path: root/aw_api/SoapBuffer.py

#!/usr/bin/python
#
# Copyright 2009 Google Inc.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#      http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
#

"""Handler class for implementing a SOAP buffer."""

__author__ = 'api.sgrinberg@gmail.com (Stan Grinberg)'

import re
import sys
from aw_api import ETREE
from aw_api import ETREE_NAME
from aw_api import MIN_ETREE_VERSION
from aw_api import MIN_PY_VERSION
from aw_api import MIN_PYXML_VERSION
from aw_api import PYXML
from aw_api import PYXML_NAME
from aw_api import Utils
from aw_api.Buffer import Buffer
from aw_api.Errors import InvalidInputError
from aw_api.Errors import MalformedBufferError
from aw_api.Errors import MissingPackageError
# Is this running on Google's App Engine?
try:
  import google.appengine
  from _xmlplus.parsers.expat import ExpatError
except:
  from xml.parsers.expat import ExpatError

# Check if PyXML is installed.
try:
  # Is this running on Google's App Engine?
  try:
    import google.appengine
    import _xmlplus
  except:
    import xml as _xmlplus
  from xml.dom import minidom
  # Test whether xml.version_info exists.
  version = _xmlplus.__dict__.get('version_info')
  PYXML_LIB = True
except (ImportError, AttributeError):
  # An anti-anti-hack/workaround for users of python2.5 (and up) to convince
  # Python not to ignore PyXML. We really do want to use _xmlplus here, even
  # though it is 'too old'.
  if ((list(map(eval, MIN_PY_VERSION.split('.'))) <
       list(sys.version_info)[0:3])):
    # The change to the environment variables is only active during the life of
    # a process.
    import os
    os.environ['PY_USE_XMLPLUS'] = 'YES'
    PYXML_LIB = True
  else:
    PYXML_LIB = False
else:
  if (not version or
      list(version) < (list(map(eval, MIN_PYXML_VERSION.split('.'))))):
    PYXML_LIB = False

# Check if ElementTree is installed.
try:
  import elementtree.ElementTree as etree
  etree.VERSION
  ETREE_LIB = True
except (ImportError, AttributeError):
  ETREE_LIB = False

if not PYXML and not ETREE:
  msg = ('PyXML v%s or ElementTree v%s or newer is required.'
         % (MIN_PYXML_VERSION, MIN_ETREE_VERSION))
  raise MissingPackageError(msg)


class SoapBuffer(Buffer):

  """Implements a SoapBuffer.

  Catches and parses outgoing and incoming SOAP XML messages.
  """

  def __init__(self, xml_parser=None, pretty_xml=False):
    """Inits SoapBuffer.

    Args:
      xml_parser: str XML parser to use.
      pretty_xml: bool whether to use pretty xml.
    """
    self.__buffer = ''
    self.__dump = {}
    self.__xml_parser = xml_parser
    # Pick a default XML parser, if none was set.
    if not self.__xml_parser:
      for parser, status in [(PYXML, PYXML_LIB), (ETREE, ETREE_LIB)]:
        if status:
          self.__xml_parser = parser
    # Validate current state of the chosen XML parser.
    if self.__xml_parser == PYXML and not PYXML_LIB:
      msg = 'PyXML v%s or newer is required.' % MIN_PYXML_VERSION
      raise MissingPackageError(msg)
    elif self.__xml_parser == ETREE and not ETREE_LIB:
      msg = 'ElementTree v%s or newer is required.' % MIN_ETREE_VERSION
      raise MissingPackageError(msg)
    elif self.__xml_parser != PYXML and self.__xml_parser != ETREE:
      msg = ('Invalid input for XML parser, expecting one of %s.'
             % sorted(set(PYXML + ETREE)))
      raise InvalidInputError(msg)
    # Set XML parser signature.
    if self.__xml_parser == PYXML:
      self.__xml_parser_sig = '%s v%s' % (PYXML_NAME, MIN_PYXML_VERSION)
    elif self.__xml_parser == ETREE:
      self.__xml_parser_sig = '%s v%s' % (ETREE_NAME, MIN_ETREE_VERSION)
    self.__pretty_xml = pretty_xml

  def write(self, str_in):
    """Append given string to a buffer.

    Args:
      str_in string to append to a buffer.
    """
    self.__buffer += str_in

  def flush(self):
    pass

  def GetBufferAsStr(self):
    """Return buffer as string.

    Returns:
      str buffer.
    """
    return str(self.__buffer)

  def IsHandshakeComplete(self):
    """Return state of the handshake.

    A successful handshake occurs when a SOAP request is sent and a SOAP
    response is recieved.

    Returns:
      bool True if successful handshake, False otherwise.
    """
    if (self.GetHeadersOut() and self.GetSOAPOut() and self.GetHeadersIn() and
        self.GetSOAPIn() and not Utils.IsHtml(self.GetSOAPIn())):
      return True
    return False

  def __GetBufferAsDict(self):
    """Parse HTTP headers and SOAP data.

    Returns:
      dict request's HTTP headers and SOAP data.
    """
    tags = (('Outgoing HTTP headers', 'dumpHeadersOut'),
            ('Outgoing SOAP', 'dumpSOAPOut'),
            ('Incoming HTTP headers', 'dumpHeadersIn'),
            ('Incoming SOAP', 'dumpSOAPIn'))
    xml_dumps = self.GetBufferAsStr().split('_' * 33)
    if len(xml_dumps) > 1:
      # The HTTP and SOAP messages were delivered via ZSI.
      for xml_part in xml_dumps:
        xml_part = xml_part.lstrip('\n').rstrip('\n')
        if not xml_part:
          continue
        xml_part = '\n'.join(xml_part.split('\n')[1:])
        in_parts = xml_part.split('\n\n', 1)
        if len(in_parts) == 1:
          from ZSI.version import Version
          xml_out = '<?xml version="1.0" encoding="UTF-8"?>\n%s' % xml_part
          if self.__pretty_xml:
            xml_out = self.__PrettyPrintXml(xml_out, 1)
          self.__dump['dumpHeadersOut'] = (
              '%s Outgoing HTTP headers %s\n'
              'User-Agent: ZSI %s (http://pywebsvcs.sf.net); %s\n'
              '%s' % ('*' * 3, '*' * 46, '.'.join(map(str, Version)),
                      self.__xml_parser_sig, '*' * 72))
          self.__dump['dumpSOAPOut'] = (
              '%s Outgoing SOAP %s\n%s\n%s'
              % ('*' * 3, '*' * 54, xml_out, '*' * 72))
        elif len(in_parts) == 2:
          sub_parts = in_parts[0].split('-' * 7)
          xml_in = ('<?xml version="1.0" encoding="UTF-8"?>\n%s'
                    % in_parts[1].lstrip('\n'))
          if self.__pretty_xml:
            xml_in = self.__PrettyPrintXml(xml_in, 1)
          self.__dump['dumpHeadersIn'] = (
              '%s Incoming HTTP headers %s\n'
              '%s\n%s' % ('*' * 3, '*' * 46,
                          (' '.join(sub_parts[0].split('\n')) +
                           sub_parts[1]).replace('\r', ''), '*' * 72))
          self.__dump['dumpSOAPIn'] = (
              '%s Incoming SOAP %s\n%s\n%s'
              % ('*' * 3, '*' * 54, xml_in, '*' * 72))
        else:
          pass
    else:
      # The HTTP and SOAP messages were delivered via SOAPpy or httplib.HTTPS.
      xml_dumps = self.GetBufferAsStr().split('*' * 72)
      for xml_part in xml_dumps:
        xml_part = xml_part.lstrip('\n').rstrip('\n')
        for name, tag in tags:
          if xml_part.find(name) > -1:
            # Insert XML parser signature into the SOAP header.
            trigger = xml_part[xml_part.lower().find('content-type'):
                               xml_part.lower().find('content-type')+12]
            if trigger and tag == 'dumpHeadersOut':
              xml_part = xml_part.replace(
                  trigger,
                  'XML-parser: %s\n%s' % (self.__xml_parser_sig, trigger))
            if self.__pretty_xml:
              doc = []
              banner = ''
              for line in xml_part.split('\n'):
                if line.rfind('SOAP %s' % ('*' * 46)) > -1:
                  banner = line
                  continue
                doc.append(line)
              if banner:
                xml_part = '%s\n%s' % (banner,
                                       self.__PrettyPrintXml('\n'.join(doc), 1))
            self.__dump[tag] = (xml_part + '\n' + '*' * 72)
            break
    return self.__dump

  def __GetDumpValue(self, dump_type):
    """Return dump value given its type.

    Args:
      dump_type: str type of the dump.

    Returns:
      str value of the dump.
    """
    dump_value = None
    try:
      if dump_type in self.__dump:
        dump_value = self.__dump[dump_type]
      else:
        dump_value = self.__GetBufferAsDict()[dump_type]
    except KeyError:
      dump_value = ''
    return dump_value

  def GetHeadersOut(self):
    """Return outgoing headers dump.

    Returns:
      str outgoing headers dump.
    """
    return self.__GetDumpValue('dumpHeadersOut')

  def GetSOAPOut(self):
    """Return SOAP out dump.

    Returns:
      str outgoing SOAP dump.
    """
    dump_value = self.__GetDumpValue('dumpSOAPOut')

    # Mask out login's password, and authToken.
    dump_value = dump_value.replace('><', '>\n<')
    for mask in ['password', 'authToken', 'ns1:authToken']:
      pattern = re.compile('>.*?</%s>' % mask)
      dump_value = pattern.sub('>xxxxxx</%s>' % mask, dump_value)
    return dump_value

  def GetHeadersIn(self):
    """Return incoming headers dump.

    Returns:
      str incoming headers dump.
    """
    return self.__GetDumpValue('dumpHeadersIn')

  def GetSOAPIn(self):
    """Return incoming SOAP dump.

    Returns:
      str incoming SOAP dump.
    """
    return self.__GetDumpValue('dumpSOAPIn')

  def GetRawSOAPIn(self):
    """Return raw incoming SOAP dump with out banners and not prettified.

    Returns:
      str raw incoming SOAP dump.
    """
    doc = ''.join(self.GetSOAPIn().split('\n')[1:-1])
    return self.__PrettyPrintXml(doc, -1)

  def __GetXmlOut(self):
    """Remove banners from outgoing SOAP XML and contstruct XML object.

    Returns:
      Document/Element object generated from string, representing XML message.
    """
    # Remove banners.
    xml_dump = self.GetSOAPOut().lstrip('\n').rstrip('\n')
    xml_parts = xml_dump.split('\n')

    # While multiple threads are used, SoapBuffer gets too greedy and tries to
    # capture all traffic that goes to sys.stdout. Part that we don't need
    # should be redirected back to sys.stdout.
    banner = '%s Outgoing SOAP %s' % ('*' * 3, '*' * 54)
    begin = 1
    for index in range(len(xml_parts)):
      if xml_parts[index] == banner:
        begin = index + 1
        break
    non_xml = '\n'.join(xml_parts[:begin-1])
    # Send data we don't need back to sys.stdout.
    if non_xml:
      print non_xml

    xml_dump = '\n'.join(xml_parts[begin:len(xml_parts)-1])

    try:
      if self.__xml_parser == PYXML:
        xml_obj = minidom.parseString(xml_dump)
      elif self.__xml_parser == ETREE:
        xml_obj = etree.fromstring(xml_dump)
    except ExpatError, e:
      msg = 'Unable to parse SOAP buffer for outgoing messages.' % e
      raise MalformedBufferError(msg)
    return xml_obj

  def __GetXmlIn(self):
    """Remove banners from incoming SOAP XML and construct XML object.

    Returns:
      Document/Element object generated from string, representing XML message.
    """
    # Remove banners.
    xml_dump = self.GetSOAPIn().lstrip('\n').rstrip('\n')
    xml_parts = xml_dump.split('\n')
    xml_dump = '\n'.join(xml_parts[1:len(xml_parts)-1])

    try:
      if self.__xml_parser == PYXML:
        xml_obj = minidom.parseString(xml_dump)
      elif self.__xml_parser == ETREE:
        xml_obj = etree.fromstring(xml_dump)
    except ExpatError, e:
      msg = 'Unable to parse SOAP buffer for incoming messages.' % e
      raise MalformedBufferError(msg)
    return xml_obj

  def __RemoveElemAttr(self, elem):
    """Remove element's attribute, if any.

    Args:
      elem: str XML element with or without an attribute.

    Returns:
      str XML element without attribute.
    """
    pattern = re.compile('({.*?})')
    return re.sub(pattern, '', elem)

  def __GetXmlNameByName(self, xml_obj, name):
    """Get XML object name from a given tag name.

    Args:
      xml_obj: Document/Element object.
      name: str tag name to look up.

    Returns:
      str XML object name, or None if tag name is not found in the XML object.
    """
    value = None
    if self.__xml_parser == PYXML:
      for node in xml_obj.getElementsByTagName(name):
        if len(node.childNodes) == 1:
          value = node.childNodes[0].nodeName
        else:
          value = node.childNodes[1].nodeName
    elif self.__xml_parser == ETREE:
      root = xml_obj
      etags = etags = name.split('/')
      for etag in etags:
        for item in root.getchildren():
          if item.tag.find(etag) > -1 or etag == '*':
            root = item
            if etag == etags[-1]:
              value = self.__RemoveElemAttr(root.tag)
    return value

  def __GetXmlValueByName(self, xml_obj, name, get_all=False):
    """Get XML object value from a given tag name.

    Args:
      xml_obj: Document/Element object.
      name: str tag name whose value to look up.
      get_all: bool whether to return all values that were found or just one.

    Returns:
      str XML object value, list if more than one value is found and if
      explicitly requested, or None if name is not found in the XML object.
    """
    values = []
    if self.__xml_parser == PYXML:
      for node in xml_obj.getElementsByTagName(name):
        values.append(node.childNodes[0].nodeValue)
    elif self.__xml_parser == ETREE:
      root = xml_obj
      etags = name.split('/')
      for etag in etags:
        for item in root.getchildren():
          if item.tag.find(etag) > -1:
            root = item
            if etag == etags[-1]:
              values.append(root.text)
    if get_all and values:
      return values
    elif values:
      return values[0]
    return None

  def IsSoap(self):
    """Whether buffer contains SOAP message(s) or not.

    Returns:
      bool True if message is a SOAP, False otherwise.
    """
    data = self.GetBufferAsStr()
    if not data:
      return False
    try:
      if self.__xml_parser == PYXML:
        if (self.GetCallResponseTime() is None and
            not self.__GetXmlIn().getElementsByTagName('soapenv:Body') and
            not self.__GetXmlIn().getElementsByTagName('soap:Body')):
          return False
      elif self.__xml_parser == ETREE:
        if self.__GetXmlNameByName(self.__GetXmlIn(), 'Body') is None:
          return False
    except Exception:
      # Data contains malformed XML message.
      return False
    return True

  def GetServiceName(self):
    """Get name of the API service that was called.

    Note: the way service is determined depends on the particular order of
    elements in the caller stack (e.g., CampaignService.Get() ->
    WebService.__ManageSoap() -> SoapBuffer.GetServiceName().)

    Returns:
      str name of the API service that was called.
    """
    return sys._getframe(3).f_locals['self'].__class__.__name__

  def GetCallName(self):
    """Get name of the API method that was called.

    Returns:
      str name of the API method that was called.
    """
    if self.__xml_parser == PYXML:
      # Remove "nsX:" if exists.
      pattern = re.compile('ns.*?:')
      value = pattern.sub('', self.__GetXmlNameByName(self.__GetXmlOut(),
                                                      'SOAP-ENV:Body'))
    elif self.__xml_parser == ETREE:
      value = self.__GetXmlNameByName(self.__GetXmlOut(), 'Body/*')
    return value

  def GetOperatorName(self):
    """Get name of the operator that was used in the API call.

    Returns:
      dict a dictionary consisting of the name of the operator mapped to the
           number of times that operator was used in the API request.
    """
    if self.__xml_parser == PYXML:
      values = self.__GetXmlValueByName(self.__GetXmlOut(),
                                        'ns1:operator',
                                        True)
    elif self.__xml_parser == ETREE:
      values = self.__GetXmlValueByName(self.__GetXmlOut(),
                                        'Body/mutate/operations/operator',
                                        True)
    # If no operator found, returns None. Otherwise, the format
    # is {'ADD': 1, 'SET': 2}.
    operator = {}
    if not values:
      operator = None
    else:
      for value in values:
        if value in operator:
          operator[str(value)] += 1
        else:
          operator[str(value)] = 1
    return operator

  def GetCallResponseTime(self):
    """Get value for responseTime header.

    Returns:
      str responseTime header value.
    """
    for name in ['responseTime', 'Header/ResponseHeader/responseTime',
                 'ns2:responseTime']:
      value = self.__GetXmlValueByName(self.__GetXmlIn(), name)
      if value:
        return value
    return None

  def GetCallOperations(self):
    """Get value for operations header.

    Returns:
      str operations header value.
    """
    for name in ['operations', 'Header/ResponseHeader/operations',
                 'ns2:operations']:
      value = self.__GetXmlValueByName(self.__GetXmlIn(), name)
      if value:
        return value
    return None

  def GetCallUnits(self):
    """Get value for units header.

    Returns:
      str units header value.
    """
    for name in ['units', 'Header/ResponseHeader/units', 'ns2:units']:
      value = self.__GetXmlValueByName(self.__GetXmlIn(), name)
      if value:
        return value
    return None

  def GetCallRequestId(self):
    """Get value for requestId header.

    Returns:
      str requestId header value.
    """
    for name in ['requestId', 'Header/ResponseHeader/requestId',
                 'ns2:requestId']:
      value = self.__GetXmlValueByName(self.__GetXmlIn(), name)
      if value:
        return value
    return None

  def GetFaultAsDict(self):
    """Parse incoming SOAP fault to load all elements into a dictionary.

    Returns:
      dict dictionary object with all fault elements.
    """
    elems = ['soapenv:Fault', 'soap:Fault']
    for elem in elems:
      xml_obj = self.__GetXmlIn()
      if self.__xml_parser == PYXML:
        try:
          fault_obj = xml_obj.getElementsByTagName(elem)[0]
        except:
          envelope = xml_obj.childNodes[0]
          header = None
          body = None
          for element in envelope.childNodes:
            if element.nodeName.lower().find('header') > -1:
              header = element.childNodes[1]
            elif element.nodeName.lower().find('body') > -1:
              body = element.childNodes[1]
          fault_obj = body
      elif self.__xml_parser == ETREE:
        fault_obj = xml_obj.getchildren()[len(
            xml_obj.getchildren())-1].getchildren()[0]

      fault = {}
      detail = {}
      errors = []
      nodes = {PYXML: ['childNodes', 'localName', 'childNodes[0].nodeValue'],
               ETREE: ['getchildren()', 'tag', 'text']}
      # Step through the Document or Element (depending on which XML parser is
      # used) and construct a dictionary representation of the fault. See
      # different examples of the API fault at
      #   - data/response_fault_stacktrace.xml
      #   - data/response_fault.xml
      #   - data/response_fault_errors.xml
      for fitem in eval('fault_obj.%s' % nodes[self.__xml_parser][0]):
        if ((self.__xml_parser == PYXML and fitem.hasChildNodes()) or
            self.__xml_parser == ETREE):
          ftag = self.__RemoveElemAttr(eval('fitem.%s'
                                            % nodes[self.__xml_parser][1]))
          if ftag != 'detail':
            fault[ftag] = eval('fitem.%s' % nodes[self.__xml_parser][2])
          else:
            for ditem in eval('fitem.%s' % nodes[self.__xml_parser][0]):
              if ((self.__xml_parser == PYXML and ditem.hasChildNodes()) or
                  self.__xml_parser == ETREE):
                dtag = self.__RemoveElemAttr(
                    eval('ditem.%s' % nodes[self.__xml_parser][1]))
                if dtag != 'fault' and dtag != 'ApiExceptionFault':
                  detail[dtag] = eval('ditem.%s' % nodes[self.__xml_parser][2])
                else:
                  for dfitem in eval('ditem.%s' % nodes[self.__xml_parser][0]):
                    if ((self.__xml_parser == PYXML
                         and dfitem.hasChildNodes()) or
                        self.__xml_parser == ETREE):
                      dftag = self.__RemoveElemAttr(
                          eval('dfitem.%s' % nodes[self.__xml_parser][1]))
                      if dftag != 'errors':
                        detail[dftag] = eval('dfitem.%s'
                                             % nodes[self.__xml_parser][2])
                      else:
                        error = {}
                        for eitem in eval('dfitem.%s'
                                          % nodes[self.__xml_parser][0]):
                          if ((self.__xml_parser == PYXML
                               and eitem.hasChildNodes()) or
                              self.__xml_parser == ETREE):
                            etag = self.__RemoveElemAttr(
                                eval('eitem.%s' % nodes[self.__xml_parser][1]))
                            # Rename type elements that have a dot in them into
                            # something more useful,
                            #   - ApiError.Type => type
                            if etag.find('.') > -1:
                              etag = etag.split('.')[1].lower()
                            error[etag] = eval(
                                'eitem.%s' % nodes[self.__xml_parser][2])
                        errors.append(error)
                        detail[dftag] = errors
            fault[ftag] = detail
      return fault

  def GetFaultAsStr(self):
    """Format all fault elements into a printable form.

    Returns:
      str string with all fault elements.
    """
    try:
      fault = self.GetFaultAsDict()
      if not fault:
        return ''
    except Exception:
      return ''

    items = []
    for fitem in fault:
      if fitem != 'detail':
        items.append('%s: %s\n' % (fitem, fault[fitem]))
      else:
        items.append('\n')
        for ditem in fault[fitem]:
          if ditem != 'errors':
            items.append(('%s: %s\n' % (ditem, fault[fitem][ditem])))
          else:
            for dfitem in fault[fitem][ditem]:
              items.append('\nError:\n')
              for eitem in dfitem:
                items.append(('  %s: %s\n' % (eitem, dfitem[eitem])))
    return ''.join(items).strip()

  def InjectXml(self, xml_in):
    """Hook into the SoapBuffer to test local SOAP XML.

    Prepares dump dict with a given input.

    Args:
      xml_in: str SOAP XML request, response, or both.
    """
    # Prepare string for use with regular expression.
    xml_in = xml_in.replace('\n', '%newline%')

    try:
      pattern = re.compile('(<SOAP-ENV:Envelope.*</SOAP-ENV:Envelope>)')
      req = pattern.findall(xml_in)
      pattern = re.compile('(<soap.*?:Envelope.*</soap.*?:Envelope>)')
      res = pattern.findall(xml_in)

      # Do we have a SOAP XML request?
      if req:
        # Rebuild original formatting of the string and dump it.
        req = req[0].replace('%newline%', '\n')
        self.__dump['dumpSOAPOut'] = (
            '%s Outgoing SOAP %s\n'
            '<?xml version="1.0" encoding="UTF-8"?>\n%s\n'
            '%s' % ('*' * 3, '*' * 54, req, '*' * 72))

      # Do we have a SOAP XML response?
      if res:
        # Rebuild original formatting of the string and dump it.
        res = res[0].replace('%newline%', '\n')
        self.__dump['dumpSOAPIn'] = (
            '%s Incoming SOAP %s\n'
            '<?xml version="1.0" encoding="UTF-8"?>\n%s\n'
            '%s' % ('*' * 3, '*' * 54, res.lstrip('\n'), '*' * 72))
    except Exception:
      msg = 'Invalid input, expecting SOAP XML request, response, or both.'
      raise InvalidInputError(msg)

  def __Indent(self, elem, level=0):
    """Add whitespace to the tree, to get its pretty-printed version.

    See, http://effbot.org/zone/element-lib.htm#prettyprint.
    """
    i = "\n" + level*" "
    if len(elem):
      if not elem.text or not elem.text.strip():
        elem.text = i + " "
      if not elem.tail or not elem.tail.strip():
        elem.tail = i
      for elem in elem:
        self.__Indent(elem, level+1)
      if not elem.tail or not elem.tail.strip():
        elem.tail = i
    else:
      if level and (not elem.tail or not elem.tail.strip()):
        elem.tail = i

  def __PrettyPrintXml(self, doc, level=0):
    """Return a pretty-printed version of the XML document.

    Args:
      doc: str an XML document.
      level: int level of prettiness, defaults to 0. If -1, remove prettiness.

    Returns:
      str pretty-printed version of the XML document.
    """
    # Make sure we have a valid doc to work with.
    if Utils.IsHtml(doc):
      return doc

    if self.__xml_parser == PYXML:
      dom = minidom.parseString(doc)
      pretty_doc = dom.toprettyxml(indent=' ', encoding='UTF-8')
    elif self.__xml_parser == ETREE:
      tree = etree.fromstring(doc)
      self.__Indent(tree)
      pretty_doc = etree.tostring(tree, 'UTF-8')

    # Adjust prettiness of data values in XML document.
    #
    # Before:  <operations>
    #            0
    #          </operations>
    #
    # After:   <operations>0</operations>
    pattern = re.compile('\n\s+\n')
    pretty_doc = pattern.sub('\n', pretty_doc)
    groups = re.findall('>(\n\s+(.*?)\n\s+)</', pretty_doc, re.M)
    for old, new in groups:
      if old and new and (new.find('<') > -1 or new.find('>') > -1):
        continue
      pretty_doc = pretty_doc.replace(old, new)

    if level == -1:
      pattern = re.compile('>\s+<')
      pretty_doc = pattern.sub('><', pretty_doc)

    return pretty_doc.strip('\n')