1 | #!/usr/bin/env python2
|
2 | """
|
3 | lazylex/html.py - Low-Level HTML Processing.
|
4 |
|
5 | See lazylex/README.md for details.
|
6 |
|
7 | TODO: This should be an Oils library eventually. It's a "lazily-parsed data
|
8 | structure" like TSV8
|
9 | """
|
10 | from __future__ import print_function
|
11 |
|
12 | import cStringIO
|
13 | import re
|
14 | import sys
|
15 |
|
16 |
|
17 | def log(msg, *args):
|
18 | msg = msg % args
|
19 | print(msg, file=sys.stderr)
|
20 |
|
21 |
|
22 | class LexError(Exception):
|
23 | """For bad lexical elements like <> or &&"""
|
24 |
|
25 | def __init__(self, s, pos):
|
26 | self.s = s
|
27 | self.pos = pos
|
28 |
|
29 | def __str__(self):
|
30 | return '(LexError %r)' % (self.s[self.pos:self.pos + 20])
|
31 |
|
32 |
|
33 | class ParseError(Exception):
|
34 | """For errors in the tag structure."""
|
35 |
|
36 | def __init__(self, msg, *args):
|
37 | self.msg = msg
|
38 | self.args = args
|
39 |
|
40 | def __str__(self):
|
41 | return '(ParseError %s)' % (self.msg % self.args)
|
42 |
|
43 |
|
44 | class Output(object):
|
45 | """Takes an underlying input buffer and an output file. Maintains a
|
46 | position in the input buffer.
|
47 |
|
48 | Print FROM the input or print new text to the output.
|
49 | """
|
50 |
|
51 | def __init__(self, s, f, left_pos=0, right_pos=0):
|
52 | self.s = s
|
53 | self.f = f
|
54 | self.pos = left_pos
|
55 | if right_pos == 0:
|
56 | self.right_pos = len(s)
|
57 | else:
|
58 | self.right_pos = right_pos
|
59 |
|
60 | def SkipTo(self, pos):
|
61 | """Skip to a position."""
|
62 | self.pos = pos
|
63 |
|
64 | def PrintUntil(self, pos):
|
65 | """Print until a position."""
|
66 | piece = self.s[self.pos:pos]
|
67 | self.f.write(piece)
|
68 | self.pos = pos
|
69 |
|
70 | def PrintTheRest(self):
|
71 | """Print until the end of the string."""
|
72 | self.PrintUntil(self.right_pos)
|
73 |
|
74 | def Print(self, s):
|
75 | """Print text to the underlying buffer."""
|
76 | self.f.write(s)
|
77 |
|
78 |
|
79 | # HTML Tokens
|
80 | (Decl, Comment, Processing, StartTag, StartEndTag, EndTag, DecChar, HexChar,
|
81 | CharEntity, RawData, Invalid, EndOfStream) = range(12)
|
82 |
|
83 |
|
84 | def MakeLexer(rules):
|
85 | return [
|
86 | # DOTALL is for the comment
|
87 | (re.compile(pat, re.VERBOSE | re.DOTALL), i) for (pat, i) in rules
|
88 | ]
|
89 |
|
90 |
|
91 | #
|
92 | # Eggex
|
93 | #
|
94 | # Tag = / ~['>']+ /
|
95 |
|
96 | # Is this valid? A single character?
|
97 | # Tag = / ~'>'* /
|
98 |
|
99 | # Maybe better: / [NOT '>']+/
|
100 | # capital letters not allowed there?
|
101 | #
|
102 | # But then this is confusing:
|
103 | # / [NOT ~digit]+/
|
104 | #
|
105 | # / [NOT digit] / is [^\d]
|
106 | # / ~digit / is \D
|
107 | #
|
108 | # Or maybe:
|
109 | #
|
110 | # / [~ digit]+ /
|
111 | # / [~ '>']+ /
|
112 | # / [NOT '>']+ /
|
113 |
|
114 | # End = / '</' Tag '>' /
|
115 | # StartEnd = / '<' Tag '/>' /
|
116 | # Start = / '<' Tag '>' /
|
117 | #
|
118 | # EntityRef = / '&' dot{* N} ';' /
|
119 |
|
120 | LEXER = [
|
121 | # TODO: instead of nongreedy matches, the loop can just do .find('-->') and
|
122 | # .find('?>')
|
123 |
|
124 | # Actually non-greedy matches are regular and can be matched in linear time
|
125 | # with RE2.
|
126 | #
|
127 | # https://news.ycombinator.com/item?id=27099798
|
128 | #
|
129 | # Maybe try combining all of these for speed.
|
130 | (r'<!-- .*? -->', Comment),
|
131 | (r'<\? .*? \?>', Processing),
|
132 |
|
133 | # NOTE: < is allowed in these.
|
134 | (r'<! [^>]+ >', Decl), # <!DOCTYPE html>
|
135 | (r'</ [^>]+ >', EndTag), # self-closing <br/> comes FIRST
|
136 | (r'< [^>]+ />', StartEndTag), # end </a>
|
137 | (r'< [^>]+ >', StartTag), # start <a>
|
138 | (r'&\# [0-9]+ ;', DecChar),
|
139 | (r'&\# x[0-9a-fA-F]+ ;', HexChar),
|
140 | (r'& [a-zA-Z]+ ;', CharEntity),
|
141 |
|
142 | # Note: > is allowed in raw data.
|
143 | # https://stackoverflow.com/questions/10462348/right-angle-bracket-in-html
|
144 | (r'[^&<]+', RawData),
|
145 | (r'.', Invalid), # error!
|
146 | ]
|
147 |
|
148 | LEXER = MakeLexer(LEXER)
|
149 |
|
150 |
|
151 | def _Tokens(s, left_pos, right_pos):
|
152 | """
|
153 | Args:
|
154 | s: string to parse
|
155 | left_pos, right_pos: Optional span boundaries.
|
156 | """
|
157 | pos = left_pos
|
158 | if right_pos == 0:
|
159 | n = len(s)
|
160 | else:
|
161 | n = right_pos
|
162 |
|
163 | while pos < n:
|
164 | # Find the FIRST pattern that matches.
|
165 | for pat, tok_id in LEXER:
|
166 | m = pat.match(s, pos)
|
167 | if m:
|
168 | end_pos = m.end()
|
169 | yield tok_id, end_pos
|
170 | pos = end_pos
|
171 | break
|
172 |
|
173 | # Zero length sentinel
|
174 | yield EndOfStream, pos
|
175 |
|
176 |
|
177 | def ValidTokens(s, left_pos=0, right_pos=0):
|
178 | """Wrapper around _Tokens to prevent callers from having to handle Invalid.
|
179 |
|
180 | I'm not combining the two functions because I might want to do a
|
181 | 'yield' transformation on Tokens()? Exceptions might complicate the
|
182 | issue?
|
183 | """
|
184 | pos = left_pos
|
185 | for tok_id, end_pos in _Tokens(s, left_pos, right_pos):
|
186 | if tok_id == Invalid:
|
187 | raise LexError(s, pos)
|
188 | yield tok_id, end_pos
|
189 | pos = end_pos
|
190 |
|
191 |
|
192 | # To match <a or </a
|
193 | # <h2 but not <2h ?
|
194 | _TAG_RE = re.compile(r'/? \s* ([a-zA-Z][a-zA-Z0-9]*)', re.VERBOSE)
|
195 |
|
196 | # To match href="foo"
|
197 |
|
198 | _ATTR_RE = re.compile(
|
199 | r'''
|
200 | \s+ # Leading whitespace is required
|
201 | ([a-z]+) # Attribute name
|
202 | (?: # Optional attribute value
|
203 | \s* = \s*
|
204 | (?:
|
205 | " ([^>"]*) " # double quoted value
|
206 | | ([a-zA-Z0-9_\-]+) # Just allow unquoted "identifiers"
|
207 | # TODO: relax this? for href=$foo
|
208 | )
|
209 | )?
|
210 | ''', re.VERBOSE)
|
211 |
|
212 | TagName, AttrName, UnquotedValue, QuotedValue = range(4)
|
213 |
|
214 |
|
215 | class TagLexer(object):
|
216 | """
|
217 | Given a tag like <a href="..."> or <link type="..." />, the TagLexer
|
218 | provides a few operations:
|
219 |
|
220 | - What is the tag?
|
221 | - Iterate through the attributes, giving (name, value_start_pos, value_end_pos)
|
222 | """
|
223 |
|
224 | def __init__(self, s):
|
225 | self.s = s
|
226 | self.start_pos = -1 # Invalid
|
227 | self.end_pos = -1
|
228 |
|
229 | def Reset(self, start_pos, end_pos):
|
230 | self.start_pos = start_pos
|
231 | self.end_pos = end_pos
|
232 |
|
233 | def TagString(self):
|
234 | return self.s[self.start_pos:self.end_pos]
|
235 |
|
236 | def TagName(self):
|
237 | # First event
|
238 | tok_id, start, end = next(self.Tokens())
|
239 | return self.s[start:end]
|
240 |
|
241 | def GetSpanForAttrValue(self, attr_name):
|
242 | # Algorithm: search for QuotedValue or UnquotedValue after AttrName
|
243 | # TODO: Could also cache these
|
244 |
|
245 | events = self.Tokens()
|
246 | val = (-1, -1)
|
247 | try:
|
248 | while True:
|
249 | tok_id, start, end = next(events)
|
250 | if tok_id == AttrName:
|
251 | name = self.s[start:end]
|
252 | if name == attr_name:
|
253 | # For HasAttr()
|
254 | #val = True
|
255 |
|
256 | # Now try to get a real value
|
257 | tok_id, start, end = next(events)
|
258 | if tok_id in (QuotedValue, UnquotedValue):
|
259 |
|
260 | # TODO: Unescape this with htmlentitydefs
|
261 | # I think we need another lexer!
|
262 | #
|
263 | # We could make a single pass?
|
264 | # Shortcut: 'if '&' in substring'
|
265 | # Then we need to unescape it
|
266 |
|
267 | val = start, end
|
268 | break
|
269 |
|
270 | except StopIteration:
|
271 | pass
|
272 | return val
|
273 |
|
274 | def GetAttr(self, attr_name):
|
275 | # Algorithm: search for QuotedValue or UnquotedValue after AttrName
|
276 | # TODO: Could also cache these
|
277 | start, end = self.GetSpanForAttrValue(attr_name)
|
278 | if start == -1:
|
279 | return None
|
280 | return self.s[start:end]
|
281 |
|
282 | def Tokens(self):
|
283 | """
|
284 | Yields a sequence of tokens: Tag (AttrName AttrValue?)*
|
285 |
|
286 | Where each Token is (Type, start_pos, end_pos)
|
287 |
|
288 | Note that start and end are NOT redundant! We skip over some unwanted
|
289 | characters.
|
290 | """
|
291 | m = _TAG_RE.match(self.s, self.start_pos + 1)
|
292 | if not m:
|
293 | raise RuntimeError('Invalid HTML tag: %r' % self.TagString())
|
294 | yield TagName, m.start(1), m.end(1)
|
295 |
|
296 | pos = m.end(0)
|
297 |
|
298 | while True:
|
299 | # don't search past the end
|
300 | m = _ATTR_RE.match(self.s, pos, self.end_pos)
|
301 | if not m:
|
302 | # A validating parser would check that > or /> is next -- there's no junk
|
303 | break
|
304 |
|
305 | yield AttrName, m.start(1), m.end(1)
|
306 |
|
307 | # Quoted is group 2, unquoted is group 3.
|
308 | if m.group(2) is not None:
|
309 | yield QuotedValue, m.start(2), m.end(2)
|
310 | elif m.group(3) is not None:
|
311 | yield UnquotedValue, m.start(3), m.end(3)
|
312 |
|
313 | # Skip past the "
|
314 | pos = m.end(0)
|
315 |
|
316 |
|
317 | def ReadUntilStartTag(it, tag_lexer, tag_name):
|
318 | """Find the next <foo>.
|
319 |
|
320 | tag_lexer is RESET.
|
321 | """
|
322 | pos = 0
|
323 | while True:
|
324 | try:
|
325 | tok_id, end_pos = next(it)
|
326 | except StopIteration:
|
327 | break
|
328 | tag_lexer.Reset(pos, end_pos)
|
329 | if tok_id == StartTag and tag_lexer.TagName() == tag_name:
|
330 | return pos, end_pos
|
331 |
|
332 | pos = end_pos
|
333 |
|
334 | raise ParseError('No start tag %r', tag_name)
|
335 |
|
336 |
|
337 | def ReadUntilEndTag(it, tag_lexer, tag_name):
|
338 | """Find the next </foo>.
|
339 |
|
340 | tag_lexer is RESET.
|
341 | """
|
342 | pos = 0
|
343 | while True:
|
344 | try:
|
345 | tok_id, end_pos = next(it)
|
346 | except StopIteration:
|
347 | break
|
348 | tag_lexer.Reset(pos, end_pos)
|
349 | if tok_id == EndTag and tag_lexer.TagName() == tag_name:
|
350 | return pos, end_pos
|
351 |
|
352 | pos = end_pos
|
353 |
|
354 | raise ParseError('No end tag %r', tag_name)
|
355 |
|
356 |
|
357 | CHAR_ENTITY = {
|
358 | 'amp': '&',
|
359 | 'lt': '<',
|
360 | 'gt': '>',
|
361 | 'quot': '"',
|
362 | }
|
363 |
|
364 |
|
365 | def ToText(s, left_pos=0, right_pos=0):
|
366 | """Given HTML, return text by unquoting > and < etc.
|
367 |
|
368 | Used by:
|
369 | doctools/oils_doc.py: PygmentsPlugin
|
370 | doctool/make_help.py: HelpIndexCards
|
371 |
|
372 | In the latter case, we cold process some tags, like:
|
373 |
|
374 | - Blue Link (not clickable, but still useful)
|
375 | - Red X
|
376 |
|
377 | That should be html.ToAnsi.
|
378 | """
|
379 | f = cStringIO.StringIO()
|
380 | out = Output(s, f, left_pos, right_pos)
|
381 |
|
382 | pos = left_pos
|
383 | for tok_id, end_pos in ValidTokens(s, left_pos, right_pos):
|
384 | if tok_id == RawData:
|
385 | out.SkipTo(pos)
|
386 | out.PrintUntil(end_pos)
|
387 |
|
388 | elif tok_id == CharEntity: # &
|
389 |
|
390 | entity = s[pos + 1:end_pos - 1]
|
391 |
|
392 | out.SkipTo(pos)
|
393 | out.Print(CHAR_ENTITY[entity])
|
394 | out.SkipTo(end_pos)
|
395 |
|
396 | # Not handling these yet
|
397 | elif tok_id == HexChar:
|
398 | raise AssertionError('Hex Char %r' % s[pos:pos + 20])
|
399 |
|
400 | elif tok_id == DecChar:
|
401 | raise AssertionError('Dec Char %r' % s[pos:pos + 20])
|
402 |
|
403 | pos = end_pos
|
404 |
|
405 | out.PrintTheRest()
|
406 | return f.getvalue()
|