-require 'uri'
+Url = Struct.new("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
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)
- 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
-
- 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)
+ extra = String.new
+
+ begin
+ debug "+ getting #{url.request_uri}"
+ @bot.httputil.get_response(url) { |resp|
+ case resp
+ when Net::HTTPSuccess
+
+ 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'])
+ title = get_title_from_html(partial)
+ first_par = Utils.ircify_first_html_par(partial, :strip => title)
+ extra << ", #{Bold}text#{Bold}: #{first_par}" unless first_par.empty?
+ 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|
+ title = get_title_from_html(part)
+ return "#{Bold}title#{Bold}: #{title}" if title
+ }
end
- else
- return "[Link Info] Error getting link (#{response.code} - #{response.message})"
- end # end of "case response"
+ # if nothing was found, provide more basic info, as for non-html pages
+ end
+
+ debug resp.to_hash.inspect
- } # end of request block
- } # end of http start block
+ enc = resp['content-encoding']
- return title
+ extra << ", #{Bold}encoding#{Bold}: #{enc}" if enc
- 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
+ size = size ? ", #{Bold}size#{Bold}: #{size} bytes" : ""
+ return "#{Bold}type#{Bold}: #{resp['content-type']}#{size}#{extra}"
+ end
+ else
+ raise UrlLinkError, "getting link (#{resp.code} - #{resp.message})"
+ end
+ }
+ 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)
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
+ 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
end
end
+ def reply_urls(m, list, max)
+ 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) rescue nil
+ if title and not url.info
+ url.info = title
+ 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
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(m, list, max)
end
end
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(m, list, max)
end
end
end
+
plugin = UrlPlugin.new
plugin.map 'urls search :channel :limit :string', :action => 'search',
:defaults => {:limit => 4},
projects / user / henk / code / ruby / rbot.git / blobdiff