Mercurial > repos > rliterman > csp2

annotate CSP2/CSP2_env/env-d9b9114564458d9d-741b3de822f2aaca6c6caa4325c4afce/lib/python3.8/xml/sax/saxutils.py @ 68:5028fdace37b

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload commit 2e9511a184a1ca667c7be0c6321a36dc4e3d116d
author jpayne
date Tue, 18 Mar 2025 16:23:26 -0400
parents
children
rev   line source
jpayne@68 1 """\
jpayne@68 2 A library of useful helper classes to the SAX classes, for the
jpayne@68 3 convenience of application and driver writers.
jpayne@68 4 """
jpayne@68 5
jpayne@68 6 import os, urllib.parse, urllib.request
jpayne@68 7 import io
jpayne@68 8 import codecs
jpayne@68 9 from . import handler
jpayne@68 10 from . import xmlreader
jpayne@68 11
jpayne@68 12 def __dict_replace(s, d):
jpayne@68 13 """Replace substrings of a string using a dictionary."""
jpayne@68 14 for key, value in d.items():
jpayne@68 15 s = s.replace(key, value)
jpayne@68 16 return s
jpayne@68 17
jpayne@68 18 def escape(data, entities={}):
jpayne@68 19 """Escape &, <, and > in a string of data.
jpayne@68 20
jpayne@68 21 You can escape other strings of data by passing a dictionary as
jpayne@68 22 the optional entities parameter. The keys and values must all be
jpayne@68 23 strings; each key will be replaced with its corresponding value.
jpayne@68 24 """
jpayne@68 25
jpayne@68 26 # must do ampersand first
jpayne@68 27 data = data.replace("&", "&amp;")
jpayne@68 28 data = data.replace(">", "&gt;")
jpayne@68 29 data = data.replace("<", "&lt;")
jpayne@68 30 if entities:
jpayne@68 31 data = __dict_replace(data, entities)
jpayne@68 32 return data
jpayne@68 33
jpayne@68 34 def unescape(data, entities={}):
jpayne@68 35 """Unescape &amp;, &lt;, and &gt; in a string of data.
jpayne@68 36
jpayne@68 37 You can unescape other strings of data by passing a dictionary as
jpayne@68 38 the optional entities parameter. The keys and values must all be
jpayne@68 39 strings; each key will be replaced with its corresponding value.
jpayne@68 40 """
jpayne@68 41 data = data.replace("&lt;", "<")
jpayne@68 42 data = data.replace("&gt;", ">")
jpayne@68 43 if entities:
jpayne@68 44 data = __dict_replace(data, entities)
jpayne@68 45 # must do ampersand last
jpayne@68 46 return data.replace("&amp;", "&")
jpayne@68 47
jpayne@68 48 def quoteattr(data, entities={}):
jpayne@68 49 """Escape and quote an attribute value.
jpayne@68 50
jpayne@68 51 Escape &, <, and > in a string of data, then quote it for use as
jpayne@68 52 an attribute value. The \" character will be escaped as well, if
jpayne@68 53 necessary.
jpayne@68 54
jpayne@68 55 You can escape other strings of data by passing a dictionary as
jpayne@68 56 the optional entities parameter. The keys and values must all be
jpayne@68 57 strings; each key will be replaced with its corresponding value.
jpayne@68 58 """
jpayne@68 59 entities = {**entities, '\n': '&#10;', '\r': '&#13;', '\t':'&#9;'}
jpayne@68 60 data = escape(data, entities)
jpayne@68 61 if '"' in data:
jpayne@68 62 if "'" in data:
jpayne@68 63 data = '"%s"' % data.replace('"', "&quot;")
jpayne@68 64 else:
jpayne@68 65 data = "'%s'" % data
jpayne@68 66 else:
jpayne@68 67 data = '"%s"' % data
jpayne@68 68 return data
jpayne@68 69
jpayne@68 70
jpayne@68 71 def _gettextwriter(out, encoding):
jpayne@68 72 if out is None:
jpayne@68 73 import sys
jpayne@68 74 return sys.stdout
jpayne@68 75
jpayne@68 76 if isinstance(out, io.TextIOBase):
jpayne@68 77 # use a text writer as is
jpayne@68 78 return out
jpayne@68 79
jpayne@68 80 if isinstance(out, (codecs.StreamWriter, codecs.StreamReaderWriter)):
jpayne@68 81 # use a codecs stream writer as is
jpayne@68 82 return out
jpayne@68 83
jpayne@68 84 # wrap a binary writer with TextIOWrapper
jpayne@68 85 if isinstance(out, io.RawIOBase):
jpayne@68 86 # Keep the original file open when the TextIOWrapper is
jpayne@68 87 # destroyed
jpayne@68 88 class _wrapper:
jpayne@68 89 __class__ = out.__class__
jpayne@68 90 def __getattr__(self, name):
jpayne@68 91 return getattr(out, name)
jpayne@68 92 buffer = _wrapper()
jpayne@68 93 buffer.close = lambda: None
jpayne@68 94 else:
jpayne@68 95 # This is to handle passed objects that aren't in the
jpayne@68 96 # IOBase hierarchy, but just have a write method
jpayne@68 97 buffer = io.BufferedIOBase()
jpayne@68 98 buffer.writable = lambda: True
jpayne@68 99 buffer.write = out.write
jpayne@68 100 try:
jpayne@68 101 # TextIOWrapper uses this methods to determine
jpayne@68 102 # if BOM (for UTF-16, etc) should be added
jpayne@68 103 buffer.seekable = out.seekable
jpayne@68 104 buffer.tell = out.tell
jpayne@68 105 except AttributeError:
jpayne@68 106 pass
jpayne@68 107 return io.TextIOWrapper(buffer, encoding=encoding,
jpayne@68 108 errors='xmlcharrefreplace',
jpayne@68 109 newline='\n',
jpayne@68 110 write_through=True)
jpayne@68 111
jpayne@68 112 class XMLGenerator(handler.ContentHandler):
jpayne@68 113
jpayne@68 114 def __init__(self, out=None, encoding="iso-8859-1", short_empty_elements=False):
jpayne@68 115 handler.ContentHandler.__init__(self)
jpayne@68 116 out = _gettextwriter(out, encoding)
jpayne@68 117 self._write = out.write
jpayne@68 118 self._flush = out.flush
jpayne@68 119 self._ns_contexts = [{}] # contains uri -> prefix dicts
jpayne@68 120 self._current_context = self._ns_contexts[-1]
jpayne@68 121 self._undeclared_ns_maps = []
jpayne@68 122 self._encoding = encoding
jpayne@68 123 self._short_empty_elements = short_empty_elements
jpayne@68 124 self._pending_start_element = False
jpayne@68 125
jpayne@68 126 def _qname(self, name):
jpayne@68 127 """Builds a qualified name from a (ns_url, localname) pair"""
jpayne@68 128 if name[0]:
jpayne@68 129 # Per http://www.w3.org/XML/1998/namespace, The 'xml' prefix is
jpayne@68 130 # bound by definition to http://www.w3.org/XML/1998/namespace. It
jpayne@68 131 # does not need to be declared and will not usually be found in
jpayne@68 132 # self._current_context.
jpayne@68 133 if 'http://www.w3.org/XML/1998/namespace' == name[0]:
jpayne@68 134 return 'xml:' + name[1]
jpayne@68 135 # The name is in a non-empty namespace
jpayne@68 136 prefix = self._current_context[name[0]]
jpayne@68 137 if prefix:
jpayne@68 138 # If it is not the default namespace, prepend the prefix
jpayne@68 139 return prefix + ":" + name[1]
jpayne@68 140 # Return the unqualified name
jpayne@68 141 return name[1]
jpayne@68 142
jpayne@68 143 def _finish_pending_start_element(self,endElement=False):
jpayne@68 144 if self._pending_start_element:
jpayne@68 145 self._write('>')
jpayne@68 146 self._pending_start_element = False
jpayne@68 147
jpayne@68 148 # ContentHandler methods
jpayne@68 149
jpayne@68 150 def startDocument(self):
jpayne@68 151 self._write('<?xml version="1.0" encoding="%s"?>\n' %
jpayne@68 152 self._encoding)
jpayne@68 153
jpayne@68 154 def endDocument(self):
jpayne@68 155 self._flush()
jpayne@68 156
jpayne@68 157 def startPrefixMapping(self, prefix, uri):
jpayne@68 158 self._ns_contexts.append(self._current_context.copy())
jpayne@68 159 self._current_context[uri] = prefix
jpayne@68 160 self._undeclared_ns_maps.append((prefix, uri))
jpayne@68 161
jpayne@68 162 def endPrefixMapping(self, prefix):
jpayne@68 163 self._current_context = self._ns_contexts[-1]
jpayne@68 164 del self._ns_contexts[-1]
jpayne@68 165
jpayne@68 166 def startElement(self, name, attrs):
jpayne@68 167 self._finish_pending_start_element()
jpayne@68 168 self._write('<' + name)
jpayne@68 169 for (name, value) in attrs.items():
jpayne@68 170 self._write(' %s=%s' % (name, quoteattr(value)))
jpayne@68 171 if self._short_empty_elements:
jpayne@68 172 self._pending_start_element = True
jpayne@68 173 else:
jpayne@68 174 self._write(">")
jpayne@68 175
jpayne@68 176 def endElement(self, name):
jpayne@68 177 if self._pending_start_element:
jpayne@68 178 self._write('/>')
jpayne@68 179 self._pending_start_element = False
jpayne@68 180 else:
jpayne@68 181 self._write('</%s>' % name)
jpayne@68 182
jpayne@68 183 def startElementNS(self, name, qname, attrs):
jpayne@68 184 self._finish_pending_start_element()
jpayne@68 185 self._write('<' + self._qname(name))
jpayne@68 186
jpayne@68 187 for prefix, uri in self._undeclared_ns_maps:
jpayne@68 188 if prefix:
jpayne@68 189 self._write(' xmlns:%s="%s"' % (prefix, uri))
jpayne@68 190 else:
jpayne@68 191 self._write(' xmlns="%s"' % uri)
jpayne@68 192 self._undeclared_ns_maps = []
jpayne@68 193
jpayne@68 194 for (name, value) in attrs.items():
jpayne@68 195 self._write(' %s=%s' % (self._qname(name), quoteattr(value)))
jpayne@68 196 if self._short_empty_elements:
jpayne@68 197 self._pending_start_element = True
jpayne@68 198 else:
jpayne@68 199 self._write(">")
jpayne@68 200
jpayne@68 201 def endElementNS(self, name, qname):
jpayne@68 202 if self._pending_start_element:
jpayne@68 203 self._write('/>')
jpayne@68 204 self._pending_start_element = False
jpayne@68 205 else:
jpayne@68 206 self._write('</%s>' % self._qname(name))
jpayne@68 207
jpayne@68 208 def characters(self, content):
jpayne@68 209 if content:
jpayne@68 210 self._finish_pending_start_element()
jpayne@68 211 if not isinstance(content, str):
jpayne@68 212 content = str(content, self._encoding)
jpayne@68 213 self._write(escape(content))
jpayne@68 214
jpayne@68 215 def ignorableWhitespace(self, content):
jpayne@68 216 if content:
jpayne@68 217 self._finish_pending_start_element()
jpayne@68 218 if not isinstance(content, str):
jpayne@68 219 content = str(content, self._encoding)
jpayne@68 220 self._write(content)
jpayne@68 221
jpayne@68 222 def processingInstruction(self, target, data):
jpayne@68 223 self._finish_pending_start_element()
jpayne@68 224 self._write('<?%s %s?>' % (target, data))
jpayne@68 225
jpayne@68 226
jpayne@68 227 class XMLFilterBase(xmlreader.XMLReader):
jpayne@68 228 """This class is designed to sit between an XMLReader and the
jpayne@68 229 client application's event handlers. By default, it does nothing
jpayne@68 230 but pass requests up to the reader and events on to the handlers
jpayne@68 231 unmodified, but subclasses can override specific methods to modify
jpayne@68 232 the event stream or the configuration requests as they pass
jpayne@68 233 through."""
jpayne@68 234
jpayne@68 235 def __init__(self, parent = None):
jpayne@68 236 xmlreader.XMLReader.__init__(self)
jpayne@68 237 self._parent = parent
jpayne@68 238
jpayne@68 239 # ErrorHandler methods
jpayne@68 240
jpayne@68 241 def error(self, exception):
jpayne@68 242 self._err_handler.error(exception)
jpayne@68 243
jpayne@68 244 def fatalError(self, exception):
jpayne@68 245 self._err_handler.fatalError(exception)
jpayne@68 246
jpayne@68 247 def warning(self, exception):
jpayne@68 248 self._err_handler.warning(exception)
jpayne@68 249
jpayne@68 250 # ContentHandler methods
jpayne@68 251
jpayne@68 252 def setDocumentLocator(self, locator):
jpayne@68 253 self._cont_handler.setDocumentLocator(locator)
jpayne@68 254
jpayne@68 255 def startDocument(self):
jpayne@68 256 self._cont_handler.startDocument()
jpayne@68 257
jpayne@68 258 def endDocument(self):
jpayne@68 259 self._cont_handler.endDocument()
jpayne@68 260
jpayne@68 261 def startPrefixMapping(self, prefix, uri):
jpayne@68 262 self._cont_handler.startPrefixMapping(prefix, uri)
jpayne@68 263
jpayne@68 264 def endPrefixMapping(self, prefix):
jpayne@68 265 self._cont_handler.endPrefixMapping(prefix)
jpayne@68 266
jpayne@68 267 def startElement(self, name, attrs):
jpayne@68 268 self._cont_handler.startElement(name, attrs)
jpayne@68 269
jpayne@68 270 def endElement(self, name):
jpayne@68 271 self._cont_handler.endElement(name)
jpayne@68 272
jpayne@68 273 def startElementNS(self, name, qname, attrs):
jpayne@68 274 self._cont_handler.startElementNS(name, qname, attrs)
jpayne@68 275
jpayne@68 276 def endElementNS(self, name, qname):
jpayne@68 277 self._cont_handler.endElementNS(name, qname)
jpayne@68 278
jpayne@68 279 def characters(self, content):
jpayne@68 280 self._cont_handler.characters(content)
jpayne@68 281
jpayne@68 282 def ignorableWhitespace(self, chars):
jpayne@68 283 self._cont_handler.ignorableWhitespace(chars)
jpayne@68 284
jpayne@68 285 def processingInstruction(self, target, data):
jpayne@68 286 self._cont_handler.processingInstruction(target, data)
jpayne@68 287
jpayne@68 288 def skippedEntity(self, name):
jpayne@68 289 self._cont_handler.skippedEntity(name)
jpayne@68 290
jpayne@68 291 # DTDHandler methods
jpayne@68 292
jpayne@68 293 def notationDecl(self, name, publicId, systemId):
jpayne@68 294 self._dtd_handler.notationDecl(name, publicId, systemId)
jpayne@68 295
jpayne@68 296 def unparsedEntityDecl(self, name, publicId, systemId, ndata):
jpayne@68 297 self._dtd_handler.unparsedEntityDecl(name, publicId, systemId, ndata)
jpayne@68 298
jpayne@68 299 # EntityResolver methods
jpayne@68 300
jpayne@68 301 def resolveEntity(self, publicId, systemId):
jpayne@68 302 return self._ent_handler.resolveEntity(publicId, systemId)
jpayne@68 303
jpayne@68 304 # XMLReader methods
jpayne@68 305
jpayne@68 306 def parse(self, source):
jpayne@68 307 self._parent.setContentHandler(self)
jpayne@68 308 self._parent.setErrorHandler(self)
jpayne@68 309 self._parent.setEntityResolver(self)
jpayne@68 310 self._parent.setDTDHandler(self)
jpayne@68 311 self._parent.parse(source)
jpayne@68 312
jpayne@68 313 def setLocale(self, locale):
jpayne@68 314 self._parent.setLocale(locale)
jpayne@68 315
jpayne@68 316 def getFeature(self, name):
jpayne@68 317 return self._parent.getFeature(name)
jpayne@68 318
jpayne@68 319 def setFeature(self, name, state):
jpayne@68 320 self._parent.setFeature(name, state)
jpayne@68 321
jpayne@68 322 def getProperty(self, name):
jpayne@68 323 return self._parent.getProperty(name)
jpayne@68 324
jpayne@68 325 def setProperty(self, name, value):
jpayne@68 326 self._parent.setProperty(name, value)
jpayne@68 327
jpayne@68 328 # XMLFilter methods
jpayne@68 329
jpayne@68 330 def getParent(self):
jpayne@68 331 return self._parent
jpayne@68 332
jpayne@68 333 def setParent(self, parent):
jpayne@68 334 self._parent = parent
jpayne@68 335
jpayne@68 336 # --- Utility functions
jpayne@68 337
jpayne@68 338 def prepare_input_source(source, base=""):
jpayne@68 339 """This function takes an InputSource and an optional base URL and
jpayne@68 340 returns a fully resolved InputSource object ready for reading."""
jpayne@68 341
jpayne@68 342 if isinstance(source, os.PathLike):
jpayne@68 343 source = os.fspath(source)
jpayne@68 344 if isinstance(source, str):
jpayne@68 345 source = xmlreader.InputSource(source)
jpayne@68 346 elif hasattr(source, "read"):
jpayne@68 347 f = source
jpayne@68 348 source = xmlreader.InputSource()
jpayne@68 349 if isinstance(f.read(0), str):
jpayne@68 350 source.setCharacterStream(f)
jpayne@68 351 else:
jpayne@68 352 source.setByteStream(f)
jpayne@68 353 if hasattr(f, "name") and isinstance(f.name, str):
jpayne@68 354 source.setSystemId(f.name)
jpayne@68 355
jpayne@68 356 if source.getCharacterStream() is None and source.getByteStream() is None:
jpayne@68 357 sysid = source.getSystemId()
jpayne@68 358 basehead = os.path.dirname(os.path.normpath(base))
jpayne@68 359 sysidfilename = os.path.join(basehead, sysid)
jpayne@68 360 if os.path.isfile(sysidfilename):
jpayne@68 361 source.setSystemId(sysidfilename)
jpayne@68 362 f = open(sysidfilename, "rb")
jpayne@68 363 else:
jpayne@68 364 source.setSystemId(urllib.parse.urljoin(base, sysid))
jpayne@68 365 f = urllib.request.urlopen(source.getSystemId())
jpayne@68 366
jpayne@68 367 source.setByteStream(f)
jpayne@68 368
jpayne@68 369 return source