def __call__(self, stream):
namespace = self.namespace
for kind, data, pos in stream:
if kind is START or kind is EMPTY:
tag, attrs = data
if tag.namespace and tag not in namespace:
continue
new_attrs = []
for attr, value in attrs:
if not attr.namespace or attr in namespace:
new_attrs.append((attr, value))
data = tag.localname, Attrs(new_attrs)
elif kind is END:
if data.namespace and data not in namespace:
continue
data = data.localname
elif kind is START_NS or kind is END_NS:
continue
yield kind, data, pos
def __call__(self, kind, data, pos, namespaces, variables):
namespace = Namespace(namespaces.get(self.prefix))
if kind is START:
if self.principal_type is ATTRIBUTE and data[1]:
return Attrs([(name, value) for name, value in data[1]
if name in namespace]) or None
else:
return data[0] in namespace
def test_multiple_bound_namespaces(self):
stream = Stream([
(Stream.START, (QName('div'), Attrs()), (None, -1, -1)),
(Stream.TEXT, '\n ', (None, -1, -1)),
(Stream.START_NS, ('x', 'http://example.org/'), (None, -1, -1)),
(Stream.START, (QName('http://example.org/}p'), Attrs()), (None, -1, -1)),
(Stream.END, QName('http://example.org/}p'), (None, -1, -1)),
(Stream.END_NS, 'x', (None, -1, -1)),
(Stream.TEXT, '\n ', (None, -1, -1)),
(Stream.START_NS, ('x', 'http://example.org/'), (None, -1, -1)),
(Stream.START, (QName('http://example.org/}p'), Attrs()), (None, -1, -1)),
(Stream.END, QName('http://example.org/}p'), (None, -1, -1)),
(Stream.END_NS, 'x', (None, -1, -1)),
(Stream.TEXT, '\n ', (None, -1, -1)),
(Stream.END, QName('div'), (None, -1, -1)),
])
output = stream.render(XMLSerializer)
self.assertEqual("""<div>
<x:p xmlns:x="http://example.org/"/>
<x:p xmlns:x="http://example.org/"/>
</div>""", output)
def handle_starttag(self, tag, attrib):
fixed_attrib = []
for name, value in attrib: # Fixup minimized attributes
if value is None:
value = unicode(name)
elif not isinstance(value, unicode):
value = value.decode(self.encoding, 'replace')
fixed_attrib.append((QName(name), stripentities(value)))
self._enqueue(START, (QName(tag), Attrs(fixed_attrib)))
if tag in self._EMPTY_ELEMS:
self._enqueue(END, QName(tag))
else:
self._open_tags.append(tag)
def test_nested_default_namespaces(self):
stream = Stream([
(Stream.START_NS, ('', 'http://example.org/'), (None, -1, -1)),
(Stream.START, (QName('http://example.org/}div'), Attrs()), (None, -1, -1)),
(Stream.TEXT, '\n ', (None, -1, -1)),
(Stream.START_NS, ('', 'http://example.org/'), (None, -1, -1)),
(Stream.START, (QName('http://example.org/}p'), Attrs()), (None, -1, -1)),
(Stream.END, QName('http://example.org/}p'), (None, -1, -1)),
(Stream.END_NS, '', (None, -1, -1)),
(Stream.TEXT, '\n ', (None, -1, -1)),
(Stream.START_NS, ('', 'http://example.org/'), (None, -1, -1)),
(Stream.START, (QName('http://example.org/}p'), Attrs()), (None, -1, -1)),
(Stream.END, QName('http://example.org/}p'), (None, -1, -1)),
(Stream.END_NS, '', (None, -1, -1)),
(Stream.TEXT, '\n ', (None, -1, -1)),
(Stream.END, QName('http://example.org/}div'), (None, -1, -1)),
(Stream.END_NS, '', (None, -1, -1))
])
output = stream.render(XMLSerializer)
self.assertEqual("""<div xmlns="http://example.org/">
<p/>
<p/>
</div>""", output)
def _eval(self, stream, ctxt):
"""Internal stream filter that evaluates any expressions in `START` and
`TEXT` events.
"""
filters = (self._flatten, self._eval)
for kind, data, pos in stream:
if kind is START and data[1]:
# Attributes may still contain expressions in start tags at
# this point, so do some evaluation
tag, attrs = data
new_attrs = []
for name, substream in attrs:
if isinstance(substream, basestring):
value = substream
else:
values = []
for subkind, subdata, subpos in self._eval(substream,
ctxt):
if subkind is TEXT:
values.append(subdata)
value = [x for x in values if x is not None]
if not value:
continue
new_attrs.append((name, u''.join(value)))
yield kind, (tag, Attrs(new_attrs)), pos
elif kind is EXPR:
result = data.evaluate(ctxt)
if result is not None:
# First check for a string, otherwise the iterable test below
# succeeds, and the string will be chopped up into individual
# characters
if isinstance(result, basestring):
yield TEXT, result, pos
elif hasattr(result, '__iter__'):
substream = _ensure(result)
for filter_ in filters:
substream = filter_(substream, ctxt)
for event in substream:
yield event
else:
yield TEXT, unicode(result), pos
else:
yield kind, data, pos
def ET(element):
"""Convert a given ElementTree element to a markup stream.
:param element: an ElementTree element
:return: a markup stream
"""
tag_name = QName(element.tag.lstrip('{'))
attrs = Attrs([(QName(attr.lstrip('{')), value)
for attr, value in element.items()])
yield START, (tag_name, attrs), (None, -1, -1)
if element.text:
yield TEXT, element.text, (None, -1, -1)
for child in element.getchildren():
for item in ET(child):
yield item
yield END, tag_name, (None, -1, -1)
if element.tail:
yield TEXT, element.tail, (None, -1, -1)
def __call__(self, stream):
prefixes = dict([(v, [k]) for k, v in self.prefixes.items()])
namespaces = {XML_NAMESPACE.uri: ['xml']}
def _push_ns(prefix, uri):
namespaces.setdefault(uri, []).append(prefix)
prefixes.setdefault(prefix, []).append(uri)
ns_attrs = []
_push_ns_attr = ns_attrs.append
def _make_ns_attr(prefix, uri):
return u'xmlns%s' % (prefix and ':%s' % prefix or ''), uri
def _gen_prefix():
val = 0
while 1:
val += 1
yield 'ns%d' % val
_gen_prefix = _gen_prefix().next
for kind, data, pos in stream:
if kind is START or kind is EMPTY:
tag, attrs = data
tagname = tag.localname
tagns = tag.namespace
if tagns:
if tagns in namespaces:
prefix = namespaces[tagns][-1]
if prefix:
tagname = u'%s:%s' % (prefix, tagname)
else:
_push_ns_attr((u'xmlns', tagns))
_push_ns('', tagns)
new_attrs = []
for attr, value in attrs:
attrname = attr.localname
attrns = attr.namespace
if attrns:
if attrns not in namespaces:
prefix = _gen_prefix()
_push_ns(prefix, attrns)
_push_ns_attr(('xmlns:%s' % prefix, attrns))
else:
prefix = namespaces[attrns][-1]
if prefix:
attrname = u'%s:%s' % (prefix, attrname)
new_attrs.append((attrname, value))
yield kind, (tagname, Attrs(ns_attrs + new_attrs)), pos
del ns_attrs[:]
elif kind is END:
tagname = data.localname
tagns = data.namespace
if tagns:
prefix = namespaces[tagns][-1]
if prefix:
tagname = u'%s:%s' % (prefix, tagname)
yield kind, tagname, pos
elif kind is START_NS:
prefix, uri = data
if uri not in namespaces:
prefix = prefixes.get(uri, [prefix])[-1]
_push_ns_attr(_make_ns_attr(prefix, uri))
_push_ns(prefix, uri)
elif kind is END_NS:
if data in prefixes:
uris = prefixes.get(data)
uri = uris.pop()
if not uris:
del prefixes[data]
if uri not in uris or uri != uris[-1]:
uri_prefixes = namespaces[uri]
uri_prefixes.pop()
if not uri_prefixes:
del namespaces[uri]
if ns_attrs:
attr = _make_ns_attr(data, uri)
if attr in ns_attrs:
ns_attrs.remove(attr)
else:
yield kind, data, pos
def _handle_start(self, tag, attrib):
attrs = Attrs([(QName(name), value)
for name, value in zip(*[iter(attrib)] * 2)])
self._enqueue(START, (QName(tag), attrs))
def _parse(self, source, encoding):
streams = [[]] # stacked lists of events of the "compiled" template
dirmap = {} # temporary mapping of directives to elements
ns_prefix = {}
depth = 0
fallbacks = []
includes = []
if not isinstance(source, Stream):
source = XMLParser(source,
filename=self.filename,
encoding=encoding)
for kind, data, pos in source:
stream = streams[-1]
if kind is START_NS:
# Strip out the namespace declaration for template directives
prefix, uri = data
ns_prefix[prefix] = uri
if uri not in (self.DIRECTIVE_NAMESPACE,
self.XINCLUDE_NAMESPACE):
stream.append((kind, data, pos))
elif kind is END_NS:
uri = ns_prefix.pop(data, None)
if uri and uri not in (self.DIRECTIVE_NAMESPACE,
self.XINCLUDE_NAMESPACE):
stream.append((kind, data, pos))
elif kind is START:
# Record any directive attributes in start tags
tag, attrs = data
directives = []
strip = False
if tag in self.DIRECTIVE_NAMESPACE:
cls = self._dir_by_name.get(tag.localname)
if cls is None:
raise BadDirectiveError(tag.localname, self.filepath,
pos[1])
value = attrs.get(getattr(cls, 'ATTRIBUTE', None), '')
directives.append((cls, value, ns_prefix.copy(), pos))
strip = True
new_attrs = []
for name, value in attrs:
if name in self.DIRECTIVE_NAMESPACE:
cls = self._dir_by_name.get(name.localname)
if cls is None:
raise BadDirectiveError(name.localname,
self.filepath, pos[1])
directives.append((cls, value, ns_prefix.copy(), pos))
else:
if value:
value = list(
interpolate(value,
self.basedir,
pos[0],
pos[1],
pos[2],
lookup=self.lookup))
if len(value) == 1 and value[0][0] is TEXT:
value = value[0][1]
else:
value = [(TEXT, u'', pos)]
new_attrs.append((name, value))
new_attrs = Attrs(new_attrs)
if directives:
index = self._dir_order.index
directives.sort(lambda a, b: cmp(index(a[0]), index(b[0])))
dirmap[(depth, tag)] = (directives, len(stream), strip)
if tag in self.XINCLUDE_NAMESPACE:
if tag.localname == 'include':
include_href = new_attrs.get('href')
if not include_href:
raise TemplateSyntaxError(
'Include misses required '
'attribute "href"', self.filepath, *pos[1:])
includes.append(include_href)
streams.append([])
elif tag.localname == 'fallback':
streams.append([])
fallbacks.append(streams[-1])
else:
stream.append((kind, (tag, new_attrs), pos))
depth += 1
elif kind is END:
depth -= 1
if fallbacks and data == self.XINCLUDE_NAMESPACE['fallback']:
assert streams.pop() is fallbacks[-1]
elif data == self.XINCLUDE_NAMESPACE['include']:
fallback = None
if len(fallbacks) == len(includes):
fallback = fallbacks.pop()
streams.pop() # discard anything between the include tags
# and the fallback element
stream = streams[-1]
stream.append((INCLUDE, (includes.pop(), fallback), pos))
else:
stream.append((kind, data, pos))
# If there have have directive attributes with the corresponding
# start tag, move the events inbetween into a "subprogram"
if (depth, data) in dirmap:
directives, start_offset, strip = dirmap.pop((depth, data))
substream = stream[start_offset:]
if strip:
substream = substream[1:-1]
stream[start_offset:] = [(SUB, (directives, substream),
pos)]
elif kind is PI and data[0] == 'python':
try:
# As Expat doesn't report whitespace between the PI target
# and the data, we have to jump through some hoops here to
# get correctly indented Python code
# Unfortunately, we'll still probably not get the line
# number quite right
lines = [
line.expandtabs() for line in data[1].splitlines()
]
first = lines[0]
rest = dedent('\n'.join(lines[1:])).rstrip()
if first.rstrip().endswith(':') and not rest[0].isspace():
rest = '\n'.join(
[' ' + line for line in rest.splitlines()])
source = '\n'.join([first, rest])
suite = Suite(source,
self.filepath,
pos[1],
lookup=self.lookup)
except SyntaxError, err:
raise TemplateSyntaxError(err, self.filepath,
pos[1] + (err.lineno or 1) - 1,
pos[2] + (err.offset or 0))
stream.append((EXEC, suite, pos))
elif kind is TEXT:
for kind, data, pos in interpolate(data,
self.basedir,
pos[0],
pos[1],
pos[2],
lookup=self.lookup):
stream.append((kind, data, pos))