4 # :title: Standard classes extensions
6 # Author:: Giuseppe "Oblomov" Bilotta <giuseppe.bilotta@gmail.com>
8 # This file collects extensions to standard Ruby classes and to some core rbot
9 # classes to be used by the various plugins
11 # Please note that global symbols have to be prefixed by :: because this plugin
12 # will be read into an anonymous module
14 # Extensions to the Module class
18 # Many plugins define Struct objects to hold their data. On rescans, lots of
19 # warnings are echoed because of the redefinitions. Using this method solves
20 # the problem, by checking if the Struct already exists, and if it has the
23 def define_structure(name, *members)
25 if Struct.const_defined?(sym)
26 kl = Struct.const_get(sym)
27 if kl.new.members.map { |member| member.intern } == members.map
28 debug "Struct #{sym} previously defined, skipping"
33 debug "Defining struct #{sym} with members #{members.inspect}"
34 const_set(sym, Struct.new(name.to_s, *members))
39 # DottedIndex mixin: extend a Hash or Array class with this module
40 # to achieve [] and []= methods that automatically split indices
41 # at dots (indices are automatically converted to symbols, too)
43 # You have to define the single_retrieve(_key_) and
44 # single_assign(_key_,_value_) methods (usually aliased at the
45 # original :[] and :[]= methods)
48 def rbot_index_split(*ar)
49 keys = ([] << ar).flatten
51 k.to_s.split('.').map { |kk| kk.to_sym rescue nil }.compact
56 keys = self.rbot_index_split(ar)
57 return self.single_retrieve(keys.first) if keys.length == 1
61 h[k] ||= self.class.new
70 keys = self.rbot_index_split(ar)
71 return self.single_assign(keys.first, val) if keys.length == 1
75 h[k] ||= self.class.new
83 # Extensions to the Array class
87 # This method returns a random element from the array, or nil if the array is
91 return nil if self.empty?
92 self[rand(self.length)]
95 # This method returns a given element from the array, deleting it from the
96 # array itself. The method returns nil if the element couldn't be found.
98 # If nil is specified, a random element is returned and deleted.
100 def delete_one(val=nil)
101 return nil if self.empty?
103 index = rand(self.length)
105 index = self.index(val)
106 return nil unless index
108 self.delete_at(index)
111 # shuffle and shuffle! are defined in Ruby >= 1.8.7
113 # This method returns a new array with the same items as
114 # the receiver, but shuffled
115 unless method_defined? :shuffle
121 # This method shuffles the items in the array
122 unless method_defined? :shuffle!
128 # This method is an advanced version of #join
129 # allowing fine control of separators:
131 # [1,2,3].conjoin(', ', ' and ')
134 # [1,2,3,4].conjoin{ |i, a, b| i % 2 == 0 ? '.' : '-' }
137 # Code lifted from the ruby facets project:
138 # <http://facets.rubyforge.org>
139 # git-rev: c8b7395255b977d3c7de268ff563e3c5bc7f1441
140 # file: lib/core/facets/array/conjoin.rb
141 def conjoin(*args, &block)
142 return first.to_s if size < 2
147 (size - 1).times do |i|
148 sep << yield(i, *slice(i, 2))
151 options = (Hash === args.last) ? args.pop : {}
152 separator = args.shift || ""
153 options[-1] = args.shift unless args.empty?
155 sep = [separator] * (size - 1)
157 if options.key?(:last)
158 options[-1] = options.delete(:last)
160 options[-1] ||= _(" and ")
162 options.each{ |i, s| sep[i] = s }
169 # Extensions to the Range class
173 # This method returns a random number between the lower and upper bound
176 len = self.last - self.first
177 len += 1 unless self.exclude_end?
178 self.first + Kernel::rand(len)
180 alias :rand :pick_one
183 # Extensions for the Numeric classes
187 # This method forces a real number to be not more than a given positive
188 # number or not less than a given positive number, or between two any given
191 def clip(left,right=0)
192 raise ArgumentError unless left.kind_of?(Numeric) and right.kind_of?(Numeric)
201 # Extensions to the String class
203 # TODO make riphtml() just call ircify_html() with stronger purify options.
207 # This method will return a purified version of the receiver, with all HTML
208 # stripped off and some of it converted to IRC formatting
210 def ircify_html(opts={})
214 txt.gsub!(/<script(?:\s+[^>]*)?>.*?<\/script>/im, "")
217 txt.gsub!(/<style(?:\s+[^>]*)?>.*?<\/style>/im, "")
219 # bold and strong -> bold
220 txt.gsub!(/<\/?(?:b|strong)(?:\s+[^>]*)?>/im, "#{Bold}")
222 # italic, emphasis and underline -> underline
223 txt.gsub!(/<\/?(?:i|em|u)(?:\s+[^>]*)?>/im, "#{Underline}")
225 ## This would be a nice addition, but the results are horrible
226 ## Maybe make it configurable?
227 # txt.gsub!(/<\/?a( [^>]*)?>/, "#{Reverse}")
228 case val = opts[:a_href]
229 when Reverse, Bold, Underline
230 txt.gsub!(/<(?:\/a\s*|a (?:[^>]*\s+)?href\s*=\s*(?:[^>]*\s*)?)>/, val)
232 # Not good for nested links, but the best we can do without something like hpricot
233 txt.gsub!(/<a (?:[^>]*\s+)?href\s*=\s*(?:([^"'>][^\s>]*)\s+|"((?:[^"]|\\")*)"|'((?:[^']|\\')*)')(?:[^>]*\s+)?>(.*?)<\/a>/) { |match|
235 debug [$1, $2, $3, $4].inspect
236 link = $1 || $2 || $3
241 warning "unknown :a_href option #{val} passed to ircify_html" if val
244 # Paragraph and br tags are converted to whitespace
245 txt.gsub!(/<\/?(p|br)(?:\s+[^>]*)?\s*\/?\s*>/i, ' ')
249 # Superscripts and subscripts are turned into ^{...} and _{...}
250 # where the {} are omitted for single characters
251 txt.gsub!(/<sup>(.*?)<\/sup>/, '^{\1}')
252 txt.gsub!(/<sub>(.*?)<\/sub>/, '_{\1}')
253 txt.gsub!(/(^|_)\{(.)\}/, '\1\2')
255 # List items are converted to *). We don't have special support for
256 # nested or ordered lists.
257 txt.gsub!(/<li>/, ' *) ')
259 # All other tags are just removed
260 txt.gsub!(/<[^>]+>/, '')
262 # Convert HTML entities. We do it now to be able to handle stuff
264 txt = Utils.decode_html_entities(txt)
266 # Keep unbreakable spaces or conver them to plain spaces?
267 case val = opts[:nbsp]
269 txt.gsub!([160].pack('U'), ' ')
271 warning "unknown :nbsp option #{val} passed to ircify_html" if val
274 # Remove double formatting options, since they only waste bytes
275 txt.gsub!(/#{Bold}(\s*)#{Bold}/, '\1')
276 txt.gsub!(/#{Underline}(\s*)#{Underline}/, '\1')
278 # Simplify whitespace that appears on both sides of a formatting option
279 txt.gsub!(/\s+(#{Bold}|#{Underline})\s+/, ' \1')
280 txt.sub!(/\s+(#{Bold}|#{Underline})\z/, '\1')
281 txt.sub!(/\A(#{Bold}|#{Underline})\s+/, '\1')
283 # And finally whitespace is squeezed
284 txt.gsub!(/\s+/, ' ')
287 if opts[:limit] && txt.size > opts[:limit]
288 txt = txt.slice(0, opts[:limit]) + "#{Reverse}...#{Reverse}"
291 # Decode entities and strip whitespace
295 # As above, but modify the receiver
297 def ircify_html!(opts={})
299 replace self.ircify_html(opts)
300 return self unless self.hash == old_hash
303 # This method will strip all HTML crud from the receiver
306 self.gsub(/<[^>]+>/, '').gsub(/&/,'&').gsub(/"/,'"').gsub(/</,'<').gsub(/>/,'>').gsub(/&ellip;/,'...').gsub(/'/, "'").gsub("\n",'')
309 # This method tries to find an HTML title in the string,
310 # and returns it if found
312 if defined? ::Hpricot
313 Hpricot(self).at("title").inner_html
315 return unless Irc::Utils::TITLE_REGEX.match(self)
320 # This method returns the IRC-formatted version of an
321 # HTML title found in the string
322 def ircify_html_title
323 self.get_html_title.ircify_html rescue nil
326 # This method is used to wrap a nonempty String by adding
327 # the prefix and postfix
328 def wrap_nonempty(pre, post, opts={})
332 "#{pre}#{self}#{post}"
338 # Extensions to the Regexp class, with some common and/or complex regular
343 # A method to build a regexp that matches a list of something separated by
344 # optional commas and/or the word "and", an optionally repeated prefix,
346 def Regexp.new_list(reg, pfx = "")
347 if pfx.kind_of?(String) and pfx.empty?
348 return %r(#{reg}(?:,?(?:\s+and)?\s+#{reg})*)
350 return %r(#{reg}(?:,?(?:\s+and)?(?:\s+#{pfx})?\s+#{reg})*)
357 # Match a list of channel anmes separated by optional commas, whitespace
358 # and optionally the word "and"
359 CHAN_LIST = Regexp.new_list(GEN_CHAN)
361 # Match "in #channel" or "on #channel" and/or "in private" (optionally
362 # shortened to "in pvt"), returning the channel name or the word 'private'
363 # or 'pvt' as capture
364 IN_CHAN = /#{IN_ON}\s+(#{GEN_CHAN})|(here)|/
365 IN_CHAN_PVT = /#{IN_CHAN}|in\s+(private|pvt)/
367 # As above, but with channel lists
368 IN_CHAN_LIST_SFX = Regexp.new_list(/#{GEN_CHAN}|here/, IN_ON)
369 IN_CHAN_LIST = /#{IN_ON}\s+#{IN_CHAN_LIST_SFX}|anywhere|everywhere/
370 IN_CHAN_LIST_PVT_SFX = Regexp.new_list(/#{GEN_CHAN}|here|private|pvt/, IN_ON)
371 IN_CHAN_LIST_PVT = /#{IN_ON}\s+#{IN_CHAN_LIST_PVT_SFX}|anywhere|everywhere/
373 # Match a list of nicknames separated by optional commas, whitespace and
374 # optionally the word "and"
375 NICK_LIST = Regexp.new_list(GEN_NICK)
385 class BasicUserMessage
387 # We extend the BasicUserMessage class with a method that parses a string
388 # which is a channel list as matched by IN_CHAN(_LIST) and co. The method
389 # returns an array of channel names, where 'private' or 'pvt' is replaced
390 # by the Symbol :"?", 'here' is replaced by the channel of the message or
391 # by :"?" (depending on whether the message target is the bot or a
392 # Channel), and 'anywhere' and 'everywhere' are replaced by Symbol :*
394 def parse_channel_list(string)
395 return [:*] if [:anywhere, :everywhere].include? string.to_sym
397 /(?:^|,?(?:\s+and)?\s+)(?:in|on\s+)?(#{Regexp::Irc::GEN_CHAN}|here|private|pvt)/
416 # The recurse depth of a message, for fake messages. 0 means an original
419 unless defined? @recurse_depth
425 # Set the recurse depth of a message, for fake messages. 0 should only
426 # be used by original messages
427 def recurse_depth=(val)
435 # Maximum fake message recursion
436 MAX_RECURSE_DEPTH = 10
438 class RecurseTooDeep < RuntimeError
442 # Sometimes plugins need to create a new fake message based on an existing
443 # message: for example, this is done by alias, linkbot, reaction and remotectl.
445 # This method simplifies the message creation, including a recursion depth
448 # In the options you can specify the :bot, the :server, the :source,
449 # the :target, the message :class and whether or not to :delegate. To
450 # initialize these entries from an existing message, you can use :from
452 # Additionally, if :from is given, the reply method of created message
453 # is overriden to reply to :from instead. The #in_thread attribute
454 # for created mesage is also copied from :from
456 # If you don't specify a :from you should specify a :source.
458 def fake_message(string, opts={})
459 if from = opts[:from]
461 :bot => from.bot, :server => from.server, :source => from.source,
462 :target => from.target, :class => from.class, :delegate => true,
463 :depth => from.recurse_depth + 1
467 :bot => @bot, :server => @bot.server, :target => @bot.myself,
468 :class => PrivMessage, :delegate => true, :depth => 1
471 raise RecurseTooDeep if o[:depth] > MAX_RECURSE_DEPTH
472 new_m = o[:class].new(o[:bot], o[:server], o[:source], o[:target], string)
473 new_m.recurse_depth = o[:depth]
475 # the created message will reply to the originating message
478 end.send(:define_method, :reply) do |*args|
479 debug "replying to '#{from.message}' with #{args.first}"
482 # the created message will follow originating message's in_thread
483 new_m.in_thread = from.in_thread if from.respond_to?(:in_thread)
485 return new_m unless o[:delegate]
486 method = o[:class].to_s.gsub(/^Irc::|Message$/,'').downcase
487 method = 'privmsg' if method == 'priv'
488 o[:bot].plugins.irc_delegate(method, new_m)