def _load(self, url):
doc = ET.parse(urllib2.urlopen(url))
extrans = doc.getroot().nsmap
targetns = doc.getroot().get('targetNamespace')
schemas = doc.findall(ns.expand('*/xs:schema'))
for s in schemas:
for k,v in extrans.items():
if k not in s.nsmap:
s.nsmap[k] = v
self.schemaloader.load(s, pathinfo=self.url)
self.documents.append(doc)
wsdls = doc.findall(ns.expand('/wsdl:import'))
for w in wsdls:
location = w.get('location')
location = urljoin(url, location)
tns = self._load(location)
if tns:
targetns = tns
return targetns
def _load(self, url):
doc = ET.parse(urllib2.urlopen(url))
extrans = doc.getroot().nsmap
targetns = doc.getroot().get('targetNamespace')
schemas = doc.findall(ns.expand('*/xs:schema'))
for s in schemas:
for k,v in extrans.items():
if k not in s.nsmap:
s.nsmap[k] = v
self.schemaloader.load(s, pathinfo=self.url)
self.documents.append(doc)
wsdls = doc.findall(ns.expand('/wsdl:import'))
for w in wsdls:
location = w.get('location')
location = urljoin(url, location)
tns = self._load(location)
if tns:
targetns = tns
return targetns
def load(self, schema, force=False, fragment=False, pathinfo='', basecls=None):
'''
Load and pre-process an XML schema.
@type schema: L{ElementTree.Element} or URL
@param schema: A schema to load.
@type force: bool
@param force: Optional. Reload an already loaded schema.
@type fragment: bool
@param fragment: Optional. The schema is actually a fragment of a
an already-loaded schema and should be integrated with it.
@type pathinfo: str
@param pathinfo: Optional. A URL to help with loading the schema.
Usually used by SchemaLoader to process xs:import directives.
@rtype: str
@return: The targetNamespace of the loaded schema
'''
# Is schema already parsed
if ET.iselement(schema):
root = schema
else:
schema = urljoin(pathinfo, schema)
root = ET.parse(u2.urlopen(schema)).getroot()
# Get the target namespace. Exit early if we already know this schema.
targetNamespace = root.get('targetNamespace')
if targetNamespace in self.schemas and not (force or fragment):
return targetNamespace
# Add a new entry to our dictionary
if not fragment:
self.schemas[targetNamespace] = { 'root': root, 'types': {}, 'elements': {}, 'groups': {}, 'validator': None, 'basecls': basecls }
# Update our "all namespaces" dictionary and get references to the
# various subdictionaries we'll need
self.allns.update(root.nsmap)
self.revns.update((v,k) for k,v in root.nsmap.items() if k not in (None, 'tns'))
types = self.schemas[targetNamespace]['types']
elements = self.schemas[targetNamespace]['elements']
groups = self.schemas[targetNamespace]['groups']
# Process includes
includes = []
while True:
# Get the list of includes
inclist = root.findall(ns.expand('xs:include'))
if not inclist:
break
for el in inclist:
# remove it from the document
root.remove(el)
# Get the schemaLocation and compute the URL
location = el.get('schemaLocation')
if location in includes:
# skip if we've processed this schema
continue
includes.append(location)
url = urljoin(pathinfo, location)
# Parse the XML and append it to the root document
# We probably *should* include it into the place where the
# xs:include node was, but for now, punt and append it
# to the end of the document
inc = ET.parse(u2.urlopen(url)).getroot()
root.extend(inc)
# Process imports
for el in root.findall(ns.expand('xs:import')):
location = el.get('schemaLocation')
if location:
self.load(location, pathinfo=pathinfo)
# Find all first-level tags we care about and reference them
# in the types/elements/groups dictionaries
for el in root.findall(ns.expand('xs:complexType')):
types[el.get('name')] = el
for el in root.findall(ns.expand('xs:simpleType')):
types[el.get('name')] = el
for el in root.findall(ns.expand('xs:element')):
elements[el.get('name')] = el
for el in root.findall(ns.expand('xs:group')):
groups[el.get('name')] = el
# If this is a schema fragment, integrate it into the
# original schema element tree in memory
if fragment:
realroot = self.schemas[targetNamespace]['root']
nsmap = dict(realroot.nsmap)
nsmap.update(root.nsmap)
attrib = dict(realroot.attrib)
attrib.update(root.attrib)
newroot = ET.Element(realroot.tag, attrib=attrib, nsmap=nsmap)
newroot.text = realroot.text
newroot.extend(realroot.getchildren())
newroot.extend(root.getchildren())
self.schemas[targetNamespace]['root'] = newroot
return targetNamespace
def invoke(self, operation, *args, **kwargs):
'''
Invoke a SOAP operation.
'''
self._reqno += 1
retxml = kwargs.pop('__retxml__', self.retxml)
timeout = kwargs.pop('__timeout__', self.timeout)
transport_options = kwargs.pop('__transport__', {})
# Create an instance of the request message and initialize
# the object from the arguments
param = self.factory(operation.imsg)
tmpl = param.__template__
for k,v in zip((t[0] for t in tmpl), args):
param[k] = v
for k,v in kwargs.items():
param[k] = v
# Build the soap envelope and set the http headers
payload = self.envelope(self.headers, param, operation.imsg)
httphdr = { 'Content-Type': 'text/xml', 'SOAPAction': operation.action }
httphdr.update(self.httphdr)
# Construct and issue the request, read the response
payload = ET.tostring(payload, pretty_print=True)
log.debug('=== SOAP REQUEST ===\n%s', re.sub(r'password>.*?<', r'password>*****<', payload ))
req = urllib2.Request(self.url, payload, httphdr)
try:
if self._inject:
xml = ET.fromstring(self._inject.next())
else:
if hasattr(self.transport, 'open'):
rsp = self.transport.open(req, timeout=timeout, **transport_options)
else:
rsp = self.transport.urlopen(req, timeout=timeout, **transport_options)
xml = ET.parse(rsp)
except urllib2.HTTPError as ex:
xml = ET.parse(ex)
log.debug('=== SOAP RESPONSE ===\n%s', xmlstr(xml))
# Get the soap body
retval = xml.find(ns.expand('soapenv:Body', self.nsmap))
if not retxml:
# Does the body contain any nodes?
if len(retval):
# Get the first child and examine it
retval = retval[0]
namespace, tag = ns.split(retval.tag)
# If it's a fault, convert it to an exception
if tag == 'Fault':
raise SoapFault(retval, self)
# Otherwise, deserialize
obj = self.factory(operation.omsg, retval)
# If the deserialized
# object has only one item, return that item, otherwise the
# whole object
#
# This is so if the return value is a single primitive type
# (like a string), you don't have to dig into an object just
# to get at the single primitive return value
if len(obj) == 1:
obj = obj[0]
retval = obj
else:
retval = None
return retval
def load(self,
schema,
force=False,
fragment=False,
pathinfo='',
basecls=None):
'''
Load and pre-process an XML schema.
@type schema: L{ElementTree.Element} or URL
@param schema: A schema to load.
@type force: bool
@param force: Optional. Reload an already loaded schema.
@type fragment: bool
@param fragment: Optional. The schema is actually a fragment of a
an already-loaded schema and should be integrated with it.
@type pathinfo: str
@param pathinfo: Optional. A URL to help with loading the schema.
Usually used by SchemaLoader to process xs:import directives.
@rtype: str
@return: The targetNamespace of the loaded schema
'''
# Is schema already parsed
if ET.iselement(schema):
root = schema
else:
schema = urljoin(pathinfo, schema)
root = ET.parse(u2.urlopen(schema)).getroot()
# Get the target namespace. Exit early if we already know this schema.
targetNamespace = root.get('targetNamespace')
if targetNamespace in self.schemas and not (force or fragment):
return targetNamespace
# Add a new entry to our dictionary
if not fragment:
self.schemas[targetNamespace] = {
'root': root,
'types': {},
'elements': {},
'groups': {},
'validator': None,
'basecls': basecls
}
# Update our "all namespaces" dictionary and get references to the
# various subdictionaries we'll need
self.allns.update(root.nsmap)
self.revns.update(
(v, k) for k, v in root.nsmap.items() if k not in (None, 'tns'))
types = self.schemas[targetNamespace]['types']
elements = self.schemas[targetNamespace]['elements']
groups = self.schemas[targetNamespace]['groups']
# Process includes
includes = []
while True:
# Get the list of includes
inclist = root.findall(ns.expand('xs:include'))
if not inclist:
break
for el in inclist:
# remove it from the document
root.remove(el)
# Get the schemaLocation and compute the URL
location = el.get('schemaLocation')
if location in includes:
# skip if we've processed this schema
continue
includes.append(location)
url = urljoin(pathinfo, location)
# Parse the XML and append it to the root document
# We probably *should* include it into the place where the
# xs:include node was, but for now, punt and append it
# to the end of the document
inc = ET.parse(u2.urlopen(url)).getroot()
root.extend(inc)
# Process imports
for el in root.findall(ns.expand('xs:import')):
location = el.get('schemaLocation')
if location:
self.load(location, pathinfo=pathinfo)
# Find all first-level tags we care about and reference them
# in the types/elements/groups dictionaries
for el in root.findall(ns.expand('xs:complexType')):
types[el.get('name')] = el
for el in root.findall(ns.expand('xs:simpleType')):
types[el.get('name')] = el
for el in root.findall(ns.expand('xs:element')):
elements[el.get('name')] = el
for el in root.findall(ns.expand('xs:group')):
groups[el.get('name')] = el
# If this is a schema fragment, integrate it into the
# original schema element tree in memory
if fragment:
realroot = self.schemas[targetNamespace]['root']
nsmap = dict(realroot.nsmap)
nsmap.update(root.nsmap)
attrib = dict(realroot.attrib)
attrib.update(root.attrib)
newroot = ET.Element(realroot.tag, attrib=attrib, nsmap=nsmap)
newroot.text = realroot.text
newroot.extend(realroot.getchildren())
newroot.extend(root.getchildren())
self.schemas[targetNamespace]['root'] = newroot
return targetNamespace
def invoke(self, operation, *args, **kwargs):
'''
Invoke a SOAP operation.
'''
self._reqno += 1
retxml = kwargs.pop('__retxml__', self.retxml)
timeout = kwargs.pop('__timeout__', self.timeout)
transport_options = kwargs.pop('__transport__', {})
# Create an instance of the request message and initialize
# the object from the arguments
param = self.factory(operation.imsg)
tmpl = param.__template__
for k,v in zip((t[0] for t in tmpl), args):
param[k] = v
for k,v in kwargs.items():
param[k] = v
# Build the soap envelope and set the http headers
payload = self.envelope(self.headers, param, operation.imsg)
httphdr = { 'Content-Type': 'text/xml', 'SOAPAction': operation.action }
httphdr.update(self.httphdr)
# Construct and issue the request, read the response
payload = ET.tostring(payload, pretty_print=True)
log.debug('=== SOAP REQUEST ===\n%s', payload)
req = urllib2.Request(self.url, payload, httphdr)
try:
if self._inject:
xml = ET.fromstring(self._inject.next())
else:
if hasattr(self.transport, 'open'):
rsp = self.transport.open(req, timeout=timeout, **transport_options)
else:
rsp = self.transport.urlopen(req, timeout=timeout, **transport_options)
xml = ET.parse(rsp)
except urllib2.HTTPError as ex:
xml = ET.parse(ex)
log.debug('=== SOAP RESPONSE ===\n%s', xmlstr(xml))
# Get the soap body
retval = xml.find(ns.expand('soapenv:Body', self.nsmap))
if not retxml:
# Does the body contain any nodes?
if len(retval):
# Get the first child and examine it
retval = retval[0]
namespace, tag = ns.split(retval.tag)
# If it's a fault, convert it to an exception
if tag == 'Fault':
raise SoapFault(retval, self)
# Otherwise, deserialize
obj = self.factory(operation.omsg, retval)
# If the deserialized
# object has only one item, return that item, otherwise the
# whole object
#
# This is so if the return value is a single primitive type
# (like a string), you don't have to dig into an object just
# to get at the single primitive return value
if len(obj) == 1:
obj = obj[0]
retval = obj
else:
retval = None
return retval
