summaryrefslogtreecommitdiffstats
path: root/Lib/htmllib.py
blob: fe2a0cbc29800da6d52d6777a5dbd40dbd8c7af1 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
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
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
# A parser for HTML documents


# HTML: HyperText Markup Language; an SGML-like syntax used by WWW to
# describe hypertext documents
#
# SGML: Standard Generalized Markup Language
#
# WWW: World-Wide Web; a distributed hypertext system develped at CERN
#
# CERN: European Particle Physics Laboratory in Geneva, Switzerland


# This file is only concerned with parsing and formatting HTML
# documents, not with the other (hypertext and networking) aspects of
# the WWW project.  (It does support highlighting of anchors.)


import os
import sys
import regex
import string
import sgmllib


class HTMLParser(sgmllib.SGMLParser):

	# Copy base class entities and add some
	entitydefs = {}
	for key in sgmllib.SGMLParser.entitydefs.keys():
		entitydefs[key] = sgmllib.SGMLParser.entitydefs[key]
	entitydefs['bullet'] = '*'

	# Provided -- handlers for tags introducing literal text
	
	def start_listing(self, attrs):
		self.setliteral('listing')
		self.literal_bgn('listing', attrs)

	def end_listing(self):
		self.literal_end('listing')

	def start_xmp(self, attrs):
		self.setliteral('xmp')
		self.literal_bgn('xmp', attrs)

	def end_xmp(self):
		self.literal_end('xmp')

	def do_plaintext(self, attrs):
		self.setnomoretags()
		self.literal_bgn('plaintext', attrs)

	# To be overridden -- begin/end literal mode
	def literal_bgn(self, tag, attrs): pass
	def literal_end(self, tag): pass


# Next level of sophistication -- collect anchors, title, nextid and isindex
class CollectingParser(HTMLParser):
	#
	def __init__(self):
		HTMLParser.__init__(self)
		self.savetext = None
		self.nextid = []
		self.isindex = 0
		self.title = ''
		self.inanchor = 0
		self.anchors = []
		self.anchornames = []
		self.anchortypes = []
	#
	def start_a(self, attrs):
		self.inanchor = 0
		href = ''
		name = ''
		type = ''
		for attrname, value in attrs:
			if attrname == 'href':
				href = value
			if attrname == 'name=':
				name = value
			if attrname == 'type=':
				type = string.lower(value)
		if not (href or name):
			return
		self.anchors.append(href)
		self.anchornames.append(name)
		self.anchortypes.append(type)
		self.inanchor = len(self.anchors)
		if not href:
			self.inanchor = -self.inanchor
	#
	def end_a(self):
		if self.inanchor > 0:
			# Don't show anchors pointing into the current document
			if self.anchors[self.inanchor-1][:1] <> '#':
				self.handle_data('[' + `self.inanchor` + ']')
		self.inanchor = 0
	#
	def start_header(self, attrs): pass
	def end_header(self): pass
	#
	# (head is the same as header)
	def start_head(self, attrs): pass
	def end_head(self): pass
	#
	def start_body(self, attrs): pass
	def end_body(self): pass
	#
	def do_nextid(self, attrs):
		self.nextid = attrs
	#
	def do_isindex(self, attrs):
		self.isindex = 1
	#
	def start_title(self, attrs):
		self.savetext = ''
	#
	def end_title(self):
		if self.savetext <> None:
			self.title = self.savetext
			self.savetext = None
	#
	def handle_data(self, text):
		if self.savetext is not None:
			self.savetext = self.savetext + text


# Formatting parser -- takes a formatter and a style sheet as arguments

# XXX The use of style sheets should change: for each tag and end tag
# there should be a style definition, and a style definition should
# encompass many more parameters: font, justification, indentation,
# vspace before, vspace after, hanging tag...

wordprog = regex.compile('[^ \t\n]*')
spaceprog = regex.compile('[ \t\n]*')

class FormattingParser(CollectingParser):

	def __init__(self, formatter, stylesheet):
		CollectingParser.__init__(self)
		self.fmt = formatter
		self.stl = stylesheet
		self.savetext = None
		self.compact = 0
		self.nofill = 0
		self.resetfont()
		self.setindent(self.stl.stdindent)

	def resetfont(self):
		self.fontstack = []
		self.stylestack = []
		self.fontset = self.stl.stdfontset
		self.style = ROMAN
		self.passfont()

	def passfont(self):
		font = self.fontset[self.style]
		self.fmt.setfont(font)

	def pushstyle(self, style):
		self.stylestack.append(self.style)
		self.style = min(style, len(self.fontset)-1)
		self.passfont()

	def popstyle(self):
		self.style = self.stylestack[-1]
		del self.stylestack[-1]
		self.passfont()

	def pushfontset(self, fontset, style):
		self.fontstack.append(self.fontset)
		self.fontset = fontset
		self.pushstyle(style)

	def popfontset(self):
		self.fontset = self.fontstack[-1]
		del self.fontstack[-1]
		self.popstyle()

	def flush(self):
		self.fmt.flush()

	def setindent(self, n):
		self.fmt.setleftindent(n)

	def needvspace(self, n):
		self.fmt.needvspace(n)

	def close(self):
		HTMLParser.close(self)
		self.fmt.flush()

	def handle_literal(self, text):
		lines = string.splitfields(text, '\n')
		for i in range(1, len(lines)):
			lines[i] = string.expandtabs(lines[i], 8)
		for line in lines[:-1]:
			self.fmt.addword(line, 0)
			self.fmt.flush()
			self.fmt.nospace = 0
		for line in lines[-1:]:
			self.fmt.addword(line, 0)

	def handle_data(self, text):
		if self.savetext is not None:
			self.savetext = self.savetext + text
			return
		if self.literal:
			self.handle_literal(text)
			return
		i = 0
		n = len(text)
		while i < n:
			j = i + wordprog.match(text, i)
			word = text[i:j]
			i = j + spaceprog.match(text, j)
			self.fmt.addword(word, i-j)
			if self.nofill and '\n' in text[j:i]:
				self.fmt.flush()
				self.fmt.nospace = 0
				i = j+1
				while text[i-1] <> '\n': i = i+1

	def literal_bgn(self, tag, attrs):
		if tag == 'plaintext':
			self.flush()
		else:
			self.needvspace(1)
		self.pushfontset(self.stl.stdfontset, FIXED)
		self.setindent(self.stl.literalindent)

	def literal_end(self, tag):
		self.needvspace(1)
		self.popfontset()
		self.setindent(self.stl.stdindent)

	def start_title(self, attrs):
		self.flush()
		self.savetext = ''
	# NB end_title is unchanged

	def do_p(self, attrs):
		if self.compact:
			self.flush()
		else:
			self.needvspace(1)

	def start_h1(self, attrs):
		self.needvspace(2)
		self.setindent(self.stl.h1indent)
		self.pushfontset(self.stl.h1fontset, BOLD)
		self.fmt.setjust('c')

	def end_h1(self):
		self.popfontset()
		self.needvspace(2)
		self.setindent(self.stl.stdindent)
		self.fmt.setjust('l')

	def start_h2(self, attrs):
		self.needvspace(1)
		self.setindent(self.stl.h2indent)
		self.pushfontset(self.stl.h2fontset, BOLD)

	def end_h2(self):
		self.popfontset()
		self.needvspace(1)
		self.setindent(self.stl.stdindent)

	def start_h3(self, attrs):
		self.needvspace(1)
		self.setindent(self.stl.stdindent)
		self.pushfontset(self.stl.h3fontset, BOLD)

	def end_h3(self):
		self.popfontset()
		self.needvspace(1)
		self.setindent(self.stl.stdindent)

	def start_h4(self, attrs):
		self.needvspace(1)
		self.setindent(self.stl.stdindent)
		self.pushfontset(self.stl.stdfontset, BOLD)

	def end_h4(self):
		self.popfontset()
		self.needvspace(1)
		self.setindent(self.stl.stdindent)

	start_h5 = start_h4
	end_h5 = end_h4

	start_h6 = start_h5
	end_h6 = end_h5

	start_h7 = start_h6
	end_h7 = end_h6

	def start_ul(self, attrs):
		self.needvspace(1)
		for attrname, value in attrs:
			if attrname == 'compact':
				self.compact = 1
				self.setindent(0)
				break
		else:
			self.setindent(self.stl.ulindent)

	start_dir = start_menu = start_ol = start_ul

	do_li = do_p

	def end_ul(self):
		self.compact = 0
		self.needvspace(1)
		self.setindent(self.stl.stdindent)

	end_dir = end_menu = end_ol = end_ul

	def start_dl(self, attrs):
		for attrname, value in attrs:
			if attrname == 'compact':
				self.compact = 1
		self.needvspace(1)

	def end_dl(self):
		self.compact = 0
		self.needvspace(1)
		self.setindent(self.stl.stdindent)

	def do_dt(self, attrs):
		if self.compact:
			self.flush()
		else:
			self.needvspace(1)
		self.setindent(self.stl.stdindent)

	def do_dd(self, attrs):
		self.fmt.addword('', 1)
		self.setindent(self.stl.ddindent)

	def start_address(self, attrs):
		self.compact = 1
		self.needvspace(1)
		self.fmt.setjust('r')

	def end_address(self):
		self.compact = 0
		self.needvspace(1)
		self.setindent(self.stl.stdindent)
		self.fmt.setjust('l')

	def start_pre(self, attrs):
		self.needvspace(1)
		self.nofill = self.nofill + 1
		self.pushstyle(FIXED)

	def end_pre(self):
		self.popstyle()
		self.nofill = self.nofill - 1
		self.needvspace(1)

	start_typewriter = start_pre
	end_typewriter = end_pre

	def do_img(self, attrs):
		self.fmt.addword('(image)', 0)

	# Physical styles

	def start_tt(self, attrs): self.pushstyle(FIXED)
	def end_tt(self): self.popstyle()

	def start_b(self, attrs): self.pushstyle(BOLD)
	def end_b(self): self.popstyle()

	def start_i(self, attrs): self.pushstyle(ITALIC)
	def end_i(self): self.popstyle()

	def start_u(self, attrs): self.pushstyle(ITALIC) # Underline???
	def end_u(self): self.popstyle()

	def start_r(self, attrs): self.pushstyle(ROMAN) # Not official
	def end_r(self): self.popstyle()

	# Logical styles

	start_em = start_i
	end_em = end_i

	start_strong = start_b
	end_strong = end_b

	start_code = start_tt
	end_code = end_tt

	start_samp = start_tt
	end_samp = end_tt

	start_kbd = start_tt
	end_kbd = end_tt

	start_file = start_tt # unofficial
	end_file = end_tt

	start_var = start_i
	end_var = end_i

	start_dfn = start_i
	end_dfn = end_i

	start_cite = start_i
	end_cite = end_i

	start_hp1 = start_i
	end_hp1 = start_i

	start_hp2 = start_b
	end_hp2 = end_b

	def unknown_starttag(self, tag, attrs):
		print '*** unknown <' + tag + '>'

	def unknown_endtag(self, tag):
		print '*** unknown </' + tag + '>'


# An extension of the formatting parser which formats anchors differently.
class AnchoringParser(FormattingParser):

	def start_a(self, attrs):
		FormattingParser.start_a(self, attrs)
		if self.inanchor:
			self.fmt.bgn_anchor(self.inanchor)

	def end_a(self):
		if self.inanchor:
			self.fmt.end_anchor(self.inanchor)
			self.inanchor = 0


# Style sheet -- this is never instantiated, but the attributes
# of the class object itself are used to specify fonts to be used
# for various paragraph styles.
# A font set is a non-empty list of fonts, in the order:
# [roman, italic, bold, fixed].
# When a style is not available the nearest lower style is used

ROMAN = 0
ITALIC = 1
BOLD = 2
FIXED = 3

class NullStylesheet:
	# Fonts -- none
	stdfontset = [None]
	h1fontset = [None]
	h2fontset = [None]
	h3fontset = [None]
	# Indents
	stdindent = 2
	ddindent = 25
	ulindent = 4
	h1indent = 0
	h2indent = 0
	literalindent = 0


class X11Stylesheet(NullStylesheet):
	stdfontset = [
		'-*-helvetica-medium-r-normal-*-*-100-100-*-*-*-*-*',
		'-*-helvetica-medium-o-normal-*-*-100-100-*-*-*-*-*',
		'-*-helvetica-bold-r-normal-*-*-100-100-*-*-*-*-*',
		'-*-courier-medium-r-normal-*-*-100-100-*-*-*-*-*',
		]
	h1fontset = [
		'-*-helvetica-medium-r-normal-*-*-180-100-*-*-*-*-*',
		'-*-helvetica-medium-o-normal-*-*-180-100-*-*-*-*-*',
		'-*-helvetica-bold-r-normal-*-*-180-100-*-*-*-*-*',
		]
	h2fontset = [
		'-*-helvetica-medium-r-normal-*-*-140-100-*-*-*-*-*',
		'-*-helvetica-medium-o-normal-*-*-140-100-*-*-*-*-*',
		'-*-helvetica-bold-r-normal-*-*-140-100-*-*-*-*-*',
		]
	h3fontset = [
		'-*-helvetica-medium-r-normal-*-*-120-100-*-*-*-*-*',
		'-*-helvetica-medium-o-normal-*-*-120-100-*-*-*-*-*',
		'-*-helvetica-bold-r-normal-*-*-120-100-*-*-*-*-*',
		]
	ddindent = 40


class MacStylesheet(NullStylesheet):
	stdfontset = [
		('Geneva', 'p', 10),
		('Geneva', 'i', 10),
		('Geneva', 'b', 10),
		('Monaco', 'p', 10),
		]
	h1fontset = [
		('Geneva', 'p', 18),
		('Geneva', 'i', 18),
		('Geneva', 'b', 18),
		('Monaco', 'p', 18),
		]
	h3fontset = [
		('Geneva', 'p', 14),
		('Geneva', 'i', 14),
		('Geneva', 'b', 14),
		('Monaco', 'p', 14),
		]
	h3fontset = [
		('Geneva', 'p', 12),
		('Geneva', 'i', 12),
		('Geneva', 'b', 12),
		('Monaco', 'p', 12),
		]


if os.name == 'mac':
	StdwinStylesheet = MacStylesheet
else:
	StdwinStylesheet = X11Stylesheet


class GLStylesheet(NullStylesheet):
	stdfontset = [
		'Helvetica 10',
		'Helvetica-Italic 10',
		'Helvetica-Bold 10',
		'Courier 10',
		]
	h1fontset = [
		'Helvetica 18',
		'Helvetica-Italic 18',
		'Helvetica-Bold 18',
		'Courier 18',
		]
	h2fontset = [
		'Helvetica 14',
		'Helvetica-Italic 14',
		'Helvetica-Bold 14',
		'Courier 14',
		]
	h3fontset = [
		'Helvetica 12',
		'Helvetica-Italic 12',
		'Helvetica-Bold 12',
		'Courier 12',
		]


# Test program -- produces no output but times how long it takes
# to send a document to a null formatter, exclusive of I/O

def test():
	import fmt
	import time
	if sys.argv[1:]: file = sys.argv[1]
	else: file = 'test.html'
	data = open(file, 'r').read()
	t0 = time.time()
	fmtr = fmt.WritingFormatter(sys.stdout, 79)
	p = FormattingParser(fmtr, NullStylesheet)
	p.feed(data)
	p.close()
	t1 = time.time()
	print
	print '*** Formatting time:', round(t1-t0, 3), 'seconds.'


# Test program using stdwin

def testStdwin():
	import stdwin, fmt
	from stdwinevents import *
	if sys.argv[1:]: file = sys.argv[1]
	else: file = 'test.html'
	data = open(file, 'r').read()
	window = stdwin.open('testStdwin')
	b = None
	while 1:
		etype, ewin, edetail = stdwin.getevent()
		if etype == WE_CLOSE:
			break
		if etype == WE_SIZE:
			window.setdocsize(0, 0)
			window.setorigin(0, 0)
			window.change((0, 0), (10000, 30000)) # XXX
		if etype == WE_DRAW:
			if not b:
				b = fmt.StdwinBackEnd(window, 1)
				f = fmt.BaseFormatter(b.d, b)
				p = FormattingParser(f, MacStylesheet)
				p.feed(data)
				p.close()
				b.finish()
			else:
				b.redraw(edetail)
	window.close()


# Test program using GL

def testGL():
	import gl, GL, fmt
	if sys.argv[1:]: file = sys.argv[1]
	else: file = 'test.html'
	data = open(file, 'r').read()
	W, H = 600, 600
	gl.foreground()
	gl.prefsize(W, H)
	wid = gl.winopen('testGL')
	gl.ortho2(0, W, H, 0)
	gl.color(GL.WHITE)
	gl.clear()
	gl.color(GL.BLACK)
	b = fmt.GLBackEnd(wid)
	f = fmt.BaseFormatter(b.d, b)
	p = FormattingParser(f, GLStylesheet)
	p.feed(data)
	p.close()
	b.finish()
	#
	import time
	time.sleep(5)


if __name__ == '__main__':
	test()