# $Id: selftest.py 2266 2005-01-29 23:42:26Z fredrik $
# elementtree selftest program
# this test script uses Python's "doctest" module to check that the
# *test script* works as expected.
import sys, StringIO
import cElementTree as ElementTree
def unserialize(text):
import StringIO
file = StringIO.StringIO(text)
tree = ElementTree.parse(file)
return tree.getroot()
def serialize(elem, encoding=None):
import StringIO
file = StringIO.StringIO()
tree = ElementTree.ElementTree(elem)
if encoding:
tree.write(file, encoding)
else:
tree.write(file)
return file.getvalue()
def summarize(elem):
return elem.tag
def summarize_list(seq):
return map(summarize, seq)
SAMPLE_XML = unserialize("""
text
""")
SAMPLE_XML_NS = unserialize("""
text
""")
# interface tests
def check_string(string):
len(string)
for char in string:
if len(char) != 1:
print "expected one-character string, got %r" % char
new_string = string + ""
new_string = string + " "
string[:0]
def check_mapping(mapping):
len(mapping)
keys = mapping.keys()
items = mapping.items()
for key in keys:
item = mapping[key]
mapping["key"] = "value"
if mapping["key"] != "value":
print "expected value string, got %r" % mapping["key"]
def check_element(element):
if not hasattr(element, "tag"):
print "no tag member"
if not hasattr(element, "attrib"):
print "no attrib member"
if not hasattr(element, "text"):
print "no text member"
if not hasattr(element, "tail"):
print "no tail member"
check_string(element.tag)
check_mapping(element.attrib)
if element.text != None:
check_string(element.text)
if element.tail != None:
check_string(element.tail)
def check_element_tree(tree):
check_element(tree.getroot())
def check_method(method):
if not callable(method):
print method, "not callable"
def version():
"""
>>> ElementTree.__version__
'1.0.1'
>>> ElementTree.VERSION
'1.0.1'
"""
def element():
"""
Test element tree interface.
>>> element = ElementTree.Element("tag")
>>> check_element(element)
>>> tree = ElementTree.ElementTree(element)
>>> check_element_tree(tree)
Make sure all standard element methods exist.
>>> check_method(element.append)
>>> check_method(element.insert)
>>> check_method(element.remove)
>>> check_method(element.getchildren)
>>> check_method(element.find)
>>> check_method(element.findall)
>>> check_method(element.findtext)
>>> check_method(element.clear)
>>> check_method(element.get)
>>> check_method(element.set)
>>> check_method(element.keys)
>>> check_method(element.items)
>>> check_method(element.getiterator)
Basic method sanity checks.
>>> serialize(element) # 1
''
>>> subelement = ElementTree.Element("subtag")
>>> element.append(subelement)
>>> serialize(element) # 2
''
>>> element.insert(0, subelement)
>>> serialize(element) # 3
''
>>> element.remove(subelement)
>>> serialize(element) # 4
''
>>> element.remove(subelement)
>>> serialize(element) # 5
''
>>> element.remove(subelement)
Traceback (most recent call last):
ValueError: list.remove(x): x not in list
>>> serialize(element) # 6
''
"""
def parsefile():
"""
Test parsing from file. Note that we're opening the files in
here; by default, the 'parse' function opens the file in binary
mode, and doctest doesn't filter out carriage returns.
>>> tree = ElementTree.parse(open("samples/simple.xml", "r"))
>>> tree.write(sys.stdout)
text
texttail
>>> tree = ElementTree.parse(open("samples/simple-ns.xml", "r"))
>>> tree.write(sys.stdout)
text
texttail
>>> parser = ElementTree.XMLParser()
>>> parser.version
'Expat 1.95.8'
>>> parser.feed(open("samples/simple.xml").read())
>>> print serialize(parser.close())
text
texttail
>>> parser = ElementTree.XMLTreeBuilder() # 1.2 compatibility
>>> parser.feed(open("samples/simple.xml").read())
>>> print serialize(parser.close())
text
texttail
>>> target = ElementTree.TreeBuilder()
>>> parser = ElementTree.XMLParser(target=target)
>>> parser.feed(open("samples/simple.xml").read())
>>> print serialize(parser.close())
text
texttail
"""
def writefile():
"""
>>> elem = ElementTree.Element("tag")
>>> elem.text = "text"
>>> serialize(elem)
'text'
>>> ElementTree.SubElement(elem, "subtag").text = "subtext"
>>> serialize(elem)
'textsubtext'
"""
def encoding():
r"""
Test encoding issues.
>>> elem = ElementTree.Element("tag")
>>> elem.text = u"abc"
>>> serialize(elem)
'abc'
>>> serialize(elem, "utf-8")
'abc'
>>> serialize(elem, "us-ascii")
'abc'
>>> serialize(elem, "iso-8859-1")
"\nabc"
>>> elem.text = "<&\"\'>"
>>> serialize(elem)
'<&"\'>'
>>> serialize(elem, "utf-8")
'<&"\'>'
>>> serialize(elem, "us-ascii") # cdata characters
'<&"\'>'
>>> serialize(elem, "iso-8859-1")
'\n<&"\'>'
>>> elem.attrib["key"] = "<&\"\'>"
>>> elem.text = None
>>> serialize(elem)
''
>>> serialize(elem, "utf-8")
''
>>> serialize(elem, "us-ascii")
''
>>> serialize(elem, "iso-8859-1")
'\n'
>>> elem.text = u'\xe5\xf6\xf6<>'
>>> elem.attrib.clear()
>>> serialize(elem)
'åöö<>'
>>> serialize(elem, "utf-8")
'\xc3\xa5\xc3\xb6\xc3\xb6<>'
>>> serialize(elem, "us-ascii")
'åöö<>'
>>> serialize(elem, "iso-8859-1")
"\n\xe5\xf6\xf6<>"
>>> elem.attrib["key"] = u'\xe5\xf6\xf6<>'
>>> elem.text = None
>>> serialize(elem)
''
>>> serialize(elem, "utf-8")
''
>>> serialize(elem, "us-ascii")
''
>>> serialize(elem, "iso-8859-1")
'\n'
"""
def qname():
"""
Test QName handling.
1) decorated tags
>>> elem = ElementTree.Element("{uri}tag")
>>> serialize(elem) # 1.1
''
2) decorated attributes
>>> elem.attrib["{uri}key"] = "value"
>>> serialize(elem) # 2.1
''
"""
def cdata():
"""
Test CDATA handling (etc).
>>> serialize(unserialize("hello"))
'hello'
>>> serialize(unserialize("hello"))
'hello'
>>> serialize(unserialize(""))
'hello'
"""
def find():
"""
Test find methods (including xpath syntax).
>>> elem = SAMPLE_XML
>>> elem.find("tag").tag
'tag'
>>> ElementTree.ElementTree(elem).find("tag").tag
'tag'
>>> elem.find("section/tag").tag
'tag'
>>> ElementTree.ElementTree(elem).find("section/tag").tag
'tag'
>>> elem.findtext("tag")
'text'
>>> elem.findtext("tog", "default")
'default'
>>> ElementTree.ElementTree(elem).findtext("tag")
'text'
>>> elem.findtext("section/tag")
'subtext'
>>> ElementTree.ElementTree(elem).findtext("section/tag")
'subtext'
>>> summarize_list(elem.findall("tag"))
['tag', 'tag']
>>> summarize_list(elem.findall("*"))
['tag', 'tag', 'section']
>>> summarize_list(elem.findall(".//tag"))
['tag', 'tag', 'tag']
>>> summarize_list(elem.findall("section/tag"))
['tag']
>>> summarize_list(elem.findall("section//tag"))
['tag']
>>> summarize_list(elem.findall("section/*"))
['tag']
>>> summarize_list(elem.findall("section//*"))
['tag']
>>> summarize_list(elem.findall("section/.//*"))
['tag']
>>> summarize_list(elem.findall("*/*"))
['tag']
>>> summarize_list(elem.findall("*//*"))
['tag']
>>> summarize_list(elem.findall("*/tag"))
['tag']
>>> summarize_list(elem.findall("*/./tag"))
['tag']
>>> summarize_list(elem.findall("./tag"))
['tag', 'tag']
>>> summarize_list(elem.findall(".//tag"))
['tag', 'tag', 'tag']
>>> summarize_list(elem.findall("././tag"))
['tag', 'tag']
>>> summarize_list(ElementTree.ElementTree(elem).findall("/tag"))
['tag', 'tag']
>>> summarize_list(ElementTree.ElementTree(elem).findall("./tag"))
['tag', 'tag']
>>> elem = SAMPLE_XML_NS
>>> summarize_list(elem.findall("tag"))
[]
>>> summarize_list(elem.findall("{http://effbot.org/ns}tag"))
['{http://effbot.org/ns}tag', '{http://effbot.org/ns}tag']
>>> summarize_list(elem.findall(".//{http://effbot.org/ns}tag"))
['{http://effbot.org/ns}tag', '{http://effbot.org/ns}tag', '{http://effbot.org/ns}tag']
"""
def copy():
"""
Test copy handling (etc).
>>> import copy
>>> e1 = unserialize("hello")
>>> e2 = copy.copy(e1)
>>> e3 = copy.deepcopy(e1)
>>> e1.find("foo").tag = "bar"
>>> serialize(e1)
'hello'
>>> serialize(e2)
'hello'
>>> serialize(e3)
'hello'
"""
def attrib():
"""
Test attribute handling.
>>> elem = ElementTree.Element("tag")
>>> elem.get("key") # 1.1
>>> elem.get("key", "default") # 1.2
'default'
>>> elem.set("key", "value")
>>> elem.get("key") # 1.3
'value'
>>> elem = ElementTree.Element("tag", key="value")
>>> elem.get("key") # 2.1
'value'
>>> elem.attrib # 2.2
{'key': 'value'}
>>> elem = ElementTree.Element("tag", {"key": "value"})
>>> elem.get("key") # 3.1
'value'
>>> elem.attrib # 3.2
{'key': 'value'}
>>> elem = ElementTree.Element("tag", {"key": "other"}, key="value")
>>> elem.get("key") # 4.1
'value'
>>> elem.attrib # 4.2
{'key': 'value'}
"""
def makeelement():
"""
Test makeelement handling.
>>> elem = ElementTree.Element("tag")
>>> subelem = elem.makeelement("subtag", {"key": "value"})
>>> elem.append(subelem)
>>> serialize(elem)
''
>>> elem.clear()
>>> serialize(elem)
''
>>> elem.append(subelem)
>>> serialize(elem)
''
"""
def iterparse():
"""
Test iterparse interface.
>>> iterparse = ElementTree.iterparse
>>> context = iterparse("samples/simple.xml")
>>> for action, elem in context:
... print action, elem.tag
end element
end element
end empty-element
end root
>>> context.root.tag
'root'
>>> context = iterparse("samples/simple-ns.xml")
>>> for action, elem in context:
... print action, elem.tag
end {namespace}element
end {namespace}element
end {namespace}empty-element
end {namespace}root
>>> events = ()
>>> context = iterparse("samples/simple.xml", events)
>>> for action, elem in context:
... print action, elem.tag
>>> events = ()
>>> context = iterparse("samples/simple.xml", events=events)
>>> for action, elem in context:
... print action, elem.tag
>>> events = ("start", "end")
>>> context = iterparse("samples/simple.xml", events)
>>> for action, elem in context:
... print action, elem.tag
start root
start element
end element
start element
end element
start empty-element
end empty-element
end root
>>> events = ("start", "end", "start-ns", "end-ns")
>>> context = iterparse("samples/simple-ns.xml", events)
>>> for action, elem in context:
... if action in ("start", "end"):
... print action, elem.tag
... else:
... print action, elem
start-ns ('', 'namespace')
start {namespace}root
start {namespace}element
end {namespace}element
start {namespace}element
end {namespace}element
start {namespace}empty-element
end {namespace}empty-element
end {namespace}root
end-ns None
"""
def custom_builder():
"""
Test parser w. custom builder.
>>> class Builder:
... def start(self, tag, attrib):
... print "start", tag
... def end(self, tag):
... print "end", tag
... def data(self, text):
... pass
>>> builder = Builder()
>>> parser = ElementTree.XMLParser(builder)
>>> parser.feed(open("samples/simple.xml", "r").read())
start root
start element
end element
start element
end element
start empty-element
end empty-element
end root
"""
def getchildren():
"""
>>> tree = ElementTree.parse(open("samples/simple.xml", "r"))
>>> for elem in tree.getiterator():
... summarize_list(elem.getchildren())
['element', 'element', 'empty-element']
[]
[]
[]
"""
ENTITY_XML = """\
%user-entities;
]>
&entity;
"""
def entity():
"""
Test entity handling.
1) bad entities
>>> ElementTree.XML("&entity;")
Traceback (most recent call last):
SyntaxError: undefined entity: line 1, column 10
2) custom entity
>>> parser = ElementTree.XMLParser()
>>> parser.entity["entity"] = "text"
>>> parser.feed(ENTITY_XML)
>>> root = parser.close()
>>> serialize(root)
'text'
"""
if __name__ == "__main__":
import doctest, selftest
failed, tested = doctest.testmod(selftest)
print tested - failed, "tests ok."