Class: Rouge::Lexer Abstract

Inherits:
Object
  • Object
show all
Includes:
Token::Tokens
Defined in:
lib/rouge/lexer.rb

Overview

This class is abstract.

A lexer transforms text into a stream of `[token, chunk]` pairs.

Direct Known Subclasses

Rouge::Lexers::PlainText, RegexLexer

Defined Under Namespace

Classes: AmbiguousGuess

Constant Summary

Constants included from Token::Tokens

Token::Tokens::Num, Token::Tokens::Str

Class Method Summary collapse

Instance Method Summary collapse

Methods included from Token::Tokens

token

Constructor Details

#initialize(opts = {}) ⇒ Lexer

Create a new lexer with the given options. Individual lexers may specify extra options. The only current globally accepted option is `:debug`.

Parameters:

  • opts (Hash) (defaults to: {})

    a customizable set of options

Options Hash (opts):

  • :debug (Object)

    Prints debug information to stdout. The particular info depends on the lexer in question. In regex lexers, this will log the state stack at the beginning of each step, along with each regex tried and each stream consumed. Try it, it's pretty useful.


316
317
318
# File 'lib/rouge/lexer.rb', line 316

def initialize(opts={})
  options(opts)
end

Class Method Details

.aliases(*args) ⇒ Object

Used to specify alternate names this lexer class may be found by.

Examples:

class Erb < Lexer
  tag 'erb'
  aliases 'eruby', 'rhtml'
end

Lexer.find('eruby') # => Erb

264
265
266
267
268
# File 'lib/rouge/lexer.rb', line 264

def aliases(*args)
  args.map!(&:to_s)
  args.each { |arg| Lexer.register(arg, self) }
  (@aliases ||= []).concat(args)
end

.allObject

Returns a list of all lexers.

Returns:

  • a list of all lexers.


90
91
92
# File 'lib/rouge/lexer.rb', line 90

def all
  registry.values.uniq
end

.analyze_text(text) ⇒ Object

This method is abstract.

Return a number between 0 and 1 indicating the likelihood that the text given should be lexed with this lexer. The default implementation returns 0. Values under 0.5 will only be used to disambiguate filename or mimetype matches.

Parameters:


421
422
423
# File 'lib/rouge/lexer.rb', line 421

def self.analyze_text(text)
  0
end

.assert_utf8!(str) ⇒ Object

Raises:

  • (EncodingError)

291
292
293
294
295
296
297
# File 'lib/rouge/lexer.rb', line 291

def assert_utf8!(str)
  return if %w(US-ASCII UTF-8).include? str.encoding.name
  raise EncodingError.new(
    "Bad encoding: #{str.encoding.names.join(',')}. " +
    "Please convert your string to UTF-8."
  )
end

.default_options(o = {}) ⇒ Object


20
21
22
23
24
# File 'lib/rouge/lexer.rb', line 20

def default_options(o={})
  @default_options ||= {}
  @default_options.merge!(o)
  @default_options
end

.demo(arg = :absent) ⇒ Object

Specify or get a small demo string for this lexer


83
84
85
86
87
# File 'lib/rouge/lexer.rb', line 83

def demo(arg=:absent)
  return @demo = arg unless arg == :absent

  @demo = File.read(demo_file)
end

.demo_file(arg = :absent) ⇒ Object

Specify or get the path name containing a small demo for this lexer (can be overriden by demo).


76
77
78
79
80
# File 'lib/rouge/lexer.rb', line 76

def demo_file(arg=:absent)
  return @demo_file = Pathname.new(arg) unless arg == :absent

  @demo_file = Pathname.new(__FILE__).dirname.join('demos', tag)
end

.desc(arg = :absent) ⇒ Object

Specify or get this lexer's description.


66
67
68
69
70
71
72
# File 'lib/rouge/lexer.rb', line 66

def desc(arg=:absent)
  if arg == :absent
    @desc
  else
    @desc = arg
  end
end

.filenames(*fnames) ⇒ Object

Specify a list of filename globs associated with this lexer.

Examples:

class Ruby < Lexer
  filenames '*.rb', '*.ruby', 'Gemfile', 'Rakefile'
end

276
277
278
# File 'lib/rouge/lexer.rb', line 276

def filenames(*fnames)
  (@filenames ||= []).concat(fnames)
end

.find(name) ⇒ Object

Given a string, return the correct lexer class.


27
28
29
# File 'lib/rouge/lexer.rb', line 27

def find(name)
  registry[name.to_s]
end

.find_fancy(str, code = nil) ⇒ Object

Find a lexer, with fancy shiny features.

  • The string you pass can include CGI-style options

    Lexer.find_fancy('erb?parent=tex')
    
  • You can pass the special name 'guess' so we guess for you, and you can pass a second argument of the code to guess by

    Lexer.find_fancy('guess', "#!/bin/bash\necho Hello, world")
    

This is used in the Redcarpet plugin as well as Rouge's own markdown lexer for highlighting internal code blocks.


45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
# File 'lib/rouge/lexer.rb', line 45

def find_fancy(str, code=nil)
  name, opts = str ? str.split('?', 2) : [nil, '']

  # parse the options hash from a cgi-style string
  opts = CGI.parse(opts || '').map do |k, vals|
    [ k.to_sym, vals.empty? ? true : vals[0] ]
  end

  opts = Hash[opts]

  lexer_class = case name
  when 'guess', nil
    self.guess(:source => code, :mimetype => opts[:mimetype])
  when String
    self.find(name)
  end

  lexer_class && lexer_class.new(opts)
end

.guess(info = {}) ⇒ Object

Guess which lexer to use based on a hash of info.

Parameters:

  • info (Hash) (defaults to: {})

    a customizable set of options

Options Hash (info):

  • :mimetype (Object)

    A mimetype to guess by

  • :filename (Object)

    A filename to guess by

  • :source (Object)

    The source itself, which, if guessing by mimetype or filename fails, will be searched for shebangs, <!DOCTYPE …> tags, and other hints.

Raises:

See Also:


143
144
145
146
147
148
149
150
# File 'lib/rouge/lexer.rb', line 143

def guess(info={})
  lexers = guesses(info)

  return Lexers::PlainText if lexers.empty?
  return lexers[0] if lexers.size == 1

  raise AmbiguousGuess.new(lexers)
end

.guess_by_filename(fname) ⇒ Object


156
157
158
# File 'lib/rouge/lexer.rb', line 156

def guess_by_filename(fname)
  guess :filename => fname
end

.guess_by_mimetype(mt) ⇒ Object


152
153
154
# File 'lib/rouge/lexer.rb', line 152

def guess_by_mimetype(mt)
  guess :mimetype => mt
end

.guess_by_source(source) ⇒ Object


160
161
162
# File 'lib/rouge/lexer.rb', line 160

def guess_by_source(source)
  guess :source => source
end

.guesses(info = {}) ⇒ Object

Guess which lexer to use based on a hash of info.

This accepts the same arguments as Lexer.guess, but will never throw an error. It will return a (possibly empty) list of potential lexers to use.


99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
# File 'lib/rouge/lexer.rb', line 99

def guesses(info={})
  mimetype, filename, source = info.values_at(:mimetype, :filename, :source)
  lexers = registry.values.uniq
  total_size = lexers.size

  lexers = filter_by_mimetype(lexers, mimetype) if mimetype
  return lexers if lexers.size == 1

  lexers = filter_by_filename(lexers, filename) if filename
  return lexers if lexers.size == 1

  if source
    # If we're filtering against *all* lexers, we only use confident return
    # values from analyze_text.  But if we've filtered down already, we can trust
    # the analysis more.
    source_threshold = lexers.size < total_size ? 0 : 0.5
    return [best_by_source(lexers, source, source_threshold)].compact
  end

  []
end

.lex(stream, opts = {}, &b) ⇒ Object

Lexes `stream` with the given options. The lex is delegated to a new instance.

See Also:


16
17
18
# File 'lib/rouge/lexer.rb', line 16

def lex(stream, opts={}, &b)
  new(opts).lex(stream, &b)
end

.mimetypes(*mts) ⇒ Object

Specify a list of mimetypes associated with this lexer.

Examples:

class Html < Lexer
  mimetypes 'text/html', 'application/xhtml+xml'
end

286
287
288
# File 'lib/rouge/lexer.rb', line 286

def mimetypes(*mts)
  (@mimetypes ||= []).concat(mts)
end

.tag(t = nil) ⇒ Object

Used to specify or get the canonical name of this lexer class.

Examples:

class MyLexer < Lexer
  tag 'foo'
end

MyLexer.tag # => 'foo'

Lexer.find('foo') # => MyLexer

248
249
250
251
252
253
# File 'lib/rouge/lexer.rb', line 248

def tag(t=nil)
  return @tag if t.nil?

  @tag = t.to_s
  Lexer.register(@tag, self)
end

Instance Method Details

#debug(&b) ⇒ Object

Leave a debug message if the `:debug` option is set. The message is given as a block because some debug messages contain calculated information that is unnecessary for lexing in the real world.

Examples:

debug { "hello, world!" }

342
343
344
345
346
347
348
349
350
351
352
353
354
355
# File 'lib/rouge/lexer.rb', line 342

def debug(&b)
  # This method is a hotspot, unfortunately.
  #
  # For performance reasons, the "debug" option of a lexer cannot
  # be changed once it has begun lexing.  This method will redefine
  # itself on the first call to a noop if "debug" is not set.
  if option(:debug)
    def self.debug; puts yield; end
  else
    def self.debug; end
  end

  debug(&b)
end

#lex(string, opts = {}, &b) ⇒ Object

Given a string, yield [token, chunk] pairs. If no block is given, an enumerator is returned.

Parameters:

  • opts (Hash) (defaults to: {})

    a customizable set of options

Options Hash (opts):

  • :continue (Object)

    Continue the lex from the previous state (i.e. don't call #reset!)


369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
# File 'lib/rouge/lexer.rb', line 369

def lex(string, opts={}, &b)
  return enum_for(:lex, string) unless block_given?

  Lexer.assert_utf8!(string)

  reset! unless opts[:continue]

  # consolidate consecutive tokens of the same type
  last_token = nil
  last_val = nil
  stream_tokens(string) do |tok, val|
    next if val.empty?

    if tok == last_token
      last_val << val
      next
    end

    b.call(last_token, last_val) if last_token
    last_token = tok
    last_val = val
  end

  b.call(last_token, last_val) if last_token
end

#option(k, v = :absent) ⇒ Object

get or specify one option for this lexer


328
329
330
331
332
333
334
# File 'lib/rouge/lexer.rb', line 328

def option(k, v=:absent)
  if v == :absent
    options[k]
  else
    options({ k => v })
  end
end

#options(o = {}) ⇒ Object

get and/or specify the options for this lexer.


321
322
323
324
325
# File 'lib/rouge/lexer.rb', line 321

def options(o={})
  (@options ||= {}).merge!(o)

  self.class.default_options.merge(@options)
end

#reset!Object

This method is abstract.

Called after each lex is finished. The default implementation is a noop.


361
362
# File 'lib/rouge/lexer.rb', line 361

def reset!
end

#stream_tokens(stream, &b) ⇒ Object

This method is abstract.

Yield `[token, chunk]` pairs, given a prepared input stream. This must be implemented.

Parameters:

  • stream (StringScanner)

    the stream


407
408
409
# File 'lib/rouge/lexer.rb', line 407

def stream_tokens(stream, &b)
  raise 'abstract'
end

#tagObject

delegated to tag


396
397
398
# File 'lib/rouge/lexer.rb', line 396

def tag
  self.class.tag
end