# -*- coding: utf-8 -*-
# ------------------------------------------------------------------------------
# Appy is a framework for building applications in the Python language.
# Copyright (C) 2007 Gaetan Delannay
# This program is free software; you can redistribute it and/or
# modify it under the terms of the GNU General Public License
# as published by the Free Software Foundation; either version 2
# of the License, or (at your option) any later version.
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
# You should have received a copy of the GNU General Public License
# along with this program; if not, write to the Free Software
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301,USA.
# ------------------------------------------------------------------------------
import xml.sax, difflib, types, cgi
from xml.parsers.expat import XML_PARAM_ENTITY_PARSING_NEVER
from xml.sax.handler import ContentHandler, ErrorHandler, feature_external_ges
from xml.sax.xmlreader import InputSource
from xml.sax import SAXParseException
from appy.shared import UnicodeBuffer
from appy.shared.errors import AppyError
from appy.shared.utils import sequenceTypes
from appy.shared.css import parseStyleAttribute
# Constants --------------------------------------------------------------------
xmlPrologue = '\n'
xhtmlPrologue = '\n'
CONVERSION_ERROR = '"%s" value "%s" could not be converted by the XML ' \
'unmarshaller.'
CUSTOM_CONVERSION_ERROR = 'Custom converter for "%s" values produced an ' \
'error while converting value "%s". %s'
XML_SPECIAL_CHARS = {'<': '<', '>': '>', '&': '&', '"': '"',
"'": '''}
XML_SPECIAL_CHARS_NO_APOS = XML_SPECIAL_CHARS.copy()
del XML_SPECIAL_CHARS_NO_APOS["'"]
XML_ENTITIES = {'lt': '<', 'gt': '>', 'amp': '&', 'quot': '"', 'apos': "'"}
HTML_ENTITIES = {
'iexcl': '¡', 'cent': '¢', 'pound': '£', 'curren': '€', 'yen': '¥',
'brvbar': 'Š', 'sect': '§', 'uml': '¨', 'copy':'©', 'ordf':'ª',
'laquo':'«', 'not':'¬', 'shy':'', 'reg':'®', 'macr':'¯', 'deg':'°',
'plusmn':'±', 'sup2':'²', 'sup3':'³', 'acute':'Ž',
'micro':'µ', 'para':'¶', 'middot':'·', 'cedil':'ž', 'sup1':'¹',
'ordm':'º', 'raquo':'»', 'frac14':'Œ', 'frac12':'œ', 'frac34':'Ÿ',
'iquest':'¿', 'Agrave':'À', 'Aacute':'Á', 'Acirc':'Â', 'Atilde':'Ã',
'Auml':'Ä', 'Aring':'Å', 'AElig':'Æ', 'Ccedil':'Ç', 'Egrave':'È',
'Eacute':'É', 'Ecirc':'Ê', 'Euml':'Ë', 'Igrave':'Ì', 'Iacute':'Í',
'Icirc':'Î', 'Iuml':'Ï', 'ETH':'Ð', 'Ntilde':'Ñ', 'Ograve':'Ò',
'Oacute':'Ó', 'Ocirc':'Ó', 'Otilde':'Õ', 'Ouml':'Ö', 'times':'×',
'Oslash':'Ø', 'Ugrave':'Ù', 'Uacute':'Ú', 'Ucirc':'Û', 'Uuml':'Ü',
'Yacute':'Ý', 'THORN':'Þ', 'szlig':'ß', 'agrave':'à', 'aacute':'á',
'acirc':'â', 'atilde':'ã', 'auml':'ä', 'aring':'å', 'aelig':'æ',
'ccedil':'ç', 'egrave':'è', 'eacute':'é', 'ecirc':'ê', 'euml':'ë',
'igrave':'ì', 'iacute':'í', 'icirc':'î', 'iuml':'ï', 'eth':'ð',
'ntilde':'ñ', 'ograve':'ò', 'oacute':'ó', 'ocirc':'ô', 'otilde':'õ',
'ouml':'ö', 'divide':'÷', 'oslash':'ø', 'ugrave':'ù', 'uacute':'ú',
'ucirc':'û', 'uuml':'ü', 'yacute':'ý', 'thorn':'þ', 'yuml':'ÿ',
'euro':'€', 'nbsp':' ', "rsquo":"'", "lsquo":"'", "ldquo":"'",
"rdquo":"'", 'ndash': '—', 'mdash': '—', 'oelig':'oe', 'quot': "'",
'mu': 'µ'}
import htmlentitydefs
for k, v in htmlentitydefs.entitydefs.iteritems():
if not HTML_ENTITIES.has_key(k) and not XML_ENTITIES.has_key(k):
HTML_ENTITIES[k] = ''
def escapeXml(s, format='xml', nsText='text'):
'''Returns p_s, whose XML special chars have been replaced with escaped XML
entities. If p_format is "odf", line breaks are converted to ODF line
breaks. In this case, it is needed to give the name of the "text"
namespace (p_nsText) as defined in the ODF document where the line breaks
must be inserted.'''
if isinstance(s, unicode):
res = u''
else:
res = ''
odf = format == 'odf'
for c in s:
if XML_SPECIAL_CHARS_NO_APOS.has_key(c):
# We do not escape 'apos': there is no particular need for that.
res += XML_SPECIAL_CHARS_NO_APOS[c]
elif odf and (c == '\n'):
res += '<%s:line-break/>' % nsText
elif odf and (c == '\r'):
pass
else:
res += c
return res
def escapeXhtml(s):
'''Return p_s, whose XHTML special chars and carriage return chars have
been replaced with corresponding XHTML entities.'''
if isinstance(s, unicode):
res = u''
else:
res = ''
for c in s:
if XML_SPECIAL_CHARS_NO_APOS.has_key(c):
res += XML_SPECIAL_CHARS_NO_APOS[c]
elif c == '\n':
res += '
'
elif c == '\r':
pass
else:
res += c
return res
# ------------------------------------------------------------------------------
class XmlElement:
'''Represents an XML tag.'''
def __init__(self, elem, attrs=None, nsUri=None):
'''An XmlElement instance may represent:
- an already parsed tag (in this case, p_elem may be prefixed with a
namespace);
- the definition of an XML element (in this case, no namespace can be
found in p_elem; but a namespace URI may be defined in p_nsUri).'''
self.elem = elem
self.attrs = attrs
if elem.find(':') != -1:
self.ns, self.name = elem.split(':')
else:
self.ns = ''
self.name = elem
self.nsUri = nsUri
def equalsTo(self, other, namespaces=None):
'''Does p_elem == p_other? If a p_namespaces dict is given, p_other must
define a nsUri.'''
res = None
if namespaces:
res = self.elem == ('%s:%s' % (namespaces[other.nsUri], other.name))
else:
res = self.elem == other.elem
return res
def __repr__(self):
res = self.elem
if self.attrs:
res += '('
for attrName, attrValue in self.attrs.items():
res += '%s="%s"' % (attrName, attrValue)
res += ')'
return res
def getFullName(self, namespaces=None):
'''Gets the name of the element including the namespace prefix.'''
if not namespaces:
res = self.elem
else:
res = '%s:%s' % (namespaces[self.nsUri], self.name)
return res
class XmlEnvironment:
'''An XML environment remembers a series of elements during a SAX parsing.
This class is an abstract class that gathers basic things like
namespaces.'''
def __init__(self):
# This dict contains the xml namespace declarations encountered so far
self.namespaces = {} # ~{s_namespaceUri:s_namespaceName}~
self.currentElem = None # The currently parsed element
self.parser = None
def manageNamespaces(self, attrs):
'''Manages namespaces definitions encountered in p_attrs.'''
for attrName, attrValue in attrs.items():
if attrName.startswith('xmlns:'):
self.namespaces[attrValue] = attrName[6:]
def ns(self, nsUri):
'''Returns the namespace corresponding to o_nsUri.'''
return self.namespaces[nsUri]
class XmlParser(ContentHandler, ErrorHandler):
'''Basic expat-based XML parser that does things like :
- remembering the currently parsed element;
- managing namespace declarations.
This parser also knows about HTML entities.'''
def __init__(self, env=None, caller=None, raiseOnError=True):
'''p_env should be an instance of a class that inherits from
XmlEnvironment: it specifies the environment to use for this SAX
parser.'''
ContentHandler.__init__(self)
if not env: env = XmlEnvironment()
self.env = env
self.env.parser = self
self.caller = caller # The class calling this parser
self.parser = xml.sax.make_parser() # Fast, standard expat parser
self.res = None # The result of parsing.
# Raise or not an error when a parsing error is encountered.
self.raiseOnError = raiseOnError
# ContentHandler methods ---------------------------------------------------
def startDocument(self):
parser = self.parser._parser
parser.UseForeignDTD(True)
parser.SetParamEntityParsing(XML_PARAM_ENTITY_PARSING_NEVER)
def setDocumentLocator(self, locator):
self.locator = locator
return self.env
def endDocument(self):
return self.env
def startElement(self, elem, attrs):
self.env.manageNamespaces(attrs)
if self.env.currentElem == None:
self.env.currentElem = XmlElement(elem, attrs=attrs)
else:
# Reuse the exiting instance in order to avoid creating one instance
# every time an elem is met in the XML file.
self.env.currentElem.__init__(elem, attrs)
return self.env
def endElement(self, elem):
self.env.currentElem.__init__(elem)
return self.env
def characters(self, content):
return self.env
def skippedEntity(self, name):
'''This method is called every time expat does not recognize an entity.
We provide here support for HTML entities.'''
if HTML_ENTITIES.has_key(name):
self.characters(HTML_ENTITIES[name].decode('utf-8'))
else:
# Put a question mark instead of raising an exception.
self.characters('?')
# ErrorHandler methods ---------------------------------------------------
def error(self, error):
if self.raiseOnError: raise error
else: print('SAX error %s' % str(error))
def fatalError(self, error):
if self.raiseOnError: raise error
else: print('SAX fatal error %s' % str(error))
def warning(self, error): pass
def parse(self, xml, source='string'):
'''Parses a XML stream.
* If p_source is "string", p_xml must be a string containing
valid XML content.
* If p_source is "file": p_xml can be:
- a string containing the path to the XML file on disk;
- a file instance opened for reading. Note that in this case, this
method will close it.
'''
try:
from cStringIO import StringIO
except ImportError:
from StringIO import StringIO
self.parser.setContentHandler(self)
self.parser.setErrorHandler(self)
self.parser.setFeature(feature_external_ges, False)
inputSource = InputSource()
if source == 'string':
inputSource.setByteStream(StringIO(xml))
else:
if not isinstance(xml, file):
xml = file(xml)
inputSource.setByteStream(xml)
self.parser.parse(inputSource)
if isinstance(xml, file): xml.close()
return self.res
# ------------------------------------------------------------------------------
from appy.shared import UnmarshalledFile
from appy import Object
try:
from DateTime import DateTime
except ImportError:
DateTime = 'unicode'
class XmlUnmarshaller(XmlParser):
'''This class allows to parse a XML file and recreate the corresponding web
of Python objects. This parser assumes that the XML file respects this
convention: any tag may define in attribute "type" storing the type of
its content, which may be:
bool * int * float * long * DateTime * tuple * list * object
If "object" is specified, it means that the tag contains sub-tags, each
one corresponding to the value of an attribute for this object.
if "tuple" is specified, it will be converted to a list.'''
def __init__(self, classes={}, tagTypes={}, conversionFunctions={},
utf8=True):
XmlParser.__init__(self)
# self.classes below is a dict whose keys are tag names and values are
# Python classes. During the unmarshalling process, when an object is
# encountered, instead of creating an instance of Object, we will create
# an instance of the class specified in self.classes.
# Root tag is named "xmlPythonData" by default by the XmlMarshaller.
# This will not work if the object in the specified tag is not an
# Object instance (ie it is a list or tuple or simple value). Note that
# we will not call the constructor of the specified class. We will
# simply create an instance of Objects and dynamically change the class
# of the created instance to this class.
if not isinstance(classes, dict) and classes:
# The user may only need to define a class for the root tag
self.classes = {'xmlPythonData': classes}
else:
self.classes = classes
# We expect that the parsed XML file will follow some conventions
# (ie, a tag that corresponds to a list has attribute type="list" or a
# tag that corresponds to an object has attribute type="object".). If
# it is not the case of p_xmlContent, you can provide the missing type
# information in p_tagTypes. Here is an example of p_tagTypes:
# {"information": "list", "days": "list", "person": "object"}.
self.tagTypes = tagTypes
# The parser assumes that data is represented in some standard way. If
# it is not the case, you may provide, in this dict, custom functions
# allowing to convert values of basic types (long, float, DateTime...).
# Every such function must take a single arg which is the value to
# convert and return the converted value. Dict keys are strings
# representing types ('bool', 'int', 'unicode', etc) and dict values are
# conversion functions. Here is an example:
# {'int': convertInteger, 'DateTime': convertDate}
# NOTE: you can even invent a new basic type, put it in self.tagTypes,
# and create a specific conversionFunction for it. This way, you can
# for example convert strings that have specific values (in this case,
# knowing that the value is a 'string' is not sufficient).
self.conversionFunctions = conversionFunctions
self.utf8 = utf8
def convertAttrs(self, attrs):
'''Converts XML attrs to a dict.'''
res = {}
for k, v in attrs.items():
if ':' in k: # An attr prefixed with a namespace. Remove this.
k = k.split(':')[-1]
res[str(k)] = v
return res
def startDocument(self):
XmlParser.startDocument(self)
self.res = None # The resulting web of Python objects (Object instances)
self.env.containerStack = [] # The stack of current "containers" where
# to store the next parsed element. A container can be a list, a tuple,
# an object (the root object of the whole web or a sub-object).
self.env.currentBasicType = None # Will hold the name of the currently
# parsed basic type (unicode, float...)
self.env.currentContent = '' # We store here the content of tags.
containerTags = ('tuple', 'list', 'dict', 'object', 'file')
numericTypes = ('bool', 'int', 'float', 'long')
def startElement(self, elem, attrs):
# Remember the name of the previous element
previousElem = None
if self.env.currentElem:
previousElem = self.env.currentElem.name
e = XmlParser.startElement(self, elem, attrs)
# Determine the type of the element.
elemType = 'unicode' # Default value
if attrs.has_key('type'):
elemType = attrs['type']
elif self.tagTypes.has_key(elem):
elemType = self.tagTypes[elem]
if elemType in self.containerTags:
# I must create a new container object.
if elemType == 'object':
newObject = Object(**self.convertAttrs(attrs))
elif elemType == 'tuple': newObject = [] # Tuples become lists
elif elemType == 'list': newObject = []
elif elemType == 'dict': newObject = {}
elif elemType == 'file':
newObject = UnmarshalledFile()
if attrs.has_key('name'):
newObject.name = attrs['name']
if attrs.has_key('mimeType'):
newObject.mimeType = attrs['mimeType']
else: newObject = Object(**self.convertAttrs(attrs))
# Store the value on the last container, or on the root object.
self.storeValue(elem, newObject)
# Push the new object on the container stack
e.containerStack.append(newObject)
else:
# If we are already parsing a basic type, it means that we were
# wrong for our diagnotsic of the containing element: it was not
# basic. We will make the assumption that the containing element is
# then an object.
if e.currentBasicType:
# Previous elem was an object: create it on the stack.
newObject = Object()
self.storeValue(previousElem, newObject)
e.containerStack.append(newObject)
e.currentBasicType = elemType
def storeValue(self, name, value):
'''Stores the newly parsed p_value (contained in tag p_name) on the
current container in environment self.env.'''
e = self.env
# Remove namespace prefix when relevant
if ':' in name: name = name.split(':')[-1]
# Change the class of the value if relevant
if (name in self.classes) and isinstance(value, Object):
value.__class__ = self.classes[name]
# Where must I store this value?
if not e.containerStack:
# I store the object at the root of the web.
self.res = value
else:
currentContainer = e.containerStack[-1]
if isinstance(currentContainer, list):
currentContainer.append(value)
elif isinstance(currentContainer, dict):
# If the current container is a dict, it means that p_value is
# a dict entry object named "entry" by convention and having
# attributes "k" and "v" that store, respectively, the key and
# the value of the entry. But this object is under construction:
# at this time, attributes "k" and "v" are not created yet. We
# will act in m_endElement, when the object will be finalized.
pass
elif isinstance(currentContainer, UnmarshalledFile):
currentContainer.content += value or ''
else:
# Current container is an object
if hasattr(currentContainer, name) and \
getattr(currentContainer, name):
# We have already encountered a sub-object with this name.
# Having several sub-objects with the same name, we will
# create a list.
attrValue = getattr(currentContainer, name)
if not isinstance(attrValue, list):
attrValue = [attrValue, value]
else:
attrValue.append(value)
else:
attrValue = value
setattr(currentContainer, name, attrValue)
def characters(self, content):
if not self.utf8:
content = content.encode('utf-8')
e = XmlParser.characters(self, content)
if e.currentBasicType:
e.currentContent += content
def endElement(self, elem):
e = XmlParser.endElement(self, elem)
if e.currentBasicType:
value = e.currentContent.strip()
if not value: value = None
else:
# If we have a custom converter for values of this type, use it.
if self.conversionFunctions.has_key(e.currentBasicType):
try:
value = self.conversionFunctions[e.currentBasicType](
value)
except Exception, err:
raise AppyError(CUSTOM_CONVERSION_ERROR % (
e.currentBasicType, value, str(err)))
# If not, try a standard conversion
elif e.currentBasicType in self.numericTypes:
try:
exec 'value = %s' % value
except SyntaxError:
raise AppyError(CONVERSION_ERROR % (
e.currentBasicType, value))
except NameError:
raise AppyError(CONVERSION_ERROR % (
e.currentBasicType, value))
# Check that the value is of the correct type. For instance,
# a float value with a comma in it could have been converted
# to a tuple instead of a float.
if not isinstance(value, eval(e.currentBasicType)):
raise AppyError(CONVERSION_ERROR % (
e.currentBasicType, value))
elif e.currentBasicType == 'DateTime':
value = DateTime(value)
elif e.currentBasicType == 'base64':
value = e.currentContent.decode('base64')
# Store the value on the last container
self.storeValue(elem, value)
# Clean the environment
e.currentBasicType = None
e.currentContent = ''
else:
elem = e.containerStack.pop()
# This element can be a temporary "entry" object representing a dict
# entry.
if e.containerStack:
lastContainer = e.containerStack[-1]
if isinstance(lastContainer, dict):
lastContainer[elem.k] = elem.v
# Alias: "unmarshall" -> "parse"
unmarshall = XmlParser.parse
# ------------------------------------------------------------------------------
class XmlMarshaller:
'''This class allows to produce a XML version of a Python object, which
respects some conventions as described in the doc of the corresponding
Unmarshaller (see above).'''
xmlEntities = {'<': '<', '>': '>', '&': '&', '"': '"',
"'": '''}
trueFalse = {True: 'True', False: 'False'}
fieldsToMarshall = 'all'
fieldsToExclude = []
atFiles = ('image', 'file') # Types of archetypes fields that contain files.
def __init__(self, cdata=False, dumpUnicode=False, conversionFunctions={},
dumpXmlPrologue=True, rootTag='xmlPythonData', namespaces={},
namespacedTags={}):
# If p_cdata is True, all string values will be dumped as XML CDATA.
self.cdata = cdata
# If p_dumpUnicode is True, the result will be unicode.
self.dumpUnicode = dumpUnicode
# The following dict stores specific conversion (=Python to XML)
# functions. A specific conversion function is useful when you are not
# happy with the way built-in converters work, or if you want to
# specify a specific way to represent, in XML, some particular Python
# object or value. In this dict, every key represents a given type
# (class names must be full-path class names); every value is a function
# accepting 2 args: first one is the UnicodeBuffer where the result is
# being dumped, while the second one is the Python object or value to
# dump.
self.conversionFunctions = conversionFunctions
# If dumpXmlPrologue is True, the XML prologue will be dumped.
self.dumpXmlPrologue = dumpXmlPrologue
# The name of the root tag
self.rootElementName = rootTag
# The namespaces that will be defined at the root of the XML message.
# It is a dict whose keys are namespace prefixes and whose values are
# namespace URLs. If you want to specify a default namespace, specify an
# entry with an empty string as a key.
self.namespaces = namespaces
# The following dict will tell which XML tags will get which namespace
# prefix ({s_tagName: s_prefix}). Special optional dict entry
# '*':s_prefix will indicate a default prefix that will be applied to
# any tag that does not have it own key in this dict.
self.namespacedTags = namespacedTags
self.objectType = None # Will be given by method m_marshal
def getTagName(self, name):
'''Returns the name of tag p_name as will be dumped. It can be p_name,
or p_name prefixed with a namespace prefix (will depend on
self.prefixedTags).'''
# Determine the prefix
prefix = ''
if name in self.namespacedTags: prefix = self.namespacedTags[name]
elif '*' in self.namespacedTags: prefix = self.namespacedTags['*']
if prefix: return '%s:%s' % (prefix, name)
return name
def isAnObject(self, instance):
'''Returns True if p_instance is a class instance, False if it is a
basic type, or tuple, sequence, etc.'''
if instance.__class__.__name__ == 'LazyMap': return False
iType = type(instance)
if iType == types.InstanceType:
return True
elif iType.__name__ == 'ImplicitAcquirerWrapper':
# This is the case with Archetype instances
return True
elif iType.__class__.__name__ == 'ExtensionClass':
return True
return False
def isAList(self, value):
'''Is p_value a list?'''
return value.__class__.__name__ in ('list', 'PersistentList', 'LazyMap')
def isADict(self, value):
'''Is p_value a dict?'''
return value.__class__.__name__ in ('dict', 'PersistentMapping')
def dumpRootTag(self, res, instance):
'''Dumps the root tag.'''
# Dumps the name of the tag.
tagName = self.getTagName(self.rootElementName)
res.write('<'); res.write(tagName)
# Dumps namespace definitions if any
for prefix, url in self.namespaces.iteritems():
if not prefix:
pre = 'xmlns' # The default namespace
else:
pre = 'xmlns:%s' % prefix
res.write(' %s="%s"' % (pre, url))
# Dumps Appy- or Plone-specific attribute
if self.objectType != 'popo':
res.write(' type="object" id="%s"' % instance.UID())
res.write('>')
return tagName
def dumpString(self, res, s):
'''Dumps a string into the result.'''
if self.cdata: res.write('')
def dumpFile(self, res, v):
'''Dumps a file into the result.'''
if not v: return
# p_value contains the (possibly binary) content of a file. We will
# encode it in Base64, in one or several parts.
partTag = self.getTagName('part')
res.write('<%s type="base64" number="1">' % partTag)
if hasattr(v, 'data'):
# The file is an Archetypes file.
valueType = v.data.__class__.__name__
if valueType == 'Pdata':
# There will be several parts.
res.write(v.data.data.encode('base64'))
# Write subsequent parts
nextPart = v.data.next
nextPartNumber = 2
while nextPart:
res.write('%s>' % partTag) # Close the previous part
res.write('<%s type="base64" number="%d">' % \
(partTag, nextPartNumber))
res.write(nextPart.data.encode('base64'))
nextPart = nextPart.next
nextPartNumber += 1
else:
res.write(v.data.encode('base64'))
else:
res.write(v.encode('base64'))
res.write('%s>' % partTag)
def dumpDict(self, res, v):
'''Dumps the XML version of dict p_v.'''
for key, value in v.iteritems():
res.write('
or