-require 'net/http'
require 'uri'
-require 'cgi'
Url = Struct.new("Url", :channel, :nick, :time, :url)
-TITLE_RE = /<\s*title\s*>(.+)<\s*\/title\s*>/im
+TITLE_RE = /<\s*?title\s*?>(.+?)<\s*?\/title\s*?>/im
+LINK_INFO = "[Link Info]"
class UrlPlugin < Plugin
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 => true,
+ :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")
+
def initialize
super
@registry.set_default(Array.new)
def get_title_from_html(pagedata)
return unless TITLE_RE.match(pagedata)
- title = $1.strip.gsub(/\s*\n+\s*/, " ")
- title = CGI::unescapeHTML title
- title = title[0..255] if title.length > 255
- "[Link Info] title: #{title}"
+ $1.ircify_html
end
def get_title_for_url(uri_str)
- # This god-awful mess is what the ruby http library has reduced me to.
- # Python's HTTP lib is so much nicer. :~(
-
- puts "+ Getting #{uri_str}"
- url = URI.parse(uri_str)
+
+ url = uri_str.kind_of?(URI) ? uri_str : URI.parse(uri_str)
return if url.scheme !~ /https?/
-
- puts "+ connecting to #{url.host}:#{url.port}"
- title = Net::HTTP.start(url.host, url.port) do |http|
- url.path = '/' if url.path == ''
- head = http.request_head(url.path)
- case head
- when Net::HTTPRedirection then
- # call self recursively if this is a redirect
- redirect_to = head['location']
- puts "+ redirect location: #{redirect_to}"
- absolute_uris = URI.extract redirect_to
- raise "wtf! redirect = #{redirect_to}" if absolute_uris.size > 1
- if absolute_uris.size == 1
- url = URI.parse absolute_uris[0]
- else
- url.path = redirect_to
- end
- puts "+ whee, redirect to #{url.to_s}!"
- title = get_title_for_url(url.to_s)
- when Net::HTTPSuccess then
- if head['content-type'] =~ /^text\//
- # content is 'text/*'
- # retrieve the title from the page
- puts "+ getting #{url.path}"
- response = http.request_get(url.path)
- return get_title_from_html(response.body)
- else
- # content isn't 'text/*'... display info about the file.
- size = head['content-length'].gsub(/(\d)(?=\d{3}+(?:\.|$))(\d{3}\..*)?/,'\1,\2')
- #lastmod = head['last-modified']
- return "[Link Info] type: #{head['content-type']}#{size ? ", size: #{size} bytes" : ""}"
- end
- when Net::HTTPClientError then
- return "Error getting link (#{response.code} - #{response.message})"
- when Net::HTTPServerError then
- return "Error getting link (#{response.code} - #{response.message})"
+
+ title = nil
+
+ begin
+ range = @bot.config['http.info_bytes']
+ response = @bot.httputil.get_response(url, :range => "bytes=0-#{range}")
+ if response.code != "206" && response.code != "200"
+ return "Error getting link (#{response.code} - #{response.message})"
+ end
+ extra = String.new
+
+ if response['content-type'] =~ /^text\//
+
+ body = response.body.slice(0, range)
+ title = String.new
+
+ # since the content is 'text/*' and is small enough to
+ # be a webpage, retrieve the title from the page
+ debug "+ getting #{url.request_uri}"
+
+ # 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']
+ title = get_title_from_html(body)
+ first_par = Utils.ircify_first_html_par(body, :strip => title)
+ extra << "\n#{LINK_INFO} text: #{first_par}" unless first_par.empty?
+ return "title: #{title}#{extra}" if title
+ else
+ title = get_title_from_html(body)
+ return "title: #{title}" if title
+ end
+
+ # if nothing was found, provide more basic info
end
+
+ debug response.to_hash.inspect
+ unless @bot.config['url.titles_only']
+ # content doesn't have title, just display info.
+ size = response['content-length'].gsub(/(\d)(?=\d{3}+(?:\.|$))(\d{3}\..*)?/,'\1,\2') rescue nil
+ size = size ? ", size: #{size} bytes" : ""
+ return "type: #{response['content-type']}#{size}#{extra}"
+ end
+ rescue Exception => e
+ error e.inspect
+ debug e.backtrace.join("\n")
+ return "Error connecting to site (#{e.message})"
end
- rescue SocketError => e
- return "[Link Info] Error connecting to site (#{e.message})"
end
def listen(m)
urlstr = $1
list = @registry[m.target]
- if @bot.config['url.say_titles']
- debug "Getting title for #{urlstr}..."
- title = get_title_for_url urlstr
- if title
- m.reply title
- debug "Title found!"
- else
- debug "Title not found!"
- end
+ if @bot.config['url.display_link_info']
+ 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
+ debug "Failed: #{e}"
+ 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)
debug "#{list.length} urls so far"
if list.length > @bot.config['url.max_urls']
string = params[:string]
max = 10 if max > 10
max = 1 if max < 1
- regex = Regexp.new(string)
+ regex = Regexp.new(string, Regexp::IGNORECASE)
list = @registry[channel].find_all {|url|
regex.match(url.url) || regex.match(url.nick)
}
projects / user / henk / code / ruby / rbot.git / blobdiff