mediacloth 0.0.2 → 0.0.3
Sign up to get free protection for your applications and to get access to all the features.
- data/lib/mediacloth/mediawikiast.rb +42 -0
- data/lib/mediacloth/mediawikihtmlgenerator.rb +100 -29
- data/lib/mediacloth/mediawikilexer.rb +292 -37
- data/lib/mediacloth/mediawikilexer.rb~ +491 -0
- data/lib/mediacloth/mediawikiparser.rb +535 -173
- data/lib/mediacloth/mediawikiparser.y +183 -15
- data/lib/mediacloth/mediawikiparser.y~ +210 -0
- data/lib/mediacloth/mediawikiwalker.rb +56 -8
- data/test/data/html1 +1 -1
- data/test/data/html10 +98 -0
- data/test/data/html3 +1 -1
- data/test/data/html4 +11 -1
- data/test/data/html5 +5 -1
- data/test/data/html7 +1 -2
- data/test/data/html8 +1 -1
- data/test/data/html9 +6 -0
- data/test/data/input1 +5 -0
- data/test/data/input10 +124 -0
- data/test/data/input4 +50 -1
- data/test/data/input5 +8 -0
- data/test/data/input7 +35 -2
- data/test/data/input9 +14 -0
- data/test/data/lex1 +5 -1
- data/test/data/lex10 +87 -0
- data/test/data/lex4 +47 -1
- data/test/data/lex5 +7 -1
- data/test/data/lex7 +35 -2
- data/test/data/lex9 +14 -0
- data/test/dataproducers/html.rb +2 -2
- data/test/dataproducers/html.rb~ +24 -0
- data/test/dataproducers/lex.rb +3 -3
- data/test/dataproducers/lex.rb~ +15 -0
- data/test/debugwalker.rb +1 -1
- data/test/htmlgenerator.rb +5 -4
- data/test/lexer.rb +40 -3
- data/test/parser.rb +0 -1
- metadata +14 -3
@@ -0,0 +1,491 @@
|
|
1
|
+
#The lexer for MediaWiki language.
|
2
|
+
#
|
3
|
+
#Standalone usage:
|
4
|
+
# file = File.new("somefile", "r")
|
5
|
+
# input = file.read
|
6
|
+
# lexer = MediaWikiLexer.new
|
7
|
+
# lexer.tokenize(input)
|
8
|
+
#
|
9
|
+
#Inside RACC-generated parser:
|
10
|
+
# ...
|
11
|
+
# ---- inner ----
|
12
|
+
# attr_accessor :lexer
|
13
|
+
# def parse(input)
|
14
|
+
# lexer.tokenize(input)
|
15
|
+
# return do_parse
|
16
|
+
# end
|
17
|
+
# def next_token
|
18
|
+
# return @lexer.lex
|
19
|
+
# end
|
20
|
+
# ...
|
21
|
+
# parser = MediaWikiParser.new
|
22
|
+
# parser.lexer = MediaWikiLexer.new
|
23
|
+
# parser.parse(input)
|
24
|
+
class MediaWikiLexer
|
25
|
+
|
26
|
+
#Initialized the lexer with a match table.
|
27
|
+
#
|
28
|
+
#The match table tells the lexer which method to invoke
|
29
|
+
#on given input char during "tokenize" phase.
|
30
|
+
def initialize
|
31
|
+
@position = 0
|
32
|
+
@pair_stack = [[false, false]] #stack of tokens for which a pair should be found
|
33
|
+
@list_stack = []
|
34
|
+
@lexer_table = Hash.new(method(:match_other))
|
35
|
+
@lexer_table["'"] = method(:match_italic_or_bold)
|
36
|
+
@lexer_table["="] = method(:match_section)
|
37
|
+
@lexer_table["["] = method(:match_link_start)
|
38
|
+
@lexer_table["]"] = method(:match_link_end)
|
39
|
+
@lexer_table[" "] = method(:match_space)
|
40
|
+
@lexer_table["*"] = method(:match_list)
|
41
|
+
@lexer_table["#"] = method(:match_list)
|
42
|
+
@lexer_table[";"] = method(:match_list)
|
43
|
+
@lexer_table[":"] = method(:match_list)
|
44
|
+
@lexer_table["-"] = method(:match_line)
|
45
|
+
@lexer_table["~"] = method(:match_signature)
|
46
|
+
@lexer_table["h"] = method(:match_inline_link)
|
47
|
+
@lexer_table["\n"] = method(:match_newline)
|
48
|
+
end
|
49
|
+
|
50
|
+
#Transforms input stream (string) into the stream of tokens.
|
51
|
+
#Tokens are collected into an array of type [ [TOKEN_SYMBOL, TOKEN_VALUE], ..., [false, false] ].
|
52
|
+
#This array can be given as input token-by token to RACC based parser with no
|
53
|
+
#modification. The last token [false, false] inficates EOF.
|
54
|
+
def tokenize(input)
|
55
|
+
@tokens = []
|
56
|
+
start_para
|
57
|
+
@cursor = 0
|
58
|
+
@text = input
|
59
|
+
@next_token = []
|
60
|
+
|
61
|
+
#This tokenizer algorithm assumes that everything that is not
|
62
|
+
#matched by the lexer is going to be :TEXT token. Otherwise it's usual
|
63
|
+
#lexer algo which call methods from the match table to define next tokens.
|
64
|
+
while (@cursor < @text.length)
|
65
|
+
@current_token = [:TEXT, ''] unless @current_token
|
66
|
+
@token_start = @cursor
|
67
|
+
@char = @text[@cursor, 1]
|
68
|
+
|
69
|
+
if @lexer_table[@char].call == :TEXT
|
70
|
+
@current_token[1] += @text[@token_start, 1]
|
71
|
+
else
|
72
|
+
#skip empty :TEXT tokens
|
73
|
+
puts "not a text: #{@next_token[0]}"
|
74
|
+
unless empty_text_token?
|
75
|
+
@tokens << @current_token
|
76
|
+
puts "chance to break para before #{@next_token[0]}"
|
77
|
+
unless para_breaker?(@next_token[0])
|
78
|
+
#if no paragraph was previously started
|
79
|
+
#then we should start it
|
80
|
+
start_para if !@para
|
81
|
+
else
|
82
|
+
#if we already have a paragraph this is the time to close it
|
83
|
+
end_para if @para
|
84
|
+
end
|
85
|
+
end
|
86
|
+
|
87
|
+
if para_breaker?(@next_token[0])
|
88
|
+
if @tokens.last and @tokens.last[0] == :PARA_START
|
89
|
+
#we need to remove para start token because no para end is possible
|
90
|
+
@tokens.pop
|
91
|
+
@para = false
|
92
|
+
end
|
93
|
+
end
|
94
|
+
|
95
|
+
@next_token[1] = @text[@token_start, @cursor - @token_start]
|
96
|
+
@tokens << @next_token
|
97
|
+
#hack to enable sub-lexing!
|
98
|
+
if @sub_tokens
|
99
|
+
@tokens += @sub_tokens
|
100
|
+
@sub_tokens = nil
|
101
|
+
end
|
102
|
+
#end of hack!
|
103
|
+
|
104
|
+
#if the next token can start the paragraph, let's try that
|
105
|
+
start_para if @tokens.last and para_starter?(@tokens.last[0])
|
106
|
+
|
107
|
+
@current_token = nil
|
108
|
+
@next_token = []
|
109
|
+
end
|
110
|
+
end
|
111
|
+
#add the last TEXT token if it exists
|
112
|
+
puts @current_token
|
113
|
+
if @current_token and not empty_text_token?
|
114
|
+
puts "here"
|
115
|
+
if para_breaker?(@current_token[0])
|
116
|
+
#if we already have a paragraph this is the time to close it
|
117
|
+
end_para if @para
|
118
|
+
end
|
119
|
+
@tokens << @current_token
|
120
|
+
end
|
121
|
+
|
122
|
+
#remove empty para start or finish the paragraph if necessary
|
123
|
+
if @tokens.last and @tokens.last[0] == :PARA_START
|
124
|
+
@tokens.pop
|
125
|
+
@para = false
|
126
|
+
else
|
127
|
+
end_para if @para
|
128
|
+
end
|
129
|
+
#RACC wants us to put this to indicate EOF
|
130
|
+
@tokens << [false, false]
|
131
|
+
@tokens
|
132
|
+
end
|
133
|
+
|
134
|
+
#Returns the next token from the stream. Useful for RACC parsers.
|
135
|
+
def lex
|
136
|
+
token = @tokens[@position]
|
137
|
+
@position += 1
|
138
|
+
return token
|
139
|
+
end
|
140
|
+
|
141
|
+
|
142
|
+
private
|
143
|
+
#Returns true if the token breaks the paragraph.
|
144
|
+
def para_breaker?(token)
|
145
|
+
[:SECTION_START, :SECTION_END,
|
146
|
+
:UL_START, :UL_END, :OL_START, :OL_END,
|
147
|
+
:DL_START, :DL_END, :HLINE, :PRE].include?(token)
|
148
|
+
end
|
149
|
+
|
150
|
+
#Returns true if the paragraph can be started after the token
|
151
|
+
def para_starter?(token)
|
152
|
+
[:SECTION_END, :UL_END, :OL_END, :DL_END, :HLINE, :PRE].include?(token)
|
153
|
+
end
|
154
|
+
|
155
|
+
#-- ================== Match methods ================== ++#
|
156
|
+
|
157
|
+
#Matches anything that was not matched. Returns :TEXT to indicate
|
158
|
+
#that matched characters should go into :TEXT token.
|
159
|
+
def match_other
|
160
|
+
@cursor += 1
|
161
|
+
return :TEXT
|
162
|
+
end
|
163
|
+
|
164
|
+
#Matches italic or bold symbols:
|
165
|
+
# "'''" { return :BOLD; }
|
166
|
+
# "''" { return :ITALIC; }
|
167
|
+
def match_italic_or_bold
|
168
|
+
if @text[@cursor, 3] == "'''" and @pair_stack.last[0] != :ITALICSTART
|
169
|
+
matchBold
|
170
|
+
@cursor += 3
|
171
|
+
return
|
172
|
+
end
|
173
|
+
if @text[@cursor, 2] == "''"
|
174
|
+
matchItalic
|
175
|
+
@cursor += 2
|
176
|
+
return
|
177
|
+
end
|
178
|
+
match_other
|
179
|
+
end
|
180
|
+
|
181
|
+
def matchBold
|
182
|
+
if @pair_stack.last[0] == :BOLDSTART
|
183
|
+
@next_token[0] = :BOLDEND
|
184
|
+
@pair_stack.pop
|
185
|
+
else
|
186
|
+
@next_token[0] = :BOLDSTART
|
187
|
+
@pair_stack.push @next_token
|
188
|
+
end
|
189
|
+
end
|
190
|
+
|
191
|
+
def matchItalic
|
192
|
+
if @pair_stack.last[0] == :ITALICSTART
|
193
|
+
@next_token[0] = :ITALICEND
|
194
|
+
@pair_stack.pop
|
195
|
+
else
|
196
|
+
@next_token[0] = :ITALICSTART
|
197
|
+
@pair_stack.push @next_token
|
198
|
+
end
|
199
|
+
end
|
200
|
+
|
201
|
+
#Matches sections
|
202
|
+
def match_section
|
203
|
+
if at_start_of_line? or (@pair_stack.last[0] == :SECTION_START)
|
204
|
+
i = 0
|
205
|
+
i += 1 while @text[@cursor+i, 1] == "="
|
206
|
+
@cursor += i
|
207
|
+
|
208
|
+
if @pair_stack.last[0] == :SECTION_START
|
209
|
+
@next_token[0] = :SECTION_END
|
210
|
+
@pair_stack.pop
|
211
|
+
else
|
212
|
+
@next_token[0] = :SECTION_START
|
213
|
+
@pair_stack.push @next_token
|
214
|
+
end
|
215
|
+
else
|
216
|
+
match_other
|
217
|
+
end
|
218
|
+
end
|
219
|
+
|
220
|
+
#Matches start of the hyperlinks
|
221
|
+
# "[[" { return INTLINKSTART; }
|
222
|
+
# "[" { return LINKSTART; }
|
223
|
+
def match_link_start
|
224
|
+
if @text[@cursor, 2] == "[["
|
225
|
+
@next_token[0] = :INTLINKSTART
|
226
|
+
@pair_stack.push @next_token
|
227
|
+
@cursor += 2
|
228
|
+
elsif @text[@cursor, 1] == "[" and html_link?(@cursor+1)
|
229
|
+
@next_token[0] = :LINKSTART
|
230
|
+
@pair_stack.push @next_token
|
231
|
+
@cursor += 1
|
232
|
+
else
|
233
|
+
match_other
|
234
|
+
end
|
235
|
+
end
|
236
|
+
|
237
|
+
#Matches end of the hyperlinks
|
238
|
+
# "]]" { return INTLINKEND; }
|
239
|
+
# "]" { return LINKEND; }
|
240
|
+
def match_link_end
|
241
|
+
if @text[@cursor, 2] == "]]" and @pair_stack.last[0] == :INTLINKSTART
|
242
|
+
@next_token[0] = :INTLINKEND
|
243
|
+
@pair_stack.pop
|
244
|
+
@cursor += 2
|
245
|
+
elsif @text[@cursor, 1] == "]" and @pair_stack.last[0] == :LINKSTART
|
246
|
+
@next_token[0] = :LINKEND
|
247
|
+
@pair_stack.pop
|
248
|
+
@cursor += 1
|
249
|
+
else
|
250
|
+
match_other
|
251
|
+
end
|
252
|
+
end
|
253
|
+
|
254
|
+
#Matches inlined unformatted html link
|
255
|
+
# "http://[^\s]*" { return [ LINKSTART TEXT LINKEND]; }
|
256
|
+
def match_inline_link
|
257
|
+
#if no link start token was detected and the text starts with http://
|
258
|
+
#then it's the inlined unformatted html link
|
259
|
+
if html_link?(@cursor) and @pair_stack.last[0] != :INTLINKSTART and
|
260
|
+
@pair_stack.last[0] != :LINKSTART
|
261
|
+
@next_token[0] = :LINKSTART
|
262
|
+
linkText = extract_till_whitespace
|
263
|
+
@sub_tokens = []
|
264
|
+
@sub_tokens << [:TEXT, linkText]
|
265
|
+
@sub_tokens << [:LINKEND, ']']
|
266
|
+
@cursor += linkText.length
|
267
|
+
@token_start = @cursor
|
268
|
+
else
|
269
|
+
match_other
|
270
|
+
end
|
271
|
+
end
|
272
|
+
|
273
|
+
#Matches space to find preformatted areas which start with a space after a newline
|
274
|
+
# "\n\s[^\n]*" { return PRE; }
|
275
|
+
def match_space
|
276
|
+
if at_start_of_line?
|
277
|
+
match_untill_eol
|
278
|
+
@next_token[0] = :PRE
|
279
|
+
strip_ws_from_token_start
|
280
|
+
else
|
281
|
+
match_other
|
282
|
+
end
|
283
|
+
end
|
284
|
+
|
285
|
+
#Matches any kind of list by using sublexing technique. MediaWiki lists are context-sensitive
|
286
|
+
#therefore we need to do some special processing with lists. The idea here is to strip
|
287
|
+
#the leftmost symbol indicating the list from the group of input lines and use separate
|
288
|
+
#lexer to process extracted fragment.
|
289
|
+
def match_list
|
290
|
+
if at_start_of_line?
|
291
|
+
list_id = @text[@cursor, 1]
|
292
|
+
sub_text = extract_list_contents(list_id)
|
293
|
+
extracted = 0
|
294
|
+
|
295
|
+
#hack to tokenize everything inside the list
|
296
|
+
@sub_tokens = []
|
297
|
+
sub_lines = ""
|
298
|
+
@sub_tokens << [:LI_START, ""]
|
299
|
+
sub_text.each do |t|
|
300
|
+
extracted += 1
|
301
|
+
if text_is_list? t
|
302
|
+
sub_lines += t
|
303
|
+
else
|
304
|
+
if not sub_lines.empty?
|
305
|
+
@sub_tokens += sub_lex(sub_lines)
|
306
|
+
sub_lines = ""
|
307
|
+
end
|
308
|
+
if @sub_tokens.last[0] != :LI_START
|
309
|
+
@sub_tokens << [:LI_END, ""]
|
310
|
+
@sub_tokens << [:LI_START, ""]
|
311
|
+
end
|
312
|
+
@sub_tokens += sub_lex(t.lstrip)
|
313
|
+
end
|
314
|
+
end
|
315
|
+
if not sub_lines.empty?
|
316
|
+
@sub_tokens += sub_lex(sub_lines)
|
317
|
+
@sub_tokens << [:LI_END, ""]
|
318
|
+
else
|
319
|
+
@sub_tokens << [:LI_END, ""]
|
320
|
+
end
|
321
|
+
|
322
|
+
#end of hack
|
323
|
+
@cursor += sub_text.length + extracted
|
324
|
+
@token_start = @cursor
|
325
|
+
|
326
|
+
case
|
327
|
+
when list_id == "*"
|
328
|
+
@next_token[0] = :UL_START
|
329
|
+
@sub_tokens << [:UL_END, ""]
|
330
|
+
when list_id == "#"
|
331
|
+
@next_token[0] = :OL_START
|
332
|
+
@sub_tokens << [:OL_END, ""]
|
333
|
+
when list_id == ";", list_id == ":"
|
334
|
+
@next_token[0] = :DL_START
|
335
|
+
@sub_tokens << [:DL_END, ""]
|
336
|
+
end
|
337
|
+
|
338
|
+
else
|
339
|
+
match_other
|
340
|
+
end
|
341
|
+
end
|
342
|
+
|
343
|
+
#Matches the line until \n
|
344
|
+
def match_untill_eol
|
345
|
+
val = @text[@cursor, 1]
|
346
|
+
while (val != "\n") and (!val.nil?)
|
347
|
+
@cursor += 1
|
348
|
+
val = @text[@cursor, 1]
|
349
|
+
end
|
350
|
+
@cursor += 1
|
351
|
+
end
|
352
|
+
|
353
|
+
#Matches hline tag that start with "-"
|
354
|
+
# "\n----" { return HLINE; }
|
355
|
+
def match_line
|
356
|
+
if at_start_of_line? and @text[@cursor, 4] == "----"
|
357
|
+
@next_token[0] = :HLINE
|
358
|
+
@cursor += 4
|
359
|
+
else
|
360
|
+
match_other
|
361
|
+
end
|
362
|
+
end
|
363
|
+
|
364
|
+
#Matches signature
|
365
|
+
# "~~~~~" { return SIGNATURE_DATE; }
|
366
|
+
# "~~~~" { return SIGNATURE_FULL; }
|
367
|
+
# "~~~" { return SIGNATURE_NAME; }
|
368
|
+
def match_signature
|
369
|
+
if @text[@cursor, 5] == "~~~~~"
|
370
|
+
@next_token[0] = :SIGNATURE_DATE
|
371
|
+
@cursor += 5
|
372
|
+
elsif @text[@cursor, 4] == "~~~~"
|
373
|
+
@next_token[0] = :SIGNATURE_FULL
|
374
|
+
@cursor += 4
|
375
|
+
elsif @text[@cursor, 3] == "~~~"
|
376
|
+
@next_token[0] = :SIGNATURE_NAME
|
377
|
+
@cursor += 3
|
378
|
+
else
|
379
|
+
match_other
|
380
|
+
end
|
381
|
+
end
|
382
|
+
|
383
|
+
#Matches new line and breaks the paragraph if two newlines are met
|
384
|
+
def match_newline
|
385
|
+
if @text[@cursor, 2] == "\n\n"
|
386
|
+
if @para
|
387
|
+
@next_token[0] = :PARA_END
|
388
|
+
# @para = false
|
389
|
+
@sub_tokens = [[:PARA_START, ""]]
|
390
|
+
@cursor += 2
|
391
|
+
return
|
392
|
+
end
|
393
|
+
end
|
394
|
+
match_other
|
395
|
+
end
|
396
|
+
|
397
|
+
#-- ================== Helper methods ================== ++#
|
398
|
+
|
399
|
+
#Checks if the token is placed at the start of the line.
|
400
|
+
def at_start_of_line?
|
401
|
+
if @cursor == 0 or @text[@cursor-1, 1] == "\n"
|
402
|
+
true
|
403
|
+
else
|
404
|
+
false
|
405
|
+
end
|
406
|
+
end
|
407
|
+
|
408
|
+
#Checks if the text at position contains the start of the html link
|
409
|
+
def html_link?(position)
|
410
|
+
return @text[position, 7] == 'http://'
|
411
|
+
end
|
412
|
+
|
413
|
+
#Adjusts @token_start to skip leading whitespaces
|
414
|
+
def strip_ws_from_token_start
|
415
|
+
@token_start += 1 while @text[@token_start, 1] == " "
|
416
|
+
end
|
417
|
+
|
418
|
+
#Returns true if the TEXT token is empty or contains newline only
|
419
|
+
def empty_text_token?
|
420
|
+
@current_token == [:TEXT, ''] or @current_token == [:TEXT, "\n"]
|
421
|
+
end
|
422
|
+
|
423
|
+
#Returns true if the text is a list, i.e. starts with one of #;*: symbols
|
424
|
+
#that indicate a list
|
425
|
+
def text_is_list?(text)
|
426
|
+
return text =~ /^[#;*:].*/
|
427
|
+
end
|
428
|
+
|
429
|
+
#Runs sublexer to tokenize sub_text
|
430
|
+
def sub_lex(sub_text, strip_paragraphs=true)
|
431
|
+
sub_lexer = MediaWikiLexer.new
|
432
|
+
sub_tokens = sub_lexer.tokenize(sub_text)
|
433
|
+
sub_tokens.pop #false token
|
434
|
+
if strip_paragraphs
|
435
|
+
#the last PARA_END token
|
436
|
+
sub_tokens.pop if sub_tokens.last[0] == :PARA_END
|
437
|
+
#the first PARA_START token
|
438
|
+
sub_tokens.delete_at(0) if sub_tokens[0][0] == :PARA_START
|
439
|
+
end
|
440
|
+
sub_tokens
|
441
|
+
end
|
442
|
+
|
443
|
+
#Extracts the text from current cursor position till the next whitespace
|
444
|
+
def extract_till_whitespace
|
445
|
+
i = @cursor
|
446
|
+
text = ""
|
447
|
+
while i < @text.length
|
448
|
+
curr = @text[i, 1]
|
449
|
+
if (curr == "\n") or (curr == "\t") or (curr == " ")
|
450
|
+
break
|
451
|
+
end
|
452
|
+
text += curr
|
453
|
+
i += 1
|
454
|
+
end
|
455
|
+
text
|
456
|
+
end
|
457
|
+
|
458
|
+
#Extract list contents of list type set by list_id variable.
|
459
|
+
#Example list:
|
460
|
+
# *a
|
461
|
+
# **a
|
462
|
+
#Extracted list with id "*" will look like:
|
463
|
+
# a
|
464
|
+
# *a
|
465
|
+
def extract_list_contents(list_id)
|
466
|
+
i = @cursor+1
|
467
|
+
list = ""
|
468
|
+
while i < @text.length
|
469
|
+
curr = @text[i, 1]
|
470
|
+
if (curr == "\n") and (@text[i+1, 1] != list_id)
|
471
|
+
list+=curr
|
472
|
+
break
|
473
|
+
end
|
474
|
+
list += curr unless (curr == list_id) and (@text[i-1, 1] == "\n")
|
475
|
+
i += 1
|
476
|
+
end
|
477
|
+
list
|
478
|
+
end
|
479
|
+
|
480
|
+
def start_para
|
481
|
+
@tokens << [:PARA_START, ""]
|
482
|
+
@para = true
|
483
|
+
end
|
484
|
+
|
485
|
+
def end_para
|
486
|
+
@tokens << [:PARA_END, ""]
|
487
|
+
@para = false
|
488
|
+
end
|
489
|
+
|
490
|
+
end
|
491
|
+
|