def to_xml(text): try: if PY2: # On python2, fromstring expects an encoded string return fromstring((text[BOM_LEN:] if text.startswith(BOM) else text).encode('utf-8')) return fromstring(text[BOM_LEN:] if text.startswith(BOM) else text) except ParseError: # Exchange servers may spit out the weirdest XML. lxml is pretty good at recovering from errors log.warning('Fallback to lxml processing of faulty XML') magical_parser = XMLParser(recover=True, resolve_entities=False) magical_parser.set_element_class_lookup(ElementDefaultClassLookup(element=RestrictedElement)) no_bom_text = text[BOM_LEN:] if text.startswith(BOM) else text try: root = parse(io.BytesIO(no_bom_text.encode('utf-8')), parser=magical_parser) except AssertionError as e: raise ParseError(*e.args) try: return fromstring(tostring(root)) except ParseError as e: if hasattr(e, 'position'): e.lineno, e.offset = e.position if not e.lineno: raise ParseError('%s' % text_type(e)) try: offending_line = no_bom_text.splitlines()[e.lineno - 1] except IndexError: raise ParseError('%s' % text_type(e)) else: offending_excerpt = offending_line[max(0, e.offset - 20):e.offset + 20] raise ParseError('%s\nOffending text: [...]%s[...]' % (text_type(e), offending_excerpt)) except TypeError: raise ParseError('This is not XML: %s' % text)
def getWaterLevelRawSixMinuteDataExt(self, beginDate, endDate, station, datum='MLLW', unit='feet', shift='GMT'): if self.logger: self.logger.debug("SOAP WSDL: %s" % (self.baseUrl)) soapClient = Client(self.baseUrl, retxml=True) if(unit == 'feet'): unit = 1 else: unit = 2 if(shift == 'GMT'): shift = 0 else: shift = 1 ret_xml = soapClient.service.getWaterLevelRawSixMin(station, beginDate, endDate, datum, unit, shift) if self.logger: self.logger.debug(ret_xml) parser = XMLParser(remove_blank_text=True, huge_tree=True) parser.set_element_class_lookup(objectify.ObjectifyElementClassLookup()) objectify.set_default_parser(parser) root = objectify.fromstring(ret_xml) objectify.deannotate(root, cleanup_namespaces=True) return(root)
def loadXml(data, **parserOptions): """Load SVG from an XML string, fileName, or file-like object. @type data: string or file-like object @param data: The serialized SVG, fileName, or file-like object that generates SVG as XML. @param **parserOptions: Arguments passed to lxml's U{XMLParser<http://lxml.de/api/lxml.etree.XMLParser-class.html>}. @rtype: SvgBinding @return: An in-memory representation of the SVG. """ if isinstance(data, basestring): if os.path.exists(data): data = open(data) else: data = StringIO(data) newParserOptions = {"huge_tree": True} newParserOptions.update(parserOptions) parserOptions = newParserOptions parser = XMLParser(**parserOptions) lookup = ElementDefaultClassLookup(element=SvgBinding) parser.set_element_class_lookup(lookup) return parse(data, parser).getroot()
def loadXml(self, data, validate=True, postValidate=True, **parserOptions): """Load a PMML model represented as an XML string, fileName, URI, or file-like object. Note that the XML file or string may be Gzip-compressed. @type data: string or file-like object @param data: The data to load. @type validate: bool @param validate: If True, validate the resulting PmmlBinding against this ModelLoader's XSD schema while loading. @type postValidate: bool @param postValidate: If True, run post-XSD validation checks. (Note: very few PmmlBinding subclasses have postValidation tests defined as of May 2013.) @param **parserOptions: Arguments passed to lxml's U{XMLParser<http://lxml.de/api/lxml.etree.XMLParser-class.html>}. @rtype: PmmlBinding @return: In-memory PMML object. """ if isinstance(data, basestring): if len(data) >= 2 and data[0:2] == "\x1f\x8b": data = gzip.GzipFile(fileobj=StringIO(data)) elif data.find("<") != -1: data = StringIO(data) if validate: if self.preparedSchema is None: self.preparedSchema = XMLSchema(self.schema) schema = self.preparedSchema else: schema = None newParserOptions = {"schema": schema, "huge_tree": True} newParserOptions.update(parserOptions) parserOptions = newParserOptions parser = XMLParser(**parserOptions) lookup = ElementNamespaceClassLookup() namespace = lookup.get_namespace(defs.PMML_NAMESPACE) for xsdElement in self.schema.xpath( "xs:element", namespaces={"xs": defs.XSD_NAMESPACE}): namespace[xsdElement.attrib["name"]] = PmmlBinding namespace.update(self.tagToClass) parser.set_element_class_lookup(lookup) # ElementNamespaceClassLookup don't work with iterparse, so we have to parse all at once and then iterwalk pmmlBinding = parse(data, parser).getroot() pmmlBinding.modelLoader = self if postValidate: for event, elem in iterwalk(pmmlBinding, events=("end", ), tag="{%s}*" % defs.PMML_NAMESPACE): if isinstance(elem, PmmlBinding): elem.postValidate() return pmmlBinding
def loadXml(self, data, validate=True, postValidate=True, **parserOptions): """Load a PMML model represented as an XML string, fileName, URI, or file-like object. Note that the XML file or string may be Gzip-compressed. @type data: string or file-like object @param data: The data to load. @type validate: bool @param validate: If True, validate the resulting PmmlBinding against this ModelLoader's XSD schema while loading. @type postValidate: bool @param postValidate: If True, run post-XSD validation checks. (Note: very few PmmlBinding subclasses have postValidation tests defined as of May 2013.) @param **parserOptions: Arguments passed to lxml's U{XMLParser<http://lxml.de/api/lxml.etree.XMLParser-class.html>}. @rtype: PmmlBinding @return: In-memory PMML object. """ if isinstance(data, basestring): if len(data) >= 2 and data[0:2] == "\x1f\x8b": data = gzip.GzipFile(fileobj=StringIO(data)) elif data.find("<") != -1: data = StringIO(data) if validate: if self.preparedSchema is None: self.preparedSchema = XMLSchema(self.schema) schema = self.preparedSchema else: schema = None newParserOptions = {"schema": schema, "huge_tree": True} newParserOptions.update(parserOptions) parserOptions = newParserOptions parser = XMLParser(**parserOptions) lookup = ElementNamespaceClassLookup() namespace = lookup.get_namespace(defs.PMML_NAMESPACE) for xsdElement in self.schema.xpath("xs:element", namespaces={"xs": defs.XSD_NAMESPACE}): namespace[xsdElement.attrib["name"]] = PmmlBinding namespace.update(self.tagToClass) parser.set_element_class_lookup(lookup) # ElementNamespaceClassLookup don't work with iterparse, so we have to parse all at once and then iterwalk pmmlBinding = parse(data, parser).getroot() pmmlBinding.modelLoader = self if postValidate: for event, elem in iterwalk(pmmlBinding, events=("end",), tag="{%s}*" % defs.PMML_NAMESPACE): if isinstance(elem, PmmlBinding): elem.postValidate() return pmmlBinding
def dictnode_to_lxml(tree, node_lookup=None, encoding=None): """ Input: A dictionary-based representation of a node tree. Output: An lxml representation of the same. Each dictionary has three attributes: name -- The type of node, a string. In html, this would be the tag name. text -- The content of the node: <b>text</b> tail -- Any content after the end of this node, but before the start of the next: <br/>tail attrs -- A dictionary of any extra attributes. children -- An ordered list of more node-dictionaries. """ if not node_lookup: from refactorlib.node import node_lookup from lxml.etree import Element, XMLParser root = None stack = [(tree, root)] while stack: node, parent = stack.pop() # sort attributes for determinism attrs = node.get('attrs', {}) attrs = {k: attrs[k] for k in sorted(attrs)} if parent is None: # We use this roundabout method becuase the encoding is always set # to 'UTF8' if we use parser.makeelement() parser = XMLParser(encoding=encoding) parser.set_element_class_lookup(node_lookup) parser.feed(b'<a/>') lxmlnode = parser.close() lxmlnode.tag = node['name'] lxmlnode.attrib.update(attrs) root = lxmlnode else: lxmlnode = Element(node['name'], attrib=attrs) parent.append(lxmlnode) lxmlnode.text = node['text'] lxmlnode.tail = node['tail'] for child in reversed(node['children']): stack.append((child, lxmlnode)) return root
def dictnode_to_lxml(tree, node_lookup=None, encoding=None): """ Input: A dictionary-based representation of a node tree. Output: An lxml representation of the same. Each dictionary has three attributes: name -- The type of node, a string. In html, this would be the tag name. text -- The content of the node: <b>text</b> tail -- Any content after the end of this node, but before the start of the next: <br/>tail attrs -- A dictionary of any extra attributes. children -- An ordered list of more node-dictionaries. """ if not node_lookup: from node import node_lookup from lxml.etree import XMLParser lxml_parser_object = XMLParser(encoding=encoding) lxml_parser_object.set_element_class_lookup(node_lookup) Element = lxml_parser_object.makeelement root = None stack = [ (tree,root) ] while stack: node, parent = stack.pop() if parent is None: # We use this roundabout method becuase the encoding is always set # to 'UTF8' if we use parser.makeelement() lxml_parser_object.feed('<trash></trash>') lxmlnode = lxml_parser_object.close() lxmlnode.tag = node['name'] lxmlnode.attrib.update(node.get('attrs', {})) root = lxmlnode else: lxmlnode = Element(node['name'], attrib=node.get('attrs', {})) parent.append(lxmlnode) lxmlnode.text = node['text'] lxmlnode.tail = node['tail'] for child in reversed(node['children']): stack.append((child, lxmlnode)) return root
def makeElementMaker(): """Obtain a factory for making in-memory SVG objects. This factory is an lxml ElementMaker, pre-loaded with the SVG namespace and this ModelLoader's current tag-to-class relationship. See the lxml documentation for how to use an ElementMaker. The C{SvgBinding} class has an C{elementMaker} attribute that should be used instead of calling this function. @see: The lxml U{ElementMaker documentation<http://lxml.de/api/lxml.builder.ElementMaker-class.html>}, which explains how to use an ElementMaker factory. """ parser = XMLParser(huge_tree=True) lookup = ElementDefaultClassLookup(element=SvgBinding) parser.set_element_class_lookup(lookup) return ElementMaker(namespace=defs.SVG_NAMESPACE, nsmap={None: defs.SVG_NAMESPACE, "xlink": defs.XLINK_NAMESPACE}, makeelement=parser.makeelement)
def makeElementMaker(): """Obtain a factory for making in-memory SVG objects. This factory is an lxml ElementMaker, pre-loaded with the SVG namespace and this ModelLoader's current tag-to-class relationship. See the lxml documentation for how to use an ElementMaker. The C{SvgBinding} class has an C{elementMaker} attribute that should be used instead of calling this function. @see: The lxml U{ElementMaker documentation<http://lxml.de/api/lxml.builder.ElementMaker-class.html>}, which explains how to use an ElementMaker factory. """ parser = XMLParser(huge_tree=True) lookup = ElementDefaultClassLookup(element=SvgBinding) parser.set_element_class_lookup(lookup) return ElementMaker(namespace=defs.SVG_NAMESPACE, nsmap={ None: defs.SVG_NAMESPACE, "xlink": defs.XLINK_NAMESPACE }, makeelement=parser.makeelement)
# Copyright (C) 2012-2018 by Dr. Dieter Maurer <*****@*****.**>; see 'LICENSE.txt' for details """Auxiliary classes to construct signature/encryption templates.""" from lxml.etree import ElementBase, \ parse as et_parse, fromstring as et_fromstring, XML as et_xml, \ XMLParser, ElementNamespaceClassLookup, ElementDefaultClassLookup from dm.xmlsec.binding import DSigNs, dsig, EncNs, enc # set up our own parser and related `etree` infrastructure parser = XMLParser() # apparently, `parser` has a `set_element_class_lookup` but not corresponding `get` #class_lookup = ElementNamespaceClassLookup(parser.get_element_class_lookup()) class_lookup = ElementNamespaceClassLookup(ElementDefaultClassLookup()) parser.set_element_class_lookup(class_lookup) Element = parser.makeelement def SubElement(node, *args, **kw): node.append(Element(*args, **kw)) def parse(file, parser=parser): return et_parse(file, parser=parser) def fromstring(s, parser=parser): return et_fromstring(s, parser=parser) def XML(s, parser=parser):
def create_lxml_context(): parser = XMLParser(no_network=True) parser.set_element_class_lookup( ElementDefaultClassLookup(element=Element, comment=Comment)) return parser
def loadJson(self, data, validate=True, postValidate=True, **parserOptions): """Load a PMML model represented as a JSON string, fileName, dict, or file-like object. There is no standard XML-to-JSON specification, so we define our own. Our specification is very similar to U{this proposal<http://www.xml.com/pub/a/2006/05/31/converting-between-xml-and-json.html>}, which collects subelements of different tagnames into different JSON lists, rather than having one long list and needing to specify the tag of each element in that list. This has the following advantages, particularly useful for PMML: - Frequent tagnames (like <Segment>) are not repeated, wasting space. - Subelements with a given tagname can be quickly queried, without having to iterate over a list that contains non-matching tagnames. It has the following disadvantages: - The relative order of subelements with different tagnames is not preserved. We therefore additionally include a JSON attribute named "#" to specify the ordering of subelements in the XML representation. Also, the specification referenced above represents single-child subelements as JSON objects and multiple children as JSON lists, but for consistency and ease of parsing, we always use lists. The last difference is that we include "#tail" as well as "#text", so that text outside of an element is preserved (rarely relevant for PMML, but included for completeness). Note that this method returns a JSON-like dictionary, not a string. To serialize to JSON, use the C{json} module from the Python Standard Library, a faster variant, or an exotic serializer such as BSON. @type data: string, dict, or file-like object @param data: The data to load. @type validate: bool @param validate: If True, validate the resulting PmmlBinding against this ModelLoader's XSD schema after loading. @type postValidate: bool @param postValidate: If True, run post-XSD validation checks. (Note: very few PmmlBinding subclasses have postValidation tests defined as of May 2013.) @param **parserOptions: Arguments passed to lxml's U{XMLParser<http://lxml.de/api/lxml.etree.XMLParser-class.html>}. @rtype: PmmlBinding @return: In-memory PMML object. @raise ValueError: If the JSON text is malformed or does not represent PMML, an error is raised. """ if hasattr(data, "read"): data = json.load(data) elif isinstance(data, basestring): if os.path.exists(data): data = json.load(open(data)) else: data = json.loads(data) if not isinstance(data, dict): raise ValueError("JSON object must be a mapping at the top level") if validate: if self.preparedSchema is None: self.preparedSchema = XMLSchema(self.schema) schema = self.preparedSchema else: schema = None parser = XMLParser(**parserOptions) lookup = ElementNamespaceClassLookup() namespace = lookup.get_namespace(defs.PMML_NAMESPACE) for xsdElement in self.schema.xpath( "xs:element", namespaces={"xs": defs.XSD_NAMESPACE}): namespace[xsdElement.attrib["name"]] = PmmlBinding namespace.update(self.tagToClass) parser.set_element_class_lookup(lookup) try: nsmap = data["#nsmap"] except KeyError: raise ValueError( "JSON object must have a \"#nsmap\" key at the top level") if "" in nsmap: nsmap[None] = nsmap[""] del nsmap[""] del data["#nsmap"] if len(data) != 1: raise ValueError( "JSON object must have exactly one PMML object at the top level" ) tag = data.keys()[0] data = data[tag] if not isinstance(data, list) or len(data) != 1: raise ValueError( "Top-level PMML object must be a list with exactly one item") data = data[0] pmmlBinding = self._loadJsonItem(tag, data, parser, nsmap) if validate: schema.assertValid(pmmlBinding) if postValidate: for event, elem in iterwalk(pmmlBinding, events=("end", ), tag="{%s}*" % defs.PMML_NAMESPACE): if isinstance(elem, PmmlBinding): elem.postValidate() return pmmlBinding
import logging import time from lxml import objectify from lxml.etree import XMLParser from suds import WebFault, TypeNotFound from suds.client import Client as sudsClient from suds.plugin import MessagePlugin from suds.xsd.query import TypeQuery, ElementQuery parser = XMLParser(remove_blank_text=True, huge_tree=True) parser.set_element_class_lookup(objectify.ObjectifyElementClassLookup()) objectify.set_default_parser(parser) logger = logging.getLogger('suds.client.lxml') logging.getLogger('suds.client').setLevel(logging.CRITICAL) # Don't show suds messages! class SoapObject: def __init__(self, name): self.__name__ = name def __len__(self): return len(self.__dict__.items()) - 1 # ignore the __name__ property def __repr__(self): return self.__str__() def __str__(self): return self.__name__
elif element.tag == "typeDescriptor": return nodes.NodeConfigTypeDesc elif element.tag == "flag": return nodes.NodeConfigFlag elif element.tag == "dependencyType": return nodes.NodeConfigDependencyType elif element.tag == "defaultType": return nodes.NodeConfigDefaultType elif element.tag == "type": return nodes.NodeConfigType else: raise TagNotFound(element) module_parser.set_element_class_lookup(_CommentLookup(_NodeClassLookup())) def _check_file(base_path, file_): """ Function used to search case-insensitively for a file/folder is a given path. :param base_path: The path to search for the file/folder in. :param file_: The file/folder to search for. :return: The file if found, raises an exception if not. """ base_file = file_ try: for item in listdir(base_path): if item.casefold() == base_file.casefold(): return item
import logging import time from lxml import objectify from lxml.etree import XMLParser from suds import WebFault, TypeNotFound from suds.client import Client as sudsClient from suds.plugin import MessagePlugin from suds.xsd.query import TypeQuery, ElementQuery parser = XMLParser(remove_blank_text=True, huge_tree=True) parser.set_element_class_lookup(objectify.ObjectifyElementClassLookup()) objectify.set_default_parser(parser) logger = logging.getLogger('suds.client.lxml') logging.getLogger('suds.client').setLevel( logging.CRITICAL) # Don't show suds messages! class SoapObject: def __init__(self, name): self.__name__ = name def __len__(self): return len(self.__dict__.items()) - 1 # ignore the __name__ property def __repr__(self): return self.__str__() def __str__(self): return self.__name__
def create_lxml_context(): parser = XMLParser(no_network=True) parser.set_element_class_lookup(ElementDefaultClassLookup(element=Element, comment=Comment)) return parser
ElementBase, parse as et_parse, fromstring as et_fromstring, XML as et_xml, XMLParser, ElementNamespaceClassLookup, ElementDefaultClassLookup, ) from dm.xmlsec.binding import DSigNs, dsig, EncNs, enc # set up our own parser and related `etree` infrastructure parser = XMLParser() # apparently, `parser` has a `set_element_class_lookup` but not corresponding `get` # class_lookup = ElementNamespaceClassLookup(parser.get_element_class_lookup()) class_lookup = ElementNamespaceClassLookup(ElementDefaultClassLookup()) parser.set_element_class_lookup(class_lookup) Element = parser.makeelement def SubElement(node, *args, **kw): node.append(Element(*args, **kw)) def parse(file, parser=parser): return et_parse(file, parser=parser) def fromstring(s, parser=parser): return et_fromstring(s, parser=parser)
def loadJson(self, data, validate=True, postValidate=True, **parserOptions): """Load a PMML model represented as a JSON string, fileName, dict, or file-like object. There is no standard XML-to-JSON specification, so we define our own. Our specification is very similar to U{this proposal<http://www.xml.com/pub/a/2006/05/31/converting-between-xml-and-json.html>}, which collects subelements of different tagnames into different JSON lists, rather than having one long list and needing to specify the tag of each element in that list. This has the following advantages, particularly useful for PMML: - Frequent tagnames (like <Segment>) are not repeated, wasting space. - Subelements with a given tagname can be quickly queried, without having to iterate over a list that contains non-matching tagnames. It has the following disadvantages: - The relative order of subelements with different tagnames is not preserved. We therefore additionally include a JSON attribute named "#" to specify the ordering of subelements in the XML representation. Also, the specification referenced above represents single-child subelements as JSON objects and multiple children as JSON lists, but for consistency and ease of parsing, we always use lists. The last difference is that we include "#tail" as well as "#text", so that text outside of an element is preserved (rarely relevant for PMML, but included for completeness). Note that this method returns a JSON-like dictionary, not a string. To serialize to JSON, use the C{json} module from the Python Standard Library, a faster variant, or an exotic serializer such as BSON. @type data: string, dict, or file-like object @param data: The data to load. @type validate: bool @param validate: If True, validate the resulting PmmlBinding against this ModelLoader's XSD schema after loading. @type postValidate: bool @param postValidate: If True, run post-XSD validation checks. (Note: very few PmmlBinding subclasses have postValidation tests defined as of May 2013.) @param **parserOptions: Arguments passed to lxml's U{XMLParser<http://lxml.de/api/lxml.etree.XMLParser-class.html>}. @rtype: PmmlBinding @return: In-memory PMML object. @raise ValueError: If the JSON text is malformed or does not represent PMML, an error is raised. """ if hasattr(data, "read"): data = json.load(data) elif isinstance(data, basestring): if os.path.exists(data): data = json.load(open(data)) else: data = json.loads(data) if not isinstance(data, dict): raise ValueError("JSON object must be a mapping at the top level") if validate: if self.preparedSchema is None: self.preparedSchema = XMLSchema(self.schema) schema = self.preparedSchema else: schema = None parser = XMLParser(**parserOptions) lookup = ElementNamespaceClassLookup() namespace = lookup.get_namespace(defs.PMML_NAMESPACE) for xsdElement in self.schema.xpath("xs:element", namespaces={"xs": defs.XSD_NAMESPACE}): namespace[xsdElement.attrib["name"]] = PmmlBinding namespace.update(self.tagToClass) parser.set_element_class_lookup(lookup) try: nsmap = data["#nsmap"] except KeyError: raise ValueError("JSON object must have a \"#nsmap\" key at the top level") if "" in nsmap: nsmap[None] = nsmap[""] del nsmap[""] del data["#nsmap"] if len(data) != 1: raise ValueError("JSON object must have exactly one PMML object at the top level") tag = data.keys()[0] data = data[tag] if not isinstance(data, list) or len(data) != 1: raise ValueError("Top-level PMML object must be a list with exactly one item") data = data[0] pmmlBinding = self._loadJsonItem(tag, data, parser, nsmap) if validate: schema.assertValid(pmmlBinding) if postValidate: for event, elem in iterwalk(pmmlBinding, events=("end",), tag="{%s}*" % defs.PMML_NAMESPACE): if isinstance(elem, PmmlBinding): elem.postValidate() return pmmlBinding