| 1 |
# |
| 2 |
# ElementTree |
| 3 |
# $Id: ElementPath.py 3375 2008-02-13 08:05:08Z fredrik $ |
| 4 |
# |
| 5 |
# limited xpath support for element trees |
| 6 |
# |
| 7 |
# history: |
| 8 |
# 2003-05-23 fl created |
| 9 |
# 2003-05-28 fl added support for // etc |
| 10 |
# 2003-08-27 fl fixed parsing of periods in element names |
| 11 |
# 2007-09-10 fl new selection engine |
| 12 |
# 2007-09-12 fl fixed parent selector |
| 13 |
# 2007-09-13 fl added iterfind; changed findall to return a list |
| 14 |
# 2007-11-30 fl added namespaces support |
| 15 |
# 2009-10-30 fl added child element value filter |
| 16 |
# |
| 17 |
# Copyright (c) 2003-2009 by Fredrik Lundh. All rights reserved. |
| 18 |
# |
| 19 |
# [email protected] |
| 20 |
# http://www.pythonware.com |
| 21 |
# |
| 22 |
# -------------------------------------------------------------------- |
| 23 |
# The ElementTree toolkit is |
| 24 |
# |
| 25 |
# Copyright (c) 1999-2009 by Fredrik Lundh |
| 26 |
# |
| 27 |
# By obtaining, using, and/or copying this software and/or its |
| 28 |
# associated documentation, you agree that you have read, understood, |
| 29 |
# and will comply with the following terms and conditions: |
| 30 |
# |
| 31 |
# Permission to use, copy, modify, and distribute this software and |
| 32 |
# its associated documentation for any purpose and without fee is |
| 33 |
# hereby granted, provided that the above copyright notice appears in |
| 34 |
# all copies, and that both that copyright notice and this permission |
| 35 |
# notice appear in supporting documentation, and that the name of |
| 36 |
# Secret Labs AB or the author not be used in advertising or publicity |
| 37 |
# pertaining to distribution of the software without specific, written |
| 38 |
# prior permission. |
| 39 |
# |
| 40 |
# SECRET LABS AB AND THE AUTHOR DISCLAIMS ALL WARRANTIES WITH REGARD |
| 41 |
# TO THIS SOFTWARE, INCLUDING ALL IMPLIED WARRANTIES OF MERCHANT- |
| 42 |
# ABILITY AND FITNESS. IN NO EVENT SHALL SECRET LABS AB OR THE AUTHOR |
| 43 |
# BE LIABLE FOR ANY SPECIAL, INDIRECT OR CONSEQUENTIAL DAMAGES OR ANY |
| 44 |
# DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, |
| 45 |
# WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS |
| 46 |
# ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE |
| 47 |
# OF THIS SOFTWARE. |
| 48 |
# -------------------------------------------------------------------- |
| 49 |
|
| 50 |
# Licensed to PSF under a Contributor Agreement. |
| 51 |
# See https://www.python.org/psf/license for licensing details. |
| 52 |
|
| 53 |
## |
| 54 |
# Implementation module for XPath support. There's usually no reason |
| 55 |
# to import this module directly; the <b>ElementTree</b> does this for |
| 56 |
# you, if needed. |
| 57 |
## |
| 58 |
|
| 59 |
import re |
| 60 |
|
| 61 |
xpath_tokenizer_re = re.compile( |
| 62 |
r"(" |
| 63 |
r"'[^']*'|\"[^\"]*\"|" |
| 64 |
r"::|" |
| 65 |
r"//?|" |
| 66 |
r"\.\.|" |
| 67 |
r"\(\)|" |
| 68 |
r"!=|" |
| 69 |
r"[/.*:\[\]\(\)@=])|" |
| 70 |
r"((?:\{[^}]+\})?[^/\[\]\(\)@!=\s]+)|" |
| 71 |
r"\s+" |
| 72 |
) |
| 73 |
|
| 74 |
def xpath_tokenizer(pattern, namespaces=None): |
| 75 |
default_namespace = namespaces.get('') if namespaces else None |
| 76 |
parsing_attribute = False |
| 77 |
for token in xpath_tokenizer_re.findall(pattern): |
| 78 |
ttype, tag = token |
| 79 |
if tag and tag[0] != "{": |
| 80 |
if ":" in tag: |
| 81 |
prefix, uri = tag.split(":", 1) |
| 82 |
try: |
| 83 |
if not namespaces: |
| 84 |
raise KeyError |
| 85 |
yield ttype, "{%s}%s" % (namespaces[prefix], uri) |
| 86 |
except KeyError: |
| 87 |
raise SyntaxError("prefix %r not found in prefix map" % prefix) from None |
| 88 |
elif default_namespace and not parsing_attribute: |
| 89 |
yield ttype, "{%s}%s" % (default_namespace, tag) |
| 90 |
else: |
| 91 |
yield token |
| 92 |
parsing_attribute = False |
| 93 |
else: |
| 94 |
yield token |
| 95 |
parsing_attribute = ttype == '@' |
| 96 |
|
| 97 |
|
| 98 |
def get_parent_map(context): |
| 99 |
parent_map = context.parent_map |
| 100 |
if parent_map is None: |
| 101 |
context.parent_map = parent_map = {} |
| 102 |
for p in context.root.iter(): |
| 103 |
for e in p: |
| 104 |
parent_map[e] = p |
| 105 |
return parent_map |
| 106 |
|
| 107 |
|
| 108 |
def _is_wildcard_tag(tag): |
| 109 |
return tag[:3] == '{*}' or tag[-2:] == '}*' |
| 110 |
|
| 111 |
|
| 112 |
def _prepare_tag(tag): |
| 113 |
_isinstance, _str = isinstance, str |
| 114 |
if tag == '{*}*': |
| 115 |
# Same as '*', but no comments or processing instructions. |
| 116 |
# It can be a surprise that '*' includes those, but there is no |
| 117 |
# justification for '{*}*' doing the same. |
| 118 |
def select(context, result): |
| 119 |
for elem in result: |
| 120 |
if _isinstance(elem.tag, _str): |
| 121 |
yield elem |
| 122 |
elif tag == '{}*': |
| 123 |
# Any tag that is not in a namespace. |
| 124 |
def select(context, result): |
| 125 |
for elem in result: |
| 126 |
el_tag = elem.tag |
| 127 |
if _isinstance(el_tag, _str) and el_tag[0] != '{': |
| 128 |
yield elem |
| 129 |
elif tag[:3] == '{*}': |
| 130 |
# The tag in any (or no) namespace. |
| 131 |
suffix = tag[2:] # '}name' |
| 132 |
no_ns = slice(-len(suffix), None) |
| 133 |
tag = tag[3:] |
| 134 |
def select(context, result): |
| 135 |
for elem in result: |
| 136 |
el_tag = elem.tag |
| 137 |
if el_tag == tag or _isinstance(el_tag, _str) and el_tag[no_ns] == suffix: |
| 138 |
yield elem |
| 139 |
elif tag[-2:] == '}*': |
| 140 |
# Any tag in the given namespace. |
| 141 |
ns = tag[:-1] |
| 142 |
ns_only = slice(None, len(ns)) |
| 143 |
def select(context, result): |
| 144 |
for elem in result: |
| 145 |
el_tag = elem.tag |
| 146 |
if _isinstance(el_tag, _str) and el_tag[ns_only] == ns: |
| 147 |
yield elem |
| 148 |
else: |
| 149 |
raise RuntimeError(f"internal parser error, got {tag}") |
| 150 |
return select |
| 151 |
|
| 152 |
|
| 153 |
def prepare_child(next, token): |
| 154 |
tag = token[1] |
| 155 |
if _is_wildcard_tag(tag): |
| 156 |
select_tag = _prepare_tag(tag) |
| 157 |
def select(context, result): |
| 158 |
def select_child(result): |
| 159 |
for elem in result: |
| 160 |
yield from elem |
| 161 |
return select_tag(context, select_child(result)) |
| 162 |
else: |
| 163 |
if tag[:2] == '{}': |
| 164 |
tag = tag[2:] # '{}tag' == 'tag' |
| 165 |
def select(context, result): |
| 166 |
for elem in result: |
| 167 |
for e in elem: |
| 168 |
if e.tag == tag: |
| 169 |
yield e |
| 170 |
return select |
| 171 |
|
| 172 |
def prepare_star(next, token): |
| 173 |
def select(context, result): |
| 174 |
for elem in result: |
| 175 |
yield from elem |
| 176 |
return select |
| 177 |
|
| 178 |
def prepare_self(next, token): |
| 179 |
def select(context, result): |
| 180 |
yield from result |
| 181 |
return select |
| 182 |
|
| 183 |
def prepare_descendant(next, token): |
| 184 |
try: |
| 185 |
token = next() |
| 186 |
except StopIteration: |
| 187 |
return |
| 188 |
if token[0] == "*": |
| 189 |
tag = "*" |
| 190 |
elif not token[0]: |
| 191 |
tag = token[1] |
| 192 |
else: |
| 193 |
raise SyntaxError("invalid descendant") |
| 194 |
|
| 195 |
if _is_wildcard_tag(tag): |
| 196 |
select_tag = _prepare_tag(tag) |
| 197 |
def select(context, result): |
| 198 |
def select_child(result): |
| 199 |
for elem in result: |
| 200 |
for e in elem.iter(): |
| 201 |
if e is not elem: |
| 202 |
yield e |
| 203 |
return select_tag(context, select_child(result)) |
| 204 |
else: |
| 205 |
if tag[:2] == '{}': |
| 206 |
tag = tag[2:] # '{}tag' == 'tag' |
| 207 |
def select(context, result): |
| 208 |
for elem in result: |
| 209 |
for e in elem.iter(tag): |
| 210 |
if e is not elem: |
| 211 |
yield e |
| 212 |
return select |
| 213 |
|
| 214 |
def prepare_parent(next, token): |
| 215 |
def select(context, result): |
| 216 |
# FIXME: raise error if .. is applied at toplevel? |
| 217 |
parent_map = get_parent_map(context) |
| 218 |
result_map = {} |
| 219 |
for elem in result: |
| 220 |
if elem in parent_map: |
| 221 |
parent = parent_map[elem] |
| 222 |
if parent not in result_map: |
| 223 |
result_map[parent] = None |
| 224 |
yield parent |
| 225 |
return select |
| 226 |
|
| 227 |
def prepare_predicate(next, token): |
| 228 |
# FIXME: replace with real parser!!! refs: |
| 229 |
# http://javascript.crockford.com/tdop/tdop.html |
| 230 |
signature = [] |
| 231 |
predicate = [] |
| 232 |
while 1: |
| 233 |
try: |
| 234 |
token = next() |
| 235 |
except StopIteration: |
| 236 |
return |
| 237 |
if token[0] == "]": |
| 238 |
break |
| 239 |
if token == ('', ''): |
| 240 |
# ignore whitespace |
| 241 |
continue |
| 242 |
if token[0] and token[0][:1] in "'\"": |
| 243 |
token = "'", token[0][1:-1] |
| 244 |
signature.append(token[0] or "-") |
| 245 |
predicate.append(token[1]) |
| 246 |
signature = "".join(signature) |
| 247 |
# use signature to determine predicate type |
| 248 |
if signature == "@-": |
| 249 |
# [@attribute] predicate |
| 250 |
key = predicate[1] |
| 251 |
def select(context, result): |
| 252 |
for elem in result: |
| 253 |
if elem.get(key) is not None: |
| 254 |
yield elem |
| 255 |
return select |
| 256 |
if signature == "@-='" or signature == "@-!='": |
| 257 |
# [@attribute='value'] or [@attribute!='value'] |
| 258 |
key = predicate[1] |
| 259 |
value = predicate[-1] |
| 260 |
def select(context, result): |
| 261 |
for elem in result: |
| 262 |
if elem.get(key) == value: |
| 263 |
yield elem |
| 264 |
def select_negated(context, result): |
| 265 |
for elem in result: |
| 266 |
if (attr_value := elem.get(key)) is not None and attr_value != value: |
| 267 |
yield elem |
| 268 |
return select_negated if '!=' in signature else select |
| 269 |
if signature == "-" and not re.match(r"\-?\d+$", predicate[0]): |
| 270 |
# [tag] |
| 271 |
tag = predicate[0] |
| 272 |
def select(context, result): |
| 273 |
for elem in result: |
| 274 |
if elem.find(tag) is not None: |
| 275 |
yield elem |
| 276 |
return select |
| 277 |
if signature == ".='" or signature == ".!='" or ( |
| 278 |
(signature == "-='" or signature == "-!='") |
| 279 |
and not re.match(r"\-?\d+$", predicate[0])): |
| 280 |
# [.='value'] or [tag='value'] or [.!='value'] or [tag!='value'] |
| 281 |
tag = predicate[0] |
| 282 |
value = predicate[-1] |
| 283 |
if tag: |
| 284 |
def select(context, result): |
| 285 |
for elem in result: |
| 286 |
for e in elem.findall(tag): |
| 287 |
if "".join(e.itertext()) == value: |
| 288 |
yield elem |
| 289 |
break |
| 290 |
def select_negated(context, result): |
| 291 |
for elem in result: |
| 292 |
for e in elem.iterfind(tag): |
| 293 |
if "".join(e.itertext()) != value: |
| 294 |
yield elem |
| 295 |
break |
| 296 |
else: |
| 297 |
def select(context, result): |
| 298 |
for elem in result: |
| 299 |
if "".join(elem.itertext()) == value: |
| 300 |
yield elem |
| 301 |
def select_negated(context, result): |
| 302 |
for elem in result: |
| 303 |
if "".join(elem.itertext()) != value: |
| 304 |
yield elem |
| 305 |
return select_negated if '!=' in signature else select |
| 306 |
if signature == "-" or signature == "-()" or signature == "-()-": |
| 307 |
# [index] or [last()] or [last()-index] |
| 308 |
if signature == "-": |
| 309 |
# [index] |
| 310 |
index = int(predicate[0]) - 1 |
| 311 |
if index < 0: |
| 312 |
raise SyntaxError("XPath position >= 1 expected") |
| 313 |
else: |
| 314 |
if predicate[0] != "last": |
| 315 |
raise SyntaxError("unsupported function") |
| 316 |
if signature == "-()-": |
| 317 |
try: |
| 318 |
index = int(predicate[2]) - 1 |
| 319 |
except ValueError: |
| 320 |
raise SyntaxError("unsupported expression") |
| 321 |
if index > -2: |
| 322 |
raise SyntaxError("XPath offset from last() must be negative") |
| 323 |
else: |
| 324 |
index = -1 |
| 325 |
def select(context, result): |
| 326 |
parent_map = get_parent_map(context) |
| 327 |
cache = {} |
| 328 |
for elem in result: |
| 329 |
try: |
| 330 |
parent = parent_map[elem] |
| 331 |
except KeyError: |
| 332 |
continue |
| 333 |
key = (parent, elem.tag) |
| 334 |
if key not in cache: |
| 335 |
# FIXME: what if the selector is "*" ? |
| 336 |
elems = parent.findall(elem.tag) |
| 337 |
try: |
| 338 |
cache[key] = elems[index] |
| 339 |
except IndexError: |
| 340 |
cache[key] = None |
| 341 |
if cache[key] is elem: |
| 342 |
yield elem |
| 343 |
return select |
| 344 |
raise SyntaxError("invalid predicate") |
| 345 |
|
| 346 |
ops = { |
| 347 |
"": prepare_child, |
| 348 |
"*": prepare_star, |
| 349 |
".": prepare_self, |
| 350 |
"..": prepare_parent, |
| 351 |
"//": prepare_descendant, |
| 352 |
"[": prepare_predicate, |
| 353 |
} |
| 354 |
|
| 355 |
_cache = {} |
| 356 |
|
| 357 |
class _SelectorContext: |
| 358 |
parent_map = None |
| 359 |
def __init__(self, root): |
| 360 |
self.root = root |
| 361 |
|
| 362 |
# -------------------------------------------------------------------- |
| 363 |
|
| 364 |
## |
| 365 |
# Generate all matching objects. |
| 366 |
|
| 367 |
def iterfind(elem, path, namespaces=None): |
| 368 |
# compile selector pattern |
| 369 |
if path[-1:] == "/": |
| 370 |
path = path + "*" # implicit all (FIXME: keep this?) |
| 371 |
|
| 372 |
cache_key = (path,) |
| 373 |
if namespaces: |
| 374 |
cache_key += tuple(sorted(namespaces.items())) |
| 375 |
|
| 376 |
try: |
| 377 |
selector = _cache[cache_key] |
| 378 |
except KeyError: |
| 379 |
if len(_cache) > 100: |
| 380 |
_cache.clear() |
| 381 |
if path[:1] == "/": |
| 382 |
raise SyntaxError("cannot use absolute path on element") |
| 383 |
next = iter(xpath_tokenizer(path, namespaces)).__next__ |
| 384 |
try: |
| 385 |
token = next() |
| 386 |
except StopIteration: |
| 387 |
return |
| 388 |
selector = [] |
| 389 |
while 1: |
| 390 |
try: |
| 391 |
selector.append(ops[token[0]](next, token)) |
| 392 |
except StopIteration: |
| 393 |
raise SyntaxError("invalid path") from None |
| 394 |
try: |
| 395 |
token = next() |
| 396 |
if token[0] == "/": |
| 397 |
token = next() |
| 398 |
except StopIteration: |
| 399 |
break |
| 400 |
_cache[cache_key] = selector |
| 401 |
# execute selector pattern |
| 402 |
result = [elem] |
| 403 |
context = _SelectorContext(elem) |
| 404 |
for select in selector: |
| 405 |
result = select(context, result) |
| 406 |
return result |
| 407 |
|
| 408 |
## |
| 409 |
# Find first matching object. |
| 410 |
|
| 411 |
def find(elem, path, namespaces=None): |
| 412 |
return next(iterfind(elem, path, namespaces), None) |
| 413 |
|
| 414 |
## |
| 415 |
# Find all matching objects. |
| 416 |
|
| 417 |
def findall(elem, path, namespaces=None): |
| 418 |
return list(iterfind(elem, path, namespaces)) |
| 419 |
|
| 420 |
## |
| 421 |
# Find text for first matching object. |
| 422 |
|
| 423 |
def findtext(elem, path, default=None, namespaces=None): |
| 424 |
try: |
| 425 |
elem = next(iterfind(elem, path, namespaces)) |
| 426 |
if elem.text is None: |
| 427 |
return "" |
| 428 |
return elem.text |
| 429 |
except StopIteration: |
| 430 |
return default |