| 1 |
"""Facility to use the Expat parser to load a minidom instance |
| 2 |
from a string or file. |
| 3 |
|
| 4 |
This avoids all the overhead of SAX and pulldom to gain performance. |
| 5 |
""" |
| 6 |
|
| 7 |
# Warning! |
| 8 |
# |
| 9 |
# This module is tightly bound to the implementation details of the |
| 10 |
# minidom DOM and can't be used with other DOM implementations. This |
| 11 |
# is due, in part, to a lack of appropriate methods in the DOM (there is |
| 12 |
# no way to create Entity and Notation nodes via the DOM Level 2 |
| 13 |
# interface), and for performance. The latter is the cause of some fairly |
| 14 |
# cryptic code. |
| 15 |
# |
| 16 |
# Performance hacks: |
| 17 |
# |
| 18 |
# - .character_data_handler() has an extra case in which continuing |
| 19 |
# data is appended to an existing Text node; this can be a |
| 20 |
# speedup since pyexpat can break up character data into multiple |
| 21 |
# callbacks even though we set the buffer_text attribute on the |
| 22 |
# parser. This also gives us the advantage that we don't need a |
| 23 |
# separate normalization pass. |
| 24 |
# |
| 25 |
# - Determining that a node exists is done using an identity comparison |
| 26 |
# with None rather than a truth test; this avoids searching for and |
| 27 |
# calling any methods on the node object if it exists. (A rather |
| 28 |
# nice speedup is achieved this way as well!) |
| 29 |
|
| 30 |
from xml.dom import xmlbuilder, minidom, Node |
| 31 |
from xml.dom import EMPTY_NAMESPACE, EMPTY_PREFIX, XMLNS_NAMESPACE |
| 32 |
from xml.parsers import expat |
| 33 |
from xml.dom.minidom import _append_child, _set_attribute_node |
| 34 |
from xml.dom.NodeFilter import NodeFilter |
| 35 |
|
| 36 |
TEXT_NODE = Node.TEXT_NODE |
| 37 |
CDATA_SECTION_NODE = Node.CDATA_SECTION_NODE |
| 38 |
DOCUMENT_NODE = Node.DOCUMENT_NODE |
| 39 |
|
| 40 |
FILTER_ACCEPT = xmlbuilder.DOMBuilderFilter.FILTER_ACCEPT |
| 41 |
FILTER_REJECT = xmlbuilder.DOMBuilderFilter.FILTER_REJECT |
| 42 |
FILTER_SKIP = xmlbuilder.DOMBuilderFilter.FILTER_SKIP |
| 43 |
FILTER_INTERRUPT = xmlbuilder.DOMBuilderFilter.FILTER_INTERRUPT |
| 44 |
|
| 45 |
theDOMImplementation = minidom.getDOMImplementation() |
| 46 |
|
| 47 |
# Expat typename -> TypeInfo |
| 48 |
_typeinfo_map = { |
| 49 |
"CDATA": minidom.TypeInfo(None, "cdata"), |
| 50 |
"ENUM": minidom.TypeInfo(None, "enumeration"), |
| 51 |
"ENTITY": minidom.TypeInfo(None, "entity"), |
| 52 |
"ENTITIES": minidom.TypeInfo(None, "entities"), |
| 53 |
"ID": minidom.TypeInfo(None, "id"), |
| 54 |
"IDREF": minidom.TypeInfo(None, "idref"), |
| 55 |
"IDREFS": minidom.TypeInfo(None, "idrefs"), |
| 56 |
"NMTOKEN": minidom.TypeInfo(None, "nmtoken"), |
| 57 |
"NMTOKENS": minidom.TypeInfo(None, "nmtokens"), |
| 58 |
} |
| 59 |
|
| 60 |
class ElementInfo(object): |
| 61 |
__slots__ = '_attr_info', '_model', 'tagName' |
| 62 |
|
| 63 |
def __init__(self, tagName, model=None): |
| 64 |
self.tagName = tagName |
| 65 |
self._attr_info = [] |
| 66 |
self._model = model |
| 67 |
|
| 68 |
def __getstate__(self): |
| 69 |
return self._attr_info, self._model, self.tagName |
| 70 |
|
| 71 |
def __setstate__(self, state): |
| 72 |
self._attr_info, self._model, self.tagName = state |
| 73 |
|
| 74 |
def getAttributeType(self, aname): |
| 75 |
for info in self._attr_info: |
| 76 |
if info[1] == aname: |
| 77 |
t = info[-2] |
| 78 |
if t[0] == "(": |
| 79 |
return _typeinfo_map["ENUM"] |
| 80 |
else: |
| 81 |
return _typeinfo_map[info[-2]] |
| 82 |
return minidom._no_type |
| 83 |
|
| 84 |
def getAttributeTypeNS(self, namespaceURI, localName): |
| 85 |
return minidom._no_type |
| 86 |
|
| 87 |
def isElementContent(self): |
| 88 |
if self._model: |
| 89 |
type = self._model[0] |
| 90 |
return type not in (expat.model.XML_CTYPE_ANY, |
| 91 |
expat.model.XML_CTYPE_MIXED) |
| 92 |
else: |
| 93 |
return False |
| 94 |
|
| 95 |
def isEmpty(self): |
| 96 |
if self._model: |
| 97 |
return self._model[0] == expat.model.XML_CTYPE_EMPTY |
| 98 |
else: |
| 99 |
return False |
| 100 |
|
| 101 |
def isId(self, aname): |
| 102 |
for info in self._attr_info: |
| 103 |
if info[1] == aname: |
| 104 |
return info[-2] == "ID" |
| 105 |
return False |
| 106 |
|
| 107 |
def isIdNS(self, euri, ename, auri, aname): |
| 108 |
# not sure this is meaningful |
| 109 |
return self.isId((auri, aname)) |
| 110 |
|
| 111 |
def _intern(builder, s): |
| 112 |
return builder._intern_setdefault(s, s) |
| 113 |
|
| 114 |
def _parse_ns_name(builder, name): |
| 115 |
assert ' ' in name |
| 116 |
parts = name.split(' ') |
| 117 |
intern = builder._intern_setdefault |
| 118 |
if len(parts) == 3: |
| 119 |
uri, localname, prefix = parts |
| 120 |
prefix = intern(prefix, prefix) |
| 121 |
qname = "%s:%s" % (prefix, localname) |
| 122 |
qname = intern(qname, qname) |
| 123 |
localname = intern(localname, localname) |
| 124 |
elif len(parts) == 2: |
| 125 |
uri, localname = parts |
| 126 |
prefix = EMPTY_PREFIX |
| 127 |
qname = localname = intern(localname, localname) |
| 128 |
else: |
| 129 |
raise ValueError("Unsupported syntax: spaces in URIs not supported: %r" % name) |
| 130 |
return intern(uri, uri), localname, prefix, qname |
| 131 |
|
| 132 |
|
| 133 |
class ExpatBuilder: |
| 134 |
"""Document builder that uses Expat to build a ParsedXML.DOM document |
| 135 |
instance.""" |
| 136 |
|
| 137 |
def __init__(self, options=None): |
| 138 |
if options is None: |
| 139 |
options = xmlbuilder.Options() |
| 140 |
self._options = options |
| 141 |
if self._options.filter is not None: |
| 142 |
self._filter = FilterVisibilityController(self._options.filter) |
| 143 |
else: |
| 144 |
self._filter = None |
| 145 |
# This *really* doesn't do anything in this case, so |
| 146 |
# override it with something fast & minimal. |
| 147 |
self._finish_start_element = id |
| 148 |
self._parser = None |
| 149 |
self.reset() |
| 150 |
|
| 151 |
def createParser(self): |
| 152 |
"""Create a new parser object.""" |
| 153 |
return expat.ParserCreate() |
| 154 |
|
| 155 |
def getParser(self): |
| 156 |
"""Return the parser object, creating a new one if needed.""" |
| 157 |
if not self._parser: |
| 158 |
self._parser = self.createParser() |
| 159 |
self._intern_setdefault = self._parser.intern.setdefault |
| 160 |
self._parser.buffer_text = True |
| 161 |
self._parser.ordered_attributes = True |
| 162 |
self._parser.specified_attributes = True |
| 163 |
self.install(self._parser) |
| 164 |
return self._parser |
| 165 |
|
| 166 |
def reset(self): |
| 167 |
"""Free all data structures used during DOM construction.""" |
| 168 |
self.document = theDOMImplementation.createDocument( |
| 169 |
EMPTY_NAMESPACE, None, None) |
| 170 |
self.curNode = self.document |
| 171 |
self._elem_info = self.document._elem_info |
| 172 |
self._cdata = False |
| 173 |
|
| 174 |
def install(self, parser): |
| 175 |
"""Install the callbacks needed to build the DOM into the parser.""" |
| 176 |
# This creates circular references! |
| 177 |
parser.StartDoctypeDeclHandler = self.start_doctype_decl_handler |
| 178 |
parser.StartElementHandler = self.first_element_handler |
| 179 |
parser.EndElementHandler = self.end_element_handler |
| 180 |
parser.ProcessingInstructionHandler = self.pi_handler |
| 181 |
if self._options.entities: |
| 182 |
parser.EntityDeclHandler = self.entity_decl_handler |
| 183 |
parser.NotationDeclHandler = self.notation_decl_handler |
| 184 |
if self._options.comments: |
| 185 |
parser.CommentHandler = self.comment_handler |
| 186 |
if self._options.cdata_sections: |
| 187 |
parser.StartCdataSectionHandler = self.start_cdata_section_handler |
| 188 |
parser.EndCdataSectionHandler = self.end_cdata_section_handler |
| 189 |
parser.CharacterDataHandler = self.character_data_handler_cdata |
| 190 |
else: |
| 191 |
parser.CharacterDataHandler = self.character_data_handler |
| 192 |
parser.ExternalEntityRefHandler = self.external_entity_ref_handler |
| 193 |
parser.XmlDeclHandler = self.xml_decl_handler |
| 194 |
parser.ElementDeclHandler = self.element_decl_handler |
| 195 |
parser.AttlistDeclHandler = self.attlist_decl_handler |
| 196 |
|
| 197 |
def parseFile(self, file): |
| 198 |
"""Parse a document from a file object, returning the document |
| 199 |
node.""" |
| 200 |
parser = self.getParser() |
| 201 |
first_buffer = True |
| 202 |
try: |
| 203 |
while buffer := file.read(16*1024): |
| 204 |
parser.Parse(buffer, False) |
| 205 |
if first_buffer and self.document.documentElement: |
| 206 |
self._setup_subset(buffer) |
| 207 |
first_buffer = False |
| 208 |
parser.Parse(b"", True) |
| 209 |
except ParseEscape: |
| 210 |
pass |
| 211 |
doc = self.document |
| 212 |
self.reset() |
| 213 |
self._parser = None |
| 214 |
return doc |
| 215 |
|
| 216 |
def parseString(self, string): |
| 217 |
"""Parse a document from a string, returning the document node.""" |
| 218 |
parser = self.getParser() |
| 219 |
try: |
| 220 |
parser.Parse(string, True) |
| 221 |
self._setup_subset(string) |
| 222 |
except ParseEscape: |
| 223 |
pass |
| 224 |
doc = self.document |
| 225 |
self.reset() |
| 226 |
self._parser = None |
| 227 |
return doc |
| 228 |
|
| 229 |
def _setup_subset(self, buffer): |
| 230 |
"""Load the internal subset if there might be one.""" |
| 231 |
if self.document.doctype: |
| 232 |
extractor = InternalSubsetExtractor() |
| 233 |
extractor.parseString(buffer) |
| 234 |
subset = extractor.getSubset() |
| 235 |
self.document.doctype.internalSubset = subset |
| 236 |
|
| 237 |
def start_doctype_decl_handler(self, doctypeName, systemId, publicId, |
| 238 |
has_internal_subset): |
| 239 |
doctype = self.document.implementation.createDocumentType( |
| 240 |
doctypeName, publicId, systemId) |
| 241 |
doctype.ownerDocument = self.document |
| 242 |
_append_child(self.document, doctype) |
| 243 |
self.document.doctype = doctype |
| 244 |
if self._filter and self._filter.acceptNode(doctype) == FILTER_REJECT: |
| 245 |
self.document.doctype = None |
| 246 |
del self.document.childNodes[-1] |
| 247 |
doctype = None |
| 248 |
self._parser.EntityDeclHandler = None |
| 249 |
self._parser.NotationDeclHandler = None |
| 250 |
if has_internal_subset: |
| 251 |
if doctype is not None: |
| 252 |
doctype.entities._seq = [] |
| 253 |
doctype.notations._seq = [] |
| 254 |
self._parser.CommentHandler = None |
| 255 |
self._parser.ProcessingInstructionHandler = None |
| 256 |
self._parser.EndDoctypeDeclHandler = self.end_doctype_decl_handler |
| 257 |
|
| 258 |
def end_doctype_decl_handler(self): |
| 259 |
if self._options.comments: |
| 260 |
self._parser.CommentHandler = self.comment_handler |
| 261 |
self._parser.ProcessingInstructionHandler = self.pi_handler |
| 262 |
if not (self._elem_info or self._filter): |
| 263 |
self._finish_end_element = id |
| 264 |
|
| 265 |
def pi_handler(self, target, data): |
| 266 |
node = self.document.createProcessingInstruction(target, data) |
| 267 |
_append_child(self.curNode, node) |
| 268 |
if self._filter and self._filter.acceptNode(node) == FILTER_REJECT: |
| 269 |
self.curNode.removeChild(node) |
| 270 |
|
| 271 |
def character_data_handler_cdata(self, data): |
| 272 |
childNodes = self.curNode.childNodes |
| 273 |
if self._cdata: |
| 274 |
if ( self._cdata_continue |
| 275 |
and childNodes[-1].nodeType == CDATA_SECTION_NODE): |
| 276 |
childNodes[-1].appendData(data) |
| 277 |
return |
| 278 |
node = self.document.createCDATASection(data) |
| 279 |
self._cdata_continue = True |
| 280 |
elif childNodes and childNodes[-1].nodeType == TEXT_NODE: |
| 281 |
node = childNodes[-1] |
| 282 |
value = node.data + data |
| 283 |
node.data = value |
| 284 |
return |
| 285 |
else: |
| 286 |
node = minidom.Text() |
| 287 |
node.data = data |
| 288 |
node.ownerDocument = self.document |
| 289 |
_append_child(self.curNode, node) |
| 290 |
|
| 291 |
def character_data_handler(self, data): |
| 292 |
childNodes = self.curNode.childNodes |
| 293 |
if childNodes and childNodes[-1].nodeType == TEXT_NODE: |
| 294 |
node = childNodes[-1] |
| 295 |
node.data = node.data + data |
| 296 |
return |
| 297 |
node = minidom.Text() |
| 298 |
node.data = node.data + data |
| 299 |
node.ownerDocument = self.document |
| 300 |
_append_child(self.curNode, node) |
| 301 |
|
| 302 |
def entity_decl_handler(self, entityName, is_parameter_entity, value, |
| 303 |
base, systemId, publicId, notationName): |
| 304 |
if is_parameter_entity: |
| 305 |
# we don't care about parameter entities for the DOM |
| 306 |
return |
| 307 |
if not self._options.entities: |
| 308 |
return |
| 309 |
node = self.document._create_entity(entityName, publicId, |
| 310 |
systemId, notationName) |
| 311 |
if value is not None: |
| 312 |
# internal entity |
| 313 |
# node *should* be readonly, but we'll cheat |
| 314 |
child = self.document.createTextNode(value) |
| 315 |
node.childNodes.append(child) |
| 316 |
self.document.doctype.entities._seq.append(node) |
| 317 |
if self._filter and self._filter.acceptNode(node) == FILTER_REJECT: |
| 318 |
del self.document.doctype.entities._seq[-1] |
| 319 |
|
| 320 |
def notation_decl_handler(self, notationName, base, systemId, publicId): |
| 321 |
node = self.document._create_notation(notationName, publicId, systemId) |
| 322 |
self.document.doctype.notations._seq.append(node) |
| 323 |
if self._filter and self._filter.acceptNode(node) == FILTER_ACCEPT: |
| 324 |
del self.document.doctype.notations._seq[-1] |
| 325 |
|
| 326 |
def comment_handler(self, data): |
| 327 |
node = self.document.createComment(data) |
| 328 |
_append_child(self.curNode, node) |
| 329 |
if self._filter and self._filter.acceptNode(node) == FILTER_REJECT: |
| 330 |
self.curNode.removeChild(node) |
| 331 |
|
| 332 |
def start_cdata_section_handler(self): |
| 333 |
self._cdata = True |
| 334 |
self._cdata_continue = False |
| 335 |
|
| 336 |
def end_cdata_section_handler(self): |
| 337 |
self._cdata = False |
| 338 |
self._cdata_continue = False |
| 339 |
|
| 340 |
def external_entity_ref_handler(self, context, base, systemId, publicId): |
| 341 |
return 1 |
| 342 |
|
| 343 |
def first_element_handler(self, name, attributes): |
| 344 |
if self._filter is None and not self._elem_info: |
| 345 |
self._finish_end_element = id |
| 346 |
self.getParser().StartElementHandler = self.start_element_handler |
| 347 |
self.start_element_handler(name, attributes) |
| 348 |
|
| 349 |
def start_element_handler(self, name, attributes): |
| 350 |
node = self.document.createElement(name) |
| 351 |
_append_child(self.curNode, node) |
| 352 |
self.curNode = node |
| 353 |
|
| 354 |
if attributes: |
| 355 |
for i in range(0, len(attributes), 2): |
| 356 |
a = minidom.Attr(attributes[i], EMPTY_NAMESPACE, |
| 357 |
None, EMPTY_PREFIX) |
| 358 |
value = attributes[i+1] |
| 359 |
a.value = value |
| 360 |
a.ownerDocument = self.document |
| 361 |
_set_attribute_node(node, a) |
| 362 |
|
| 363 |
if node is not self.document.documentElement: |
| 364 |
self._finish_start_element(node) |
| 365 |
|
| 366 |
def _finish_start_element(self, node): |
| 367 |
if self._filter: |
| 368 |
# To be general, we'd have to call isSameNode(), but this |
| 369 |
# is sufficient for minidom: |
| 370 |
if node is self.document.documentElement: |
| 371 |
return |
| 372 |
filt = self._filter.startContainer(node) |
| 373 |
if filt == FILTER_REJECT: |
| 374 |
# ignore this node & all descendents |
| 375 |
Rejecter(self) |
| 376 |
elif filt == FILTER_SKIP: |
| 377 |
# ignore this node, but make it's children become |
| 378 |
# children of the parent node |
| 379 |
Skipper(self) |
| 380 |
else: |
| 381 |
return |
| 382 |
self.curNode = node.parentNode |
| 383 |
node.parentNode.removeChild(node) |
| 384 |
node.unlink() |
| 385 |
|
| 386 |
# If this ever changes, Namespaces.end_element_handler() needs to |
| 387 |
# be changed to match. |
| 388 |
# |
| 389 |
def end_element_handler(self, name): |
| 390 |
curNode = self.curNode |
| 391 |
self.curNode = curNode.parentNode |
| 392 |
self._finish_end_element(curNode) |
| 393 |
|
| 394 |
def _finish_end_element(self, curNode): |
| 395 |
info = self._elem_info.get(curNode.tagName) |
| 396 |
if info: |
| 397 |
self._handle_white_text_nodes(curNode, info) |
| 398 |
if self._filter: |
| 399 |
if curNode is self.document.documentElement: |
| 400 |
return |
| 401 |
if self._filter.acceptNode(curNode) == FILTER_REJECT: |
| 402 |
self.curNode.removeChild(curNode) |
| 403 |
curNode.unlink() |
| 404 |
|
| 405 |
def _handle_white_text_nodes(self, node, info): |
| 406 |
if (self._options.whitespace_in_element_content |
| 407 |
or not info.isElementContent()): |
| 408 |
return |
| 409 |
|
| 410 |
# We have element type information and should remove ignorable |
| 411 |
# whitespace; identify for text nodes which contain only |
| 412 |
# whitespace. |
| 413 |
L = [] |
| 414 |
for child in node.childNodes: |
| 415 |
if child.nodeType == TEXT_NODE and not child.data.strip(): |
| 416 |
L.append(child) |
| 417 |
|
| 418 |
# Remove ignorable whitespace from the tree. |
| 419 |
for child in L: |
| 420 |
node.removeChild(child) |
| 421 |
|
| 422 |
def element_decl_handler(self, name, model): |
| 423 |
info = self._elem_info.get(name) |
| 424 |
if info is None: |
| 425 |
self._elem_info[name] = ElementInfo(name, model) |
| 426 |
else: |
| 427 |
assert info._model is None |
| 428 |
info._model = model |
| 429 |
|
| 430 |
def attlist_decl_handler(self, elem, name, type, default, required): |
| 431 |
info = self._elem_info.get(elem) |
| 432 |
if info is None: |
| 433 |
info = ElementInfo(elem) |
| 434 |
self._elem_info[elem] = info |
| 435 |
info._attr_info.append( |
| 436 |
[None, name, None, None, default, 0, type, required]) |
| 437 |
|
| 438 |
def xml_decl_handler(self, version, encoding, standalone): |
| 439 |
self.document.version = version |
| 440 |
self.document.encoding = encoding |
| 441 |
# This is still a little ugly, thanks to the pyexpat API. ;-( |
| 442 |
if standalone >= 0: |
| 443 |
if standalone: |
| 444 |
self.document.standalone = True |
| 445 |
else: |
| 446 |
self.document.standalone = False |
| 447 |
|
| 448 |
|
| 449 |
# Don't include FILTER_INTERRUPT, since that's checked separately |
| 450 |
# where allowed. |
| 451 |
_ALLOWED_FILTER_RETURNS = (FILTER_ACCEPT, FILTER_REJECT, FILTER_SKIP) |
| 452 |
|
| 453 |
class FilterVisibilityController(object): |
| 454 |
"""Wrapper around a DOMBuilderFilter which implements the checks |
| 455 |
to make the whatToShow filter attribute work.""" |
| 456 |
|
| 457 |
__slots__ = 'filter', |
| 458 |
|
| 459 |
def __init__(self, filter): |
| 460 |
self.filter = filter |
| 461 |
|
| 462 |
def startContainer(self, node): |
| 463 |
mask = self._nodetype_mask[node.nodeType] |
| 464 |
if self.filter.whatToShow & mask: |
| 465 |
val = self.filter.startContainer(node) |
| 466 |
if val == FILTER_INTERRUPT: |
| 467 |
raise ParseEscape |
| 468 |
if val not in _ALLOWED_FILTER_RETURNS: |
| 469 |
raise ValueError( |
| 470 |
"startContainer() returned illegal value: " + repr(val)) |
| 471 |
return val |
| 472 |
else: |
| 473 |
return FILTER_ACCEPT |
| 474 |
|
| 475 |
def acceptNode(self, node): |
| 476 |
mask = self._nodetype_mask[node.nodeType] |
| 477 |
if self.filter.whatToShow & mask: |
| 478 |
val = self.filter.acceptNode(node) |
| 479 |
if val == FILTER_INTERRUPT: |
| 480 |
raise ParseEscape |
| 481 |
if val == FILTER_SKIP: |
| 482 |
# move all child nodes to the parent, and remove this node |
| 483 |
parent = node.parentNode |
| 484 |
for child in node.childNodes[:]: |
| 485 |
parent.appendChild(child) |
| 486 |
# node is handled by the caller |
| 487 |
return FILTER_REJECT |
| 488 |
if val not in _ALLOWED_FILTER_RETURNS: |
| 489 |
raise ValueError( |
| 490 |
"acceptNode() returned illegal value: " + repr(val)) |
| 491 |
return val |
| 492 |
else: |
| 493 |
return FILTER_ACCEPT |
| 494 |
|
| 495 |
_nodetype_mask = { |
| 496 |
Node.ELEMENT_NODE: NodeFilter.SHOW_ELEMENT, |
| 497 |
Node.ATTRIBUTE_NODE: NodeFilter.SHOW_ATTRIBUTE, |
| 498 |
Node.TEXT_NODE: NodeFilter.SHOW_TEXT, |
| 499 |
Node.CDATA_SECTION_NODE: NodeFilter.SHOW_CDATA_SECTION, |
| 500 |
Node.ENTITY_REFERENCE_NODE: NodeFilter.SHOW_ENTITY_REFERENCE, |
| 501 |
Node.ENTITY_NODE: NodeFilter.SHOW_ENTITY, |
| 502 |
Node.PROCESSING_INSTRUCTION_NODE: NodeFilter.SHOW_PROCESSING_INSTRUCTION, |
| 503 |
Node.COMMENT_NODE: NodeFilter.SHOW_COMMENT, |
| 504 |
Node.DOCUMENT_NODE: NodeFilter.SHOW_DOCUMENT, |
| 505 |
Node.DOCUMENT_TYPE_NODE: NodeFilter.SHOW_DOCUMENT_TYPE, |
| 506 |
Node.DOCUMENT_FRAGMENT_NODE: NodeFilter.SHOW_DOCUMENT_FRAGMENT, |
| 507 |
Node.NOTATION_NODE: NodeFilter.SHOW_NOTATION, |
| 508 |
} |
| 509 |
|
| 510 |
|
| 511 |
class FilterCrutch(object): |
| 512 |
__slots__ = '_builder', '_level', '_old_start', '_old_end' |
| 513 |
|
| 514 |
def __init__(self, builder): |
| 515 |
self._level = 0 |
| 516 |
self._builder = builder |
| 517 |
parser = builder._parser |
| 518 |
self._old_start = parser.StartElementHandler |
| 519 |
self._old_end = parser.EndElementHandler |
| 520 |
parser.StartElementHandler = self.start_element_handler |
| 521 |
parser.EndElementHandler = self.end_element_handler |
| 522 |
|
| 523 |
class Rejecter(FilterCrutch): |
| 524 |
__slots__ = () |
| 525 |
|
| 526 |
def __init__(self, builder): |
| 527 |
FilterCrutch.__init__(self, builder) |
| 528 |
parser = builder._parser |
| 529 |
for name in ("ProcessingInstructionHandler", |
| 530 |
"CommentHandler", |
| 531 |
"CharacterDataHandler", |
| 532 |
"StartCdataSectionHandler", |
| 533 |
"EndCdataSectionHandler", |
| 534 |
"ExternalEntityRefHandler", |
| 535 |
): |
| 536 |
setattr(parser, name, None) |
| 537 |
|
| 538 |
def start_element_handler(self, *args): |
| 539 |
self._level = self._level + 1 |
| 540 |
|
| 541 |
def end_element_handler(self, *args): |
| 542 |
if self._level == 0: |
| 543 |
# restore the old handlers |
| 544 |
parser = self._builder._parser |
| 545 |
self._builder.install(parser) |
| 546 |
parser.StartElementHandler = self._old_start |
| 547 |
parser.EndElementHandler = self._old_end |
| 548 |
else: |
| 549 |
self._level = self._level - 1 |
| 550 |
|
| 551 |
class Skipper(FilterCrutch): |
| 552 |
__slots__ = () |
| 553 |
|
| 554 |
def start_element_handler(self, *args): |
| 555 |
node = self._builder.curNode |
| 556 |
self._old_start(*args) |
| 557 |
if self._builder.curNode is not node: |
| 558 |
self._level = self._level + 1 |
| 559 |
|
| 560 |
def end_element_handler(self, *args): |
| 561 |
if self._level == 0: |
| 562 |
# We're popping back out of the node we're skipping, so we |
| 563 |
# shouldn't need to do anything but reset the handlers. |
| 564 |
self._builder._parser.StartElementHandler = self._old_start |
| 565 |
self._builder._parser.EndElementHandler = self._old_end |
| 566 |
self._builder = None |
| 567 |
else: |
| 568 |
self._level = self._level - 1 |
| 569 |
self._old_end(*args) |
| 570 |
|
| 571 |
|
| 572 |
# framework document used by the fragment builder. |
| 573 |
# Takes a string for the doctype, subset string, and namespace attrs string. |
| 574 |
|
| 575 |
_FRAGMENT_BUILDER_INTERNAL_SYSTEM_ID = \ |
| 576 |
"http://xml.python.org/entities/fragment-builder/internal" |
| 577 |
|
| 578 |
_FRAGMENT_BUILDER_TEMPLATE = ( |
| 579 |
'''\ |
| 580 |
<!DOCTYPE wrapper |
| 581 |
%%s [ |
| 582 |
<!ENTITY fragment-builder-internal |
| 583 |
SYSTEM "%s"> |
| 584 |
%%s |
| 585 |
]> |
| 586 |
<wrapper %%s |
| 587 |
>&fragment-builder-internal;</wrapper>''' |
| 588 |
% _FRAGMENT_BUILDER_INTERNAL_SYSTEM_ID) |
| 589 |
|
| 590 |
|
| 591 |
class FragmentBuilder(ExpatBuilder): |
| 592 |
"""Builder which constructs document fragments given XML source |
| 593 |
text and a context node. |
| 594 |
|
| 595 |
The context node is expected to provide information about the |
| 596 |
namespace declarations which are in scope at the start of the |
| 597 |
fragment. |
| 598 |
""" |
| 599 |
|
| 600 |
def __init__(self, context, options=None): |
| 601 |
if context.nodeType == DOCUMENT_NODE: |
| 602 |
self.originalDocument = context |
| 603 |
self.context = context |
| 604 |
else: |
| 605 |
self.originalDocument = context.ownerDocument |
| 606 |
self.context = context |
| 607 |
ExpatBuilder.__init__(self, options) |
| 608 |
|
| 609 |
def reset(self): |
| 610 |
ExpatBuilder.reset(self) |
| 611 |
self.fragment = None |
| 612 |
|
| 613 |
def parseFile(self, file): |
| 614 |
"""Parse a document fragment from a file object, returning the |
| 615 |
fragment node.""" |
| 616 |
return self.parseString(file.read()) |
| 617 |
|
| 618 |
def parseString(self, string): |
| 619 |
"""Parse a document fragment from a string, returning the |
| 620 |
fragment node.""" |
| 621 |
self._source = string |
| 622 |
parser = self.getParser() |
| 623 |
doctype = self.originalDocument.doctype |
| 624 |
ident = "" |
| 625 |
if doctype: |
| 626 |
subset = doctype.internalSubset or self._getDeclarations() |
| 627 |
if doctype.publicId: |
| 628 |
ident = ('PUBLIC "%s" "%s"' |
| 629 |
% (doctype.publicId, doctype.systemId)) |
| 630 |
elif doctype.systemId: |
| 631 |
ident = 'SYSTEM "%s"' % doctype.systemId |
| 632 |
else: |
| 633 |
subset = "" |
| 634 |
nsattrs = self._getNSattrs() # get ns decls from node's ancestors |
| 635 |
document = _FRAGMENT_BUILDER_TEMPLATE % (ident, subset, nsattrs) |
| 636 |
try: |
| 637 |
parser.Parse(document, True) |
| 638 |
except: |
| 639 |
self.reset() |
| 640 |
raise |
| 641 |
fragment = self.fragment |
| 642 |
self.reset() |
| 643 |
## self._parser = None |
| 644 |
return fragment |
| 645 |
|
| 646 |
def _getDeclarations(self): |
| 647 |
"""Re-create the internal subset from the DocumentType node. |
| 648 |
|
| 649 |
This is only needed if we don't already have the |
| 650 |
internalSubset as a string. |
| 651 |
""" |
| 652 |
doctype = self.context.ownerDocument.doctype |
| 653 |
s = "" |
| 654 |
if doctype: |
| 655 |
for i in range(doctype.notations.length): |
| 656 |
notation = doctype.notations.item(i) |
| 657 |
if s: |
| 658 |
s = s + "\n " |
| 659 |
s = "%s<!NOTATION %s" % (s, notation.nodeName) |
| 660 |
if notation.publicId: |
| 661 |
s = '%s PUBLIC "%s"\n "%s">' \ |
| 662 |
% (s, notation.publicId, notation.systemId) |
| 663 |
else: |
| 664 |
s = '%s SYSTEM "%s">' % (s, notation.systemId) |
| 665 |
for i in range(doctype.entities.length): |
| 666 |
entity = doctype.entities.item(i) |
| 667 |
if s: |
| 668 |
s = s + "\n " |
| 669 |
s = "%s<!ENTITY %s" % (s, entity.nodeName) |
| 670 |
if entity.publicId: |
| 671 |
s = '%s PUBLIC "%s"\n "%s"' \ |
| 672 |
% (s, entity.publicId, entity.systemId) |
| 673 |
elif entity.systemId: |
| 674 |
s = '%s SYSTEM "%s"' % (s, entity.systemId) |
| 675 |
else: |
| 676 |
s = '%s "%s"' % (s, entity.firstChild.data) |
| 677 |
if entity.notationName: |
| 678 |
s = "%s NOTATION %s" % (s, entity.notationName) |
| 679 |
s = s + ">" |
| 680 |
return s |
| 681 |
|
| 682 |
def _getNSattrs(self): |
| 683 |
return "" |
| 684 |
|
| 685 |
def external_entity_ref_handler(self, context, base, systemId, publicId): |
| 686 |
if systemId == _FRAGMENT_BUILDER_INTERNAL_SYSTEM_ID: |
| 687 |
# this entref is the one that we made to put the subtree |
| 688 |
# in; all of our given input is parsed in here. |
| 689 |
old_document = self.document |
| 690 |
old_cur_node = self.curNode |
| 691 |
parser = self._parser.ExternalEntityParserCreate(context) |
| 692 |
# put the real document back, parse into the fragment to return |
| 693 |
self.document = self.originalDocument |
| 694 |
self.fragment = self.document.createDocumentFragment() |
| 695 |
self.curNode = self.fragment |
| 696 |
try: |
| 697 |
parser.Parse(self._source, True) |
| 698 |
finally: |
| 699 |
self.curNode = old_cur_node |
| 700 |
self.document = old_document |
| 701 |
self._source = None |
| 702 |
return -1 |
| 703 |
else: |
| 704 |
return ExpatBuilder.external_entity_ref_handler( |
| 705 |
self, context, base, systemId, publicId) |
| 706 |
|
| 707 |
|
| 708 |
class Namespaces: |
| 709 |
"""Mix-in class for builders; adds support for namespaces.""" |
| 710 |
|
| 711 |
def _initNamespaces(self): |
| 712 |
# list of (prefix, uri) ns declarations. Namespace attrs are |
| 713 |
# constructed from this and added to the element's attrs. |
| 714 |
self._ns_ordered_prefixes = [] |
| 715 |
|
| 716 |
def createParser(self): |
| 717 |
"""Create a new namespace-handling parser.""" |
| 718 |
parser = expat.ParserCreate(namespace_separator=" ") |
| 719 |
parser.namespace_prefixes = True |
| 720 |
return parser |
| 721 |
|
| 722 |
def install(self, parser): |
| 723 |
"""Insert the namespace-handlers onto the parser.""" |
| 724 |
ExpatBuilder.install(self, parser) |
| 725 |
if self._options.namespace_declarations: |
| 726 |
parser.StartNamespaceDeclHandler = ( |
| 727 |
self.start_namespace_decl_handler) |
| 728 |
|
| 729 |
def start_namespace_decl_handler(self, prefix, uri): |
| 730 |
"""Push this namespace declaration on our storage.""" |
| 731 |
self._ns_ordered_prefixes.append((prefix, uri)) |
| 732 |
|
| 733 |
def start_element_handler(self, name, attributes): |
| 734 |
if ' ' in name: |
| 735 |
uri, localname, prefix, qname = _parse_ns_name(self, name) |
| 736 |
else: |
| 737 |
uri = EMPTY_NAMESPACE |
| 738 |
qname = name |
| 739 |
localname = None |
| 740 |
prefix = EMPTY_PREFIX |
| 741 |
node = minidom.Element(qname, uri, prefix, localname) |
| 742 |
node.ownerDocument = self.document |
| 743 |
_append_child(self.curNode, node) |
| 744 |
self.curNode = node |
| 745 |
|
| 746 |
if self._ns_ordered_prefixes: |
| 747 |
for prefix, uri in self._ns_ordered_prefixes: |
| 748 |
if prefix: |
| 749 |
a = minidom.Attr(_intern(self, 'xmlns:' + prefix), |
| 750 |
XMLNS_NAMESPACE, prefix, "xmlns") |
| 751 |
else: |
| 752 |
a = minidom.Attr("xmlns", XMLNS_NAMESPACE, |
| 753 |
"xmlns", EMPTY_PREFIX) |
| 754 |
a.value = uri |
| 755 |
a.ownerDocument = self.document |
| 756 |
_set_attribute_node(node, a) |
| 757 |
del self._ns_ordered_prefixes[:] |
| 758 |
|
| 759 |
if attributes: |
| 760 |
node._ensure_attributes() |
| 761 |
_attrs = node._attrs |
| 762 |
_attrsNS = node._attrsNS |
| 763 |
for i in range(0, len(attributes), 2): |
| 764 |
aname = attributes[i] |
| 765 |
value = attributes[i+1] |
| 766 |
if ' ' in aname: |
| 767 |
uri, localname, prefix, qname = _parse_ns_name(self, aname) |
| 768 |
a = minidom.Attr(qname, uri, localname, prefix) |
| 769 |
_attrs[qname] = a |
| 770 |
_attrsNS[(uri, localname)] = a |
| 771 |
else: |
| 772 |
a = minidom.Attr(aname, EMPTY_NAMESPACE, |
| 773 |
aname, EMPTY_PREFIX) |
| 774 |
_attrs[aname] = a |
| 775 |
_attrsNS[(EMPTY_NAMESPACE, aname)] = a |
| 776 |
a.ownerDocument = self.document |
| 777 |
a.value = value |
| 778 |
a.ownerElement = node |
| 779 |
|
| 780 |
if __debug__: |
| 781 |
# This only adds some asserts to the original |
| 782 |
# end_element_handler(), so we only define this when -O is not |
| 783 |
# used. If changing one, be sure to check the other to see if |
| 784 |
# it needs to be changed as well. |
| 785 |
# |
| 786 |
def end_element_handler(self, name): |
| 787 |
curNode = self.curNode |
| 788 |
if ' ' in name: |
| 789 |
uri, localname, prefix, qname = _parse_ns_name(self, name) |
| 790 |
assert (curNode.namespaceURI == uri |
| 791 |
and curNode.localName == localname |
| 792 |
and curNode.prefix == prefix), \ |
| 793 |
"element stack messed up! (namespace)" |
| 794 |
else: |
| 795 |
assert curNode.nodeName == name, \ |
| 796 |
"element stack messed up - bad nodeName" |
| 797 |
assert curNode.namespaceURI == EMPTY_NAMESPACE, \ |
| 798 |
"element stack messed up - bad namespaceURI" |
| 799 |
self.curNode = curNode.parentNode |
| 800 |
self._finish_end_element(curNode) |
| 801 |
|
| 802 |
|
| 803 |
class ExpatBuilderNS(Namespaces, ExpatBuilder): |
| 804 |
"""Document builder that supports namespaces.""" |
| 805 |
|
| 806 |
def reset(self): |
| 807 |
ExpatBuilder.reset(self) |
| 808 |
self._initNamespaces() |
| 809 |
|
| 810 |
|
| 811 |
class FragmentBuilderNS(Namespaces, FragmentBuilder): |
| 812 |
"""Fragment builder that supports namespaces.""" |
| 813 |
|
| 814 |
def reset(self): |
| 815 |
FragmentBuilder.reset(self) |
| 816 |
self._initNamespaces() |
| 817 |
|
| 818 |
def _getNSattrs(self): |
| 819 |
"""Return string of namespace attributes from this element and |
| 820 |
ancestors.""" |
| 821 |
# XXX This needs to be re-written to walk the ancestors of the |
| 822 |
# context to build up the namespace information from |
| 823 |
# declarations, elements, and attributes found in context. |
| 824 |
# Otherwise we have to store a bunch more data on the DOM |
| 825 |
# (though that *might* be more reliable -- not clear). |
| 826 |
attrs = "" |
| 827 |
context = self.context |
| 828 |
L = [] |
| 829 |
while context: |
| 830 |
if hasattr(context, '_ns_prefix_uri'): |
| 831 |
for prefix, uri in context._ns_prefix_uri.items(): |
| 832 |
# add every new NS decl from context to L and attrs string |
| 833 |
if prefix in L: |
| 834 |
continue |
| 835 |
L.append(prefix) |
| 836 |
if prefix: |
| 837 |
declname = "xmlns:" + prefix |
| 838 |
else: |
| 839 |
declname = "xmlns" |
| 840 |
if attrs: |
| 841 |
attrs = "%s\n %s='%s'" % (attrs, declname, uri) |
| 842 |
else: |
| 843 |
attrs = " %s='%s'" % (declname, uri) |
| 844 |
context = context.parentNode |
| 845 |
return attrs |
| 846 |
|
| 847 |
|
| 848 |
class ParseEscape(Exception): |
| 849 |
"""Exception raised to short-circuit parsing in InternalSubsetExtractor.""" |
| 850 |
pass |
| 851 |
|
| 852 |
class InternalSubsetExtractor(ExpatBuilder): |
| 853 |
"""XML processor which can rip out the internal document type subset.""" |
| 854 |
|
| 855 |
subset = None |
| 856 |
|
| 857 |
def getSubset(self): |
| 858 |
"""Return the internal subset as a string.""" |
| 859 |
return self.subset |
| 860 |
|
| 861 |
def parseFile(self, file): |
| 862 |
try: |
| 863 |
ExpatBuilder.parseFile(self, file) |
| 864 |
except ParseEscape: |
| 865 |
pass |
| 866 |
|
| 867 |
def parseString(self, string): |
| 868 |
try: |
| 869 |
ExpatBuilder.parseString(self, string) |
| 870 |
except ParseEscape: |
| 871 |
pass |
| 872 |
|
| 873 |
def install(self, parser): |
| 874 |
parser.StartDoctypeDeclHandler = self.start_doctype_decl_handler |
| 875 |
parser.StartElementHandler = self.start_element_handler |
| 876 |
|
| 877 |
def start_doctype_decl_handler(self, name, publicId, systemId, |
| 878 |
has_internal_subset): |
| 879 |
if has_internal_subset: |
| 880 |
parser = self.getParser() |
| 881 |
self.subset = [] |
| 882 |
parser.DefaultHandler = self.subset.append |
| 883 |
parser.EndDoctypeDeclHandler = self.end_doctype_decl_handler |
| 884 |
else: |
| 885 |
raise ParseEscape() |
| 886 |
|
| 887 |
def end_doctype_decl_handler(self): |
| 888 |
s = ''.join(self.subset).replace('\r\n', '\n').replace('\r', '\n') |
| 889 |
self.subset = s |
| 890 |
raise ParseEscape() |
| 891 |
|
| 892 |
def start_element_handler(self, name, attrs): |
| 893 |
raise ParseEscape() |
| 894 |
|
| 895 |
|
| 896 |
def parse(file, namespaces=True): |
| 897 |
"""Parse a document, returning the resulting Document node. |
| 898 |
|
| 899 |
'file' may be either a file name or an open file object. |
| 900 |
""" |
| 901 |
if namespaces: |
| 902 |
builder = ExpatBuilderNS() |
| 903 |
else: |
| 904 |
builder = ExpatBuilder() |
| 905 |
|
| 906 |
if isinstance(file, str): |
| 907 |
with open(file, 'rb') as fp: |
| 908 |
result = builder.parseFile(fp) |
| 909 |
else: |
| 910 |
result = builder.parseFile(file) |
| 911 |
return result |
| 912 |
|
| 913 |
|
| 914 |
def parseString(string, namespaces=True): |
| 915 |
"""Parse a document from a string, returning the resulting |
| 916 |
Document node. |
| 917 |
""" |
| 918 |
if namespaces: |
| 919 |
builder = ExpatBuilderNS() |
| 920 |
else: |
| 921 |
builder = ExpatBuilder() |
| 922 |
return builder.parseString(string) |
| 923 |
|
| 924 |
|
| 925 |
def parseFragment(file, context, namespaces=True): |
| 926 |
"""Parse a fragment of a document, given the context from which it |
| 927 |
was originally extracted. context should be the parent of the |
| 928 |
node(s) which are in the fragment. |
| 929 |
|
| 930 |
'file' may be either a file name or an open file object. |
| 931 |
""" |
| 932 |
if namespaces: |
| 933 |
builder = FragmentBuilderNS(context) |
| 934 |
else: |
| 935 |
builder = FragmentBuilder(context) |
| 936 |
|
| 937 |
if isinstance(file, str): |
| 938 |
with open(file, 'rb') as fp: |
| 939 |
result = builder.parseFile(fp) |
| 940 |
else: |
| 941 |
result = builder.parseFile(file) |
| 942 |
return result |
| 943 |
|
| 944 |
|
| 945 |
def parseFragmentString(string, context, namespaces=True): |
| 946 |
"""Parse a fragment of a document from a string, given the context |
| 947 |
from which it was originally extracted. context should be the |
| 948 |
parent of the node(s) which are in the fragment. |
| 949 |
""" |
| 950 |
if namespaces: |
| 951 |
builder = FragmentBuilderNS(context) |
| 952 |
else: |
| 953 |
builder = FragmentBuilder(context) |
| 954 |
return builder.parseString(string) |
| 955 |
|
| 956 |
|
| 957 |
def makeBuilder(options): |
| 958 |
"""Create a builder based on an Options object.""" |
| 959 |
if options.namespaces: |
| 960 |
return ExpatBuilderNS(options) |
| 961 |
else: |
| 962 |
return ExpatBuilder(options) |