(url.request_uri(), @bot.httputil.headers) { |response|
-
-        case response
-          when Net::HTTPRedirection
-            # call self recursively if this is a redirect
-            redirect_to = response['location']  || '/'
-            debug "+ redirect location: #{redirect_to.inspect}"
-            url = URI.join(url.to_s, redirect_to)
-            debug "+ whee, redirecting to #{url.to_s}!"
-            return get_title_for_url(url, depth-1)
-          when Net::HTTPSuccess
-            if response['content-type'] =~ /^text\//
-              # since the content is 'text/*' and is small enough to
-              # be a webpage, retrieve the title from the page
-              debug "+ getting #{url.request_uri}"
-              # was 5*10^4 ... seems to much to me ... 4k should be enough for everybody ;)
-              data = read_data_from_response(response, 4096)
-              return get_title_from_html(data)
+    title = nil
+    extra = String.new
+
+    begin
+      debug "+ getting #{url.request_uri}"
+      @bot.httputil.get_response(url) { |resp|
+        case resp
+        when Net::HTTPSuccess
+
+          debug resp.to_hash
+
+          if resp['content-type'] =~ /^text\/|(?:x|ht)ml/
+            # The page is text or HTML, so we can try finding a title and, if
+            # requested, the first par.
+            #
+            # We act differently depending on whether we want the first par or
+            # not: in the first case we download the initial part and the parse
+            # it; in the second case we only download as much as we need to find
+            # the title
+            #
+            if @bot.config['url.first_par']
+              partial = resp.partial_body(@bot.config['http.info_bytes'])
+              logopts[:title] = title = get_title_from_html(partial)
+              first_par = Utils.ircify_first_html_par(partial, :strip => title)
+              unless first_par.empty?
+                logopts[:extra] = first_par
+                extra << ", #{Bold}text#{Bold}: #{first_par}"
+              end
+              call_event(:url_added, url.to_s, logopts)
+              return "#{Bold}title#{Bold}: #{title}#{extra}" if title
             else
-              # content doesn't have title, just display info.
-              size = response['content-length'].gsub(/(\d)(?=\d{3}+(?:\.|$))(\d{3}\..*)?/,'\1,\2')
-              size = size ? ", size: #{size} bytes" : ""
-              return "[Link Info] type: #{response['content-type']}#{size}"
+              resp.partial_body(@bot.config['http.info_bytes']) { |part|
+                logopts[:title] = title = get_title_from_html(part)
+                call_event(:url_added, url.to_s, logopts)
+                return "#{Bold}title#{Bold}: #{title}" if title
+              }
             end
+          # if nothing was found, provide more basic info, as for non-html pages
           else
-            return "[Link Info] Error getting link (#{response.code} - #{response.message})"
-          end # end of "case response"
-
-      } # end of request block
-    } # end of http start block
+            resp.no_cache = true
+          end
 
-    return title
+          enc = resp['content-encoding']
+          logopts[:extra] = String.new
+          logopts[:extra] << "Content Type: #{resp['content-type']}"
+          if enc
+            logopts[:extra] << ", encoding: #{enc}"
+            extra << ", #{Bold}encoding#{Bold}: #{enc}"
+          end
 
-  rescue SocketError => e
-    return "[Link Info] Error connecting to site (#{e.message})"
+          unless @bot.config['url.titles_only']
+            # content doesn't have title, just display info.
+            size = resp['content-length'].gsub(/(\d)(?=\d{3}+(?:\.|$))(\d{3}\..*)?/,'\1,\2') rescue nil
+            if size
+              logopts[:extra] << ", size: #{size} bytes"
+              size = ", #{Bold}size#{Bold}: #{size} bytes"
+            end
+            call_event(:url_added, url.to_s, logopts)
+            return "#{Bold}type#{Bold}: #{resp['content-type']}#{size}#{extra}"
+          end
+          call_event(:url_added, url.to_s, logopts)
+        else
+          raise UrlLinkError, "getting link (#{resp.code} - #{resp.message})"
+        end
+      }
+      return nil
+    rescue Exception => e
+      case e
+      when UrlLinkError
+        raise e
+      else
+        error e
+        raise "connecting to site/processing information (#{e.message})"
+      end
+    end
   end
 
   def listen(m)
@@ -116,25 +133,28 @@ class UrlPlugin < Plugin
         urlstr = $1
         list = @registry[m.target]
 
+        title = nil
         if @bot.config['url.display_link_info']
-          debug "Getting title for #{urlstr}..."
-          begin
-          title = get_title_for_url urlstr
-          if title
-            m.reply title
-            debug "Title found!"
-          else
-            debug "Title not found!"
-          end
-          rescue => e
-            debug "Failed: #{e}"
+          Thread.start do
+            debug "Getting title for #{urlstr}..."
+            begin
+              title = get_title_for_url urlstr, m.source.nick, m.channel
+              if title
+                m.reply "#{LINK_INFO} #{title}", :overlong => :truncate
+                debug "Title found!"
+              else
+                debug "Title not found!"
+              end
+            rescue => e
+              m.reply "Error #{e.message}"
+            end
           end
         end
 
         # check to see if this url is already listed
         return if list.find {|u| u.url == urlstr }
 
-        url = Url.new(m.target, m.sourcenick, Time.new, urlstr)
+        url = Url.new(m.target, m.sourcenick, Time.new, urlstr, title)
         debug "#{list.length} urls so far"
         if list.length > @bot.config['url.max_urls']
           list.pop
@@ -147,6 +167,31 @@ class UrlPlugin < Plugin
     end
   end
 
+  def reply_urls(opts={})
+    list = opts[:list]
+    max = opts[:max]
+    channel = opts[:channel]
+    m = opts[:msg]
+    return unless list and max and m
+    list[0..(max-1)].each do |url|
+      disp = "[#{url.time.strftime('%Y/%m/%d %H:%M:%S')}] <#{url.nick}> #{url.url}"
+      if @bot.config['url.info_on_list']
+        title = url.info || get_title_for_url(url.url, url.nick, channel) rescue nil
+        # If the url info was missing and we now have some, try to upgrade it
+        if channel and title and not url.info
+          ll = @registry[channel]
+          debug ll
+          if el = ll.find { |u| u.url == url.url }
+            el.info = title
+            @registry[channel] = ll
+          end
+        end
+        disp << " --> #{title}" if title
+      end
+      m.reply disp, :overlong => :truncate
+    end
+  end
+
   def urls(m, params)
     channel = params[:channel] ? params[:channel] : m.target
     max = params[:limit].to_i
@@ -156,9 +201,7 @@ class UrlPlugin < Plugin
     if list.empty?
       m.reply "no urls seen yet for channel #{channel}"
     else
-      list[0..(max-1)].each do |url|
-        m.reply "[#{url.time.strftime('%Y/%m/%d %H:%M:%S')}] <#{url.nick}> #{url.url}"
-      end
+      reply_urls :msg => m, :channel => channel, :list => list, :max => max
     end
   end
 
@@ -170,17 +213,17 @@ class UrlPlugin < Plugin
     max = 1 if max < 1
     regex = Regexp.new(string, Regexp::IGNORECASE)
     list = @registry[channel].find_all {|url|
-      regex.match(url.url) || regex.match(url.nick)
+      regex.match(url.url) || regex.match(url.nick) ||
+        (@bot.config['url.info_on_list'] && regex.match(url.info))
     }
     if list.empty?
       m.reply "no matches for channel #{channel}"
     else
-      list[0..(max-1)].each do |url|
-        m.reply "[#{url.time.strftime('%Y/%m/%d %H:%M:%S')}] <#{url.nick}> #{url.url}"
-      end
+      reply_urls :msg => m, :channel => channel, :list => list, :max => max
     end
   end
 end
+
 plugin = UrlPlugin.new
 plugin.map 'urls search :channel :limit :string', :action => 'search',
                           :defaults => {:limit => 4},