Пример #1
0
 def FindLocalHREF(self, href, elt, headers=1):
     '''Find a local HREF in the data elements.
     '''
     if href[0] != '#':
         raise EvaluateException(
             'Absolute HREF ("%s") not implemented' % href,
             self.Backtrace(elt))
     frag = href[1:]
     # Already found?
     e = self.id_cache.get(frag)
     if e: return e
     # Do a breadth-first search, in the data first.  Most likely
     # to find multi-ref targets shallow in the data area.
     list = self.data_elements[:] + [self.body_root]
     if headers: list.extend(self.header_elements)
     while list:
         e = list.pop()
         if e.nodeType == _Node.ELEMENT_NODE:
             nodeid = _find_id(e)
             if nodeid:
                 self.id_cache[nodeid] = e
                 if nodeid == frag: return e
         list += _children(e)
     raise EvaluateException('''Can't find node for HREF "%s"''' % href,
                             self.Backtrace(elt))
Пример #2
0
 def parse(self, elt, ps):
     href = _find_href(elt)
     if href:
         if _children(elt):
             raise EvaluateException('Array has content and HREF',
                                     ps.Backtrace(elt))
         elt = ps.FindLocalHREF(href, elt)
     if self.nilled(elt, ps): return Nilled
     if not _find_arraytype(elt) and self.undeclared is False:
         raise EvaluateException('Array expected', ps.Backtrace(elt))
     t = _find_type(elt)
     if t:
         pass  # XXX should check the type, but parsing that is hairy.
     offset = self.parse_offset(elt, ps)
     v, vlen = [], 0
     if offset and not self.sparse:
         while vlen < offset:
             vlen += 1
             v.append(self.fill)
     for c in _child_elements(elt):
         item = self.ofwhat.parse(c, ps)
         position = self.parse_position(c, ps) or offset
         if self.sparse:
             v.append((position, item))
         else:
             while offset < position:
                 offset += 1
                 v.append(self.fill)
             v.append(item)
         offset += 1
     return v
Пример #3
0
 def parse(self, elt, ps):
     href = _find_href(elt)
     if href:
         if _children(elt):
             raise EvaluateException('Array has content and HREF',
                     ps.Backtrace(elt))
         elt = ps.FindLocalHREF(href, elt)
     if self.nilled(elt, ps): return Nilled
     if not _find_arraytype(elt) and self.undeclared is False:
         raise EvaluateException('Array expected', ps.Backtrace(elt))
     t = _find_type(elt)
     if t:
         pass # XXX should check the type, but parsing that is hairy.
     offset = self.parse_offset(elt, ps)
     v, vlen = [], 0
     if offset and not self.sparse:
         while vlen < offset:
             vlen += 1
             v.append(self.fill)
     for c in _child_elements(elt):
         item = self.ofwhat.parse(c, ps)
         position = self.parse_position(c, ps) or offset
         if self.sparse:
             v.append((position, item))
         else:
             while offset < position:
                 offset += 1
                 v.append(self.fill)
             v.append(item)
         offset += 1
     return v
Пример #4
0
 def FindLocalHREF(self, href, elt, headers=1):
     '''Find a local HREF in the data elements.
     '''
     if href[0] != '#':
         raise EvaluateException(
             'Absolute HREF ("%s") not implemented' % href,
             self.Backtrace(elt))
     frag = href[1:]
     # Already found?
     e = self.id_cache.get(frag)
     if e: return e
     # Do a breadth-first search, in the data first.  Most likely
     # to find multi-ref targets shallow in the data area.
     list = self.data_elements[:] + [self.body_root]
     if headers: list.extend(self.header_elements)
     while list:
         e = list.pop()
         if e.nodeType == _Node.ELEMENT_NODE:
             nodeid = _find_id(e)
             if nodeid:
                 self.id_cache[nodeid] = e
                 if nodeid == frag: return e
         list += _children(e)
     raise EvaluateException('''Can't find node for HREF "%s"''' % href,
             self.Backtrace(elt))
Пример #5
0
 def _check_for_pi_nodes(self, list, inheader):
     '''Raise an exception if any of the list descendants are PI nodes.
     '''
     list = list[:]
     while list:
         elt = list.pop()
         t = elt.nodeType
         if t == _Node.PROCESSING_INSTRUCTION_NODE:
             raise ParseException('Found processing instruction "<?' + \
                     elt.nodeName + '...>"',
                     inheader, elt.parentNode, self.dom)
         elif t == _Node.DOCUMENT_TYPE_NODE:
             raise ParseException('Found DTD', inheader,
                     elt.parentNode, self.dom)
         list += _children(elt)
Пример #6
0
 def _check_for_pi_nodes(self, list, inheader):
     '''Raise an exception if any of the list descendants are PI nodes.
     '''
     list = list[:]
     while list:
         elt = list.pop()
         t = elt.nodeType
         if t == _Node.PROCESSING_INSTRUCTION_NODE:
             raise ParseException('Found processing instruction "<?' + \
                     elt.nodeName + '...>"',
                     inheader, elt.parentNode, self.dom)
         elif t == _Node.DOCUMENT_TYPE_NODE:
             raise ParseException('Found DTD', inheader, elt.parentNode,
                                  self.dom)
         list += _children(elt)
Пример #7
0
 def _check_for_legal_children(self, name, elt, mustqualify=1):
     '''Check if all children of this node are elements or whitespace-only
     text nodes.
     '''
     inheader = name == "Header"
     for n in _children(elt):
         t = n.nodeType
         if t == _Node.COMMENT_NODE: continue
         if t != _Node.ELEMENT_NODE:
             if t == _Node.TEXT_NODE and n.nodeValue.strip() == "":
                 continue
             raise ParseException("Non-element child in " + name, 
                     inheader, elt, self.dom)
         if mustqualify and not n.namespaceURI:
             raise ParseException('Unqualified element "' + \
                     n.nodeName + '" in ' + name, inheader, elt, self.dom)
Пример #8
0
 def _check_for_legal_children(self, name, elt, mustqualify=1):
     '''Check if all children of this node are elements or whitespace-only
     text nodes.
     '''
     inheader = name == "Header"
     for n in _children(elt):
         t = n.nodeType
         if t == _Node.COMMENT_NODE: continue
         if t != _Node.ELEMENT_NODE:
             if t == _Node.TEXT_NODE and n.nodeValue.strip() == "":
                 continue
             raise ParseException("Non-element child in " + name, inheader,
                                  elt, self.dom)
         if mustqualify and not n.namespaceURI:
             raise ParseException('Unqualified element "' + \
                     n.nodeName + '" in ' + name, inheader, elt, self.dom)
Пример #9
0
    def parse(self, elt, ps):
        debug = self.logger.debugOn()
        debug and self.logger.debug('parse')

        xtype = self.checkname(elt, ps)
        if self.type and xtype not in [self.type, (None, None)]:
            if not isinstance(self, TypeDefinition):
                raise EvaluateException(\
                    'ComplexType for %s has wrong type(%s), looking for %s' %
                        (self.pname, self.checktype(elt,ps), self.type),
                                        ps.Backtrace(elt))
            else:
                #TODO: mabye change MRO to handle this
                debug and self.logger.debug('delegate to substitute type')
                what = TypeDefinition.getSubstituteType(self, elt, ps)
                return what.parse(elt, ps)

        href = _find_href(elt)
        if href:
            if _children(elt):
                raise EvaluateException('Struct has content and HREF',
                                        ps.Backtrace(elt))
            elt = ps.FindLocalHREF(href, elt)
        c = _child_elements(elt)
        count = len(c)
        if self.nilled(elt, ps): return Nilled

        # Create the object.
        v = {}

        # parse all attributes contained in attribute_typecode_dict (user-defined attributes),
        # the values (if not None) will be keyed in self.attributes dictionary.
        attributes = self.parse_attributes(elt, ps)
        if attributes:
            v[self.attrs_aname] = attributes

        #MIXED
        if self.mixed is True:
            v[self.mixed_aname] = self.simple_value(elt, ps, mixed=True)

        # Clone list of kids (we null it out as we process)
        c, crange = c[:], range(len(c))
        # Loop over all items we're expecting

        if debug:
            self.logger.debug("ofwhat: %s", str(self.ofwhat))

        any = None
        for i, what in [(i, self.ofwhat[i]) for i in range(len(self.ofwhat))]:

            # retrieve typecode if it is hidden
            if callable(what): what = what()

            # Loop over all available kids
            if debug:
                self.logger.debug("what: (%s,%s)", what.nspname, what.pname)

            for j, c_elt in [(j, c[j]) for j in crange if c[j]]:
                # Parse value, and mark this one done.
                if debug:
                    self.logger.debug("child node: (%s,%s)",
                                      c_elt.namespaceURI, c_elt.tagName)

                match = False
                if what.name_match(c_elt):
                    match = True
                    value = what.parse(c_elt, ps)
                else:
                    # substitutionGroup head must be a global element declaration
                    # if successful delegate to matching GED
                    subwhat = _get_substitute_element(what, c_elt, ps)
                    if subwhat:
                        match = True
                        value = subwhat.parse(c_elt, ps)

                    if debug:
                        self.logger.debug("substitutionGroup: %s", subwhat)

                if match:
                    if what.maxOccurs > 1:
                        if v.has_key(what.aname):
                            v[what.aname].append(value)
                        else:
                            v[what.aname] = [value]
                        c[j] = None
                        continue
                    else:
                        v[what.aname] = value
                    c[j] = None
                    break

                if debug:
                    self.logger.debug("no element (%s,%s)", what.nspname,
                                      what.pname)

                # No match; if it was supposed to be here, that's an error.
                if self.inorder is True and i == j:
                    raise EvaluateException('Out of order complexType',
                                            ps.Backtrace(c_elt))
            else:
                # only supporting 1 <any> declaration in content.
                if isinstance(what, AnyElement):
                    any = what
                elif hasattr(what, 'default'):
                    v[what.aname] = what.default
                elif what.minOccurs > 0 and not v.has_key(what.aname):
                    raise EvaluateException('Element "' + what.aname + \
                        '" missing from complexType', ps.Backtrace(elt))

        # Look for wildcards and unprocessed children
        # XXX Stick all this stuff in "any", hope for no collisions
        if any is not None:
            occurs = 0
            v[any.aname] = []
            for j, c_elt in [(j, c[j]) for j in crange if c[j]]:
                value = any.parse(c_elt, ps)
                if any.maxOccurs == UNBOUNDED or any.maxOccurs > 1:
                    v[any.aname].append(value)
                else:
                    v[any.aname] = value

                occurs += 1

            # No such thing as nillable <any>
            if any.maxOccurs == 1 and occurs == 0:
                v[any.aname] = None
            elif occurs < any.minOccurs or (any.maxOccurs != UNBOUNDED
                                            and any.maxOccurs < occurs):
                raise EvaluateException(
                    'occurances of <any> elements(#%d) bound by (%d,%s)' %
                    (occurs, any.minOccurs, str(any.maxOccurs)),
                    ps.Backtrace(elt))

        if not self.pyclass:
            return v

        # type definition must be informed of element tag (nspname,pname),
        # element declaration is initialized with a tag.
        try:
            pyobj = self.pyclass()
        except Exception, e:
            raise TypeError("Constructing element (%s,%s) with pyclass(%s), %s" \
                %(self.nspname, self.pname, self.pyclass.__name__, str(e)))
Пример #10
0
    def parse(self, elt, ps):
        debug = self.logger.debugOn()
        debug and self.logger.debug('parse')
        
        xtype = self.checkname(elt, ps)
        if self.type and xtype not in [ self.type, (None,None) ]:
            if not isinstance(self, TypeDefinition):
                raise EvaluateException(\
                    'ComplexType for %s has wrong type(%s), looking for %s' %
                        (self.pname, self.checktype(elt,ps), self.type), 
                                        ps.Backtrace(elt))
            else:
                #TODO: mabye change MRO to handle this 
                debug and self.logger.debug('delegate to substitute type')
                what = TypeDefinition.getSubstituteType(self, elt, ps)
                return what.parse(elt, ps)
            
        href = _find_href(elt)
        if href:
            if _children(elt):
                raise EvaluateException('Struct has content and HREF',
                        ps.Backtrace(elt))
            elt = ps.FindLocalHREF(href, elt)
        c = _child_elements(elt)
        count = len(c)
        if self.nilled(elt, ps): return Nilled

        # Create the object.
        v = {}

        # parse all attributes contained in attribute_typecode_dict (user-defined attributes),
        # the values (if not None) will be keyed in self.attributes dictionary.
        attributes = self.parse_attributes(elt, ps)
        if attributes:
            v[self.attrs_aname] = attributes

        #MIXED
        if self.mixed is True:
            v[self.mixed_aname] = self.simple_value(elt,ps, mixed=True)

        # Clone list of kids (we null it out as we process)
        c, crange = c[:], range(len(c))
        # Loop over all items we're expecting
        
        if debug:
            self.logger.debug("ofwhat: %s",str(self.ofwhat))
            
        any = None
        for i,what in [ (i, self.ofwhat[i]) for i in range(len(self.ofwhat)) ]:
            
            # retrieve typecode if it is hidden
            if callable(what): what = what()
            
            # Loop over all available kids
            if debug: 
                self.logger.debug("what: (%s,%s)", what.nspname, what.pname)
                
            for j,c_elt in [ (j, c[j]) for j in crange if c[j] ]:
                # Parse value, and mark this one done. 
                if debug:
                    self.logger.debug("child node: (%s,%s)", c_elt.namespaceURI, c_elt.tagName)

                match = False
                if what.name_match(c_elt):
                    match = True
                    value = what.parse(c_elt, ps)
                else:
                    # substitutionGroup head must be a global element declaration
                    # if successful delegate to matching GED
                    subwhat = _get_substitute_element(what, c_elt, ps)
                    if subwhat:
                        match = True
                        value = subwhat.parse(c_elt, ps)

                    if debug: 
                        self.logger.debug("substitutionGroup: %s", subwhat)

                if match:
                    if what.maxOccurs > 1:
                        if v.has_key(what.aname):
                            v[what.aname].append(value)
                        else:
                            v[what.aname] = [value]
                        c[j] = None
                        continue
                    else:
                        v[what.aname] = value
                    c[j] = None
                    break

                if debug:
                    self.logger.debug("no element (%s,%s)", what.nspname, what.pname)

                # No match; if it was supposed to be here, that's an error.
                if self.inorder is True and i == j:
                    raise EvaluateException('Out of order complexType',
                            ps.Backtrace(c_elt))
            else:
                # only supporting 1 <any> declaration in content.
                if isinstance(what,AnyElement):
                    any = what
                elif hasattr(what, 'default'):
                    v[what.aname] = what.default
                elif what.minOccurs > 0 and not v.has_key(what.aname):
                    raise EvaluateException('Element "' + what.aname + \
                        '" missing from complexType', ps.Backtrace(elt))

        # Look for wildcards and unprocessed children
        # XXX Stick all this stuff in "any", hope for no collisions
        if any is not None:
            occurs = 0
            v[any.aname] = []
            for j,c_elt in [ (j, c[j]) for j in crange if c[j] ]:
                value = any.parse(c_elt, ps)
                if any.maxOccurs == UNBOUNDED or any.maxOccurs > 1:
                    v[any.aname].append(value)
                else:
                    v[any.aname] = value

                occurs += 1

            # No such thing as nillable <any>
            if any.maxOccurs == 1 and occurs == 0:
                v[any.aname] = None
            elif occurs < any.minOccurs or (any.maxOccurs!=UNBOUNDED and any.maxOccurs<occurs):
                raise EvaluateException('occurances of <any> elements(#%d) bound by (%d,%s)' %(
                    occurs, any.minOccurs,str(any.maxOccurs)), ps.Backtrace(elt))

        if not self.pyclass: 
            return v

        # type definition must be informed of element tag (nspname,pname),
        # element declaration is initialized with a tag.
        try:
            pyobj = self.pyclass()
        except Exception, e:
            raise TypeError("Constructing element (%s,%s) with pyclass(%s), %s" \
                %(self.nspname, self.pname, self.pyclass.__name__, str(e)))
Пример #11
0
class ParsedSoap:
    '''A Parsed SOAP object.
        Convert the text to a DOM tree and parse SOAP elements.
        Instance data:
            reader -- the DOM reader
            dom -- the DOM object
            ns_cache -- dictionary (by id(node)) of namespace dictionaries
            id_cache -- dictionary (by XML ID attr) of elements
            envelope -- the node holding the SOAP Envelope
            header -- the node holding the SOAP Header (or None)
            body -- the node holding the SOAP Body
            body_root -- the serialization root in the SOAP Body
            data_elements -- list of non-root elements in the SOAP Body
            trailer_elements -- list of elements following the SOAP body
    '''
    defaultReaderClass = DefaultReader

    def __init__(self,
                 input,
                 readerclass=None,
                 keepdom=False,
                 trailers=False,
                 resolver=None,
                 envelope=True,
                 **kw):
        '''Initialize.
        Keyword arguments:
            trailers -- allow trailer elments (default is zero)
            resolver -- function (bound method) to resolve URI's
            readerclass -- factory class to create a reader
            keepdom -- do not release the DOM
            envelope -- look for a SOAP envelope.
        '''

        self.readerclass = readerclass
        self.keepdom = keepdom
        if not self.readerclass:
            self.readerclass = self.defaultReaderClass

        try:
            self.reader = self.readerclass()
            if type(input) in _stringtypes:
                self.dom = self.reader.fromString(input)
            else:
                self.dom = self.reader.fromStream(input)
        except Exception, e:
            # Is this in the header?  Your guess is as good as mine.
            #raise ParseException("Can't parse document (" + \
            #    str(e.__class__) + "): " + str(e), 0)
            raise

        self.ns_cache = {
            id(self.dom): {
                'xml': XMLNS.XML,
                'xmlns': XMLNS.BASE,
                '': ''
            }
        }
        self.trailers, self.resolver, self.id_cache = trailers, resolver, {}

        # Exactly one child element
        c = [
            E for E in _children(self.dom) if E.nodeType == _Node.ELEMENT_NODE
        ]
        if len(c) == 0:
            raise ParseException("Document has no Envelope", 0)
        if len(c) != 1:
            raise ParseException("Document has extra child elements", 0)

        if envelope is False:
            self.body_root = c[0]
            return

        # And that one child must be the Envelope
        elt = c[0]
        if elt.localName != "Envelope" \
        or elt.namespaceURI != SOAP.ENV:
            raise ParseException('Document has "' + elt.localName + \
                '" element, not Envelope', 0)
        self._check_for_legal_children("Envelope", elt)
        for a in _attrs(elt):
            name = a.nodeName
            if name.find(":") == -1 and name not in ["xmlns", "id"]:
                raise ParseException('Unqualified attribute "' + \
                        name + '" in Envelope', 0)
        self.envelope = elt
        if not _valid_encoding(self.envelope):
            raise ParseException("Envelope has invalid encoding", 0)

        # Get Envelope's child elements.
        c = [
            E for E in _children(self.envelope)
            if E.nodeType == _Node.ELEMENT_NODE
        ]
        if len(c) == 0:
            raise ParseException("Envelope is empty (no Body)", 0)

        # Envelope's first child might be the header; if so, nip it off.
        elt = c[0]
        if elt.localName == "Header" \
        and elt.namespaceURI == SOAP.ENV:
            self._check_for_legal_children("Header", elt)
            self._check_for_pi_nodes(_children(elt), 1)
            self.header = c.pop(0)
            self.header_elements = _child_elements(self.header)
        else:
            self.header, self.header_elements = None, []

        # Now the first child must be the body
        if len(c) == 0:
            raise ParseException("Envelope has header but no Body", 0)
        elt = c.pop(0)
        if elt.localName != "Body" \
        or elt.namespaceURI != SOAP.ENV:
            if self.header:
                raise ParseException('Header followed by "' + \
                        elt.localName + \
                        '" element, not Body', 0, elt, self.dom)
            else:
                raise ParseException('Document has "' + \
                        elt.localName + \
                        '" element, not Body', 0, elt, self.dom)
        self._check_for_legal_children("Body", elt, 0)
        self._check_for_pi_nodes(_children(elt), 0)
        self.body = elt
        if not _valid_encoding(self.body):
            raise ParseException("Body has invalid encoding", 0)

        # Trailer elements.
        if not self.trailers:
            if len(c):
                raise ParseException("Element found after Body", 0, elt,
                                     self.dom)
            # Don't set self.trailer_elements = []; if user didn't ask
            # for trailers we *want* to throw an exception.
        else:
            self.trailer_elements = c
            for elt in self.trailer_elements:
                if not elt.namespaceURI:
                    raise ParseException('Unqualified trailer element', 0, elt,
                                         self.dom)

        # Find the serialization root.  Divide the Body children into
        # root (root=1), no (root=0), maybe (no root attribute).
        self.body_root, no, maybe = None, [], []
        for elt in _child_elements(self.body):
            root = _find_root(elt)
            if root == "1":
                if self.body_root:
                    raise ParseException("Multiple seralization roots found",
                                         0, elt, self.dom)
                self.body_root = elt
            elif root == "0":
                no.append(elt)
            elif not root:
                maybe.append(elt)
            else:
                raise ParseException('Illegal value for root attribute', 0,
                                     elt, self.dom)

        # If we didn't find a root, get the first one that didn't
        # say "not me", unless they all said "not me."
        if self.body_root is None:
            if len(maybe):
                self.body_root = maybe[0]
            else:
                raise ParseException('No serialization root found', 0,
                                     self.body, self.dom)
        if not _valid_encoding(self.body_root):
            raise ParseException("Invalid encoding", 0, elt, self.dom)

        # Now get all the non-roots (in order!).
        rootid = id(self.body_root)
        self.data_elements = [
            E for E in _child_elements(self.body) if id(E) != rootid
        ]
        self._check_for_pi_nodes(self.data_elements, 0)