annotate genshi/core.py @ 932:e53161c2773c

Merge r1140 from py3k: add support for python 3 to core genshi components (genshi.core, genshi.input and genshi.output): * default input and output encodings changed from UTF-8 to None (i.e. unicode strings) * Namespace and QName objects do not call stringrepr in __repr__ in Python 3 since repr() returns a unicode string there. * track changes to expat parser in Python 3 (mostly it accepts bytes instead of strings)
author hodgestar
date Fri, 18 Mar 2011 09:08:12 +0000
parents 8cef75b02ac1
children
rev   line source
860
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
1 # -*- coding: utf-8 -*-
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
2 #
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
3 # Copyright (C) 2006-2009 Edgewall Software
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
4 # All rights reserved.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
5 #
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
6 # This software is licensed as described in the file COPYING, which
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
7 # you should have received as part of this distribution. The terms
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
8 # are also available at http://genshi.edgewall.org/wiki/License.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
9 #
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
10 # This software consists of voluntary contributions made by many
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
11 # individuals. For the exact contribution history, see the revision
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
12 # history and logs, available at http://genshi.edgewall.org/log/.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
13
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
14 """Core classes for markup processing."""
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
15
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
16 try:
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
17 reduce # builtin in Python < 3
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
18 except NameError:
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
19 from functools import reduce
932
e53161c2773c Merge r1140 from py3k:
hodgestar
parents: 923
diff changeset
20 import sys
860
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
21 from itertools import chain
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
22 import operator
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
23
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
24 from genshi.util import plaintext, stripentities, striptags, stringrepr
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
25
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
26 __all__ = ['Stream', 'Markup', 'escape', 'unescape', 'Attrs', 'Namespace',
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
27 'QName']
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
28 __docformat__ = 'restructuredtext en'
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
29
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
30
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
31 class StreamEventKind(str):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
32 """A kind of event on a markup stream."""
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
33 __slots__ = []
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
34 _instances = {}
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
35
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
36 def __new__(cls, val):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
37 return cls._instances.setdefault(val, str.__new__(cls, val))
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
38
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
39
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
40 class Stream(object):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
41 """Represents a stream of markup events.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
42
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
43 This class is basically an iterator over the events.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
44
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
45 Stream events are tuples of the form::
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
46
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
47 (kind, data, position)
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
48
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
49 where ``kind`` is the event kind (such as `START`, `END`, `TEXT`, etc),
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
50 ``data`` depends on the kind of event, and ``position`` is a
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
51 ``(filename, line, offset)`` tuple that contains the location of the
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
52 original element or text in the input. If the original location is unknown,
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
53 ``position`` is ``(None, -1, -1)``.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
54
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
55 Also provided are ways to serialize the stream to text. The `serialize()`
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
56 method will return an iterator over generated strings, while `render()`
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
57 returns the complete generated text at once. Both accept various parameters
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
58 that impact the way the stream is serialized.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
59 """
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
60 __slots__ = ['events', 'serializer']
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
61
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
62 START = StreamEventKind('START') #: a start tag
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
63 END = StreamEventKind('END') #: an end tag
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
64 TEXT = StreamEventKind('TEXT') #: literal text
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
65 XML_DECL = StreamEventKind('XML_DECL') #: XML declaration
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
66 DOCTYPE = StreamEventKind('DOCTYPE') #: doctype declaration
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
67 START_NS = StreamEventKind('START_NS') #: start namespace mapping
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
68 END_NS = StreamEventKind('END_NS') #: end namespace mapping
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
69 START_CDATA = StreamEventKind('START_CDATA') #: start CDATA section
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
70 END_CDATA = StreamEventKind('END_CDATA') #: end CDATA section
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
71 PI = StreamEventKind('PI') #: processing instruction
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
72 COMMENT = StreamEventKind('COMMENT') #: comment
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
73
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
74 def __init__(self, events, serializer=None):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
75 """Initialize the stream with a sequence of markup events.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
76
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
77 :param events: a sequence or iterable providing the events
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
78 :param serializer: the default serialization method to use for this
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
79 stream
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
80
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
81 :note: Changed in 0.5: added the `serializer` argument
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
82 """
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
83 self.events = events #: The underlying iterable producing the events
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
84 self.serializer = serializer #: The default serializion method
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
85
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
86 def __iter__(self):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
87 return iter(self.events)
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
88
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
89 def __or__(self, function):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
90 """Override the "bitwise or" operator to apply filters or serializers
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
91 to the stream, providing a syntax similar to pipes on Unix shells.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
92
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
93 Assume the following stream produced by the `HTML` function:
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
94
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
95 >>> from genshi.input import HTML
932
e53161c2773c Merge r1140 from py3k:
hodgestar
parents: 923
diff changeset
96 >>> html = HTML('''<p onclick="alert('Whoa')">Hello, world!</p>''', encoding='utf-8')
860
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
97 >>> print(html)
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
98 <p onclick="alert('Whoa')">Hello, world!</p>
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
99
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
100 A filter such as the HTML sanitizer can be applied to that stream using
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
101 the pipe notation as follows:
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
102
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
103 >>> from genshi.filters import HTMLSanitizer
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
104 >>> sanitizer = HTMLSanitizer()
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
105 >>> print(html | sanitizer)
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
106 <p>Hello, world!</p>
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
107
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
108 Filters can be any function that accepts and produces a stream (where
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
109 a stream is anything that iterates over events):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
110
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
111 >>> def uppercase(stream):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
112 ... for kind, data, pos in stream:
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
113 ... if kind is TEXT:
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
114 ... data = data.upper()
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
115 ... yield kind, data, pos
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
116 >>> print(html | sanitizer | uppercase)
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
117 <p>HELLO, WORLD!</p>
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
118
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
119 Serializers can also be used with this notation:
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
120
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
121 >>> from genshi.output import TextSerializer
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
122 >>> output = TextSerializer()
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
123 >>> print(html | sanitizer | uppercase | output)
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
124 HELLO, WORLD!
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
125
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
126 Commonly, serializers should be used at the end of the "pipeline";
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
127 using them somewhere in the middle may produce unexpected results.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
128
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
129 :param function: the callable object that should be applied as a filter
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
130 :return: the filtered stream
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
131 :rtype: `Stream`
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
132 """
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
133 return Stream(_ensure(function(self)), serializer=self.serializer)
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
134
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
135 def filter(self, *filters):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
136 """Apply filters to the stream.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
137
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
138 This method returns a new stream with the given filters applied. The
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
139 filters must be callables that accept the stream object as parameter,
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
140 and return the filtered stream.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
141
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
142 The call::
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
143
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
144 stream.filter(filter1, filter2)
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
145
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
146 is equivalent to::
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
147
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
148 stream | filter1 | filter2
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
149
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
150 :param filters: one or more callable objects that should be applied as
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
151 filters
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
152 :return: the filtered stream
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
153 :rtype: `Stream`
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
154 """
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
155 return reduce(operator.or_, (self,) + filters)
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
156
932
e53161c2773c Merge r1140 from py3k:
hodgestar
parents: 923
diff changeset
157 def render(self, method=None, encoding=None, out=None, **kwargs):
860
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
158 """Return a string representation of the stream.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
159
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
160 Any additional keyword arguments are passed to the serializer, and thus
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
161 depend on the `method` parameter value.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
162
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
163 :param method: determines how the stream is serialized; can be either
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
164 "xml", "xhtml", "html", "text", or a custom serializer
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
165 class; if `None`, the default serialization method of
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
166 the stream is used
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
167 :param encoding: how the output string should be encoded; if set to
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
168 `None`, this method returns a `unicode` object
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
169 :param out: a file-like object that the output should be written to
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
170 instead of being returned as one big string; note that if
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
171 this is a file or socket (or similar), the `encoding` must
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
172 not be `None` (that is, the output must be encoded)
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
173 :return: a `str` or `unicode` object (depending on the `encoding`
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
174 parameter), or `None` if the `out` parameter is provided
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
175 :rtype: `basestring`
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
176
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
177 :see: XMLSerializer, XHTMLSerializer, HTMLSerializer, TextSerializer
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
178 :note: Changed in 0.5: added the `out` parameter
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
179 """
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
180 from genshi.output import encode
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
181 if method is None:
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
182 method = self.serializer or 'xml'
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
183 generator = self.serialize(method=method, **kwargs)
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
184 return encode(generator, method=method, encoding=encoding, out=out)
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
185
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
186 def select(self, path, namespaces=None, variables=None):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
187 """Return a new stream that contains the events matching the given
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
188 XPath expression.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
189
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
190 >>> from genshi import HTML
932
e53161c2773c Merge r1140 from py3k:
hodgestar
parents: 923
diff changeset
191 >>> stream = HTML('<doc><elem>foo</elem><elem>bar</elem></doc>', encoding='utf-8')
860
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
192 >>> print(stream.select('elem'))
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
193 <elem>foo</elem><elem>bar</elem>
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
194 >>> print(stream.select('elem/text()'))
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
195 foobar
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
196
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
197 Note that the outermost element of the stream becomes the *context
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
198 node* for the XPath test. That means that the expression "doc" would
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
199 not match anything in the example above, because it only tests against
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
200 child elements of the outermost element:
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
201
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
202 >>> print(stream.select('doc'))
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
203 <BLANKLINE>
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
204
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
205 You can use the "." expression to match the context node itself
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
206 (although that usually makes little sense):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
207
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
208 >>> print(stream.select('.'))
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
209 <doc><elem>foo</elem><elem>bar</elem></doc>
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
210
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
211 :param path: a string containing the XPath expression
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
212 :param namespaces: mapping of namespace prefixes used in the path
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
213 :param variables: mapping of variable names to values
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
214 :return: the selected substream
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
215 :rtype: `Stream`
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
216 :raises PathSyntaxError: if the given path expression is invalid or not
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
217 supported
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
218 """
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
219 from genshi.path import Path
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
220 return Path(path).select(self, namespaces, variables)
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
221
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
222 def serialize(self, method='xml', **kwargs):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
223 """Generate strings corresponding to a specific serialization of the
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
224 stream.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
225
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
226 Unlike the `render()` method, this method is a generator that returns
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
227 the serialized output incrementally, as opposed to returning a single
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
228 string.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
229
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
230 Any additional keyword arguments are passed to the serializer, and thus
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
231 depend on the `method` parameter value.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
232
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
233 :param method: determines how the stream is serialized; can be either
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
234 "xml", "xhtml", "html", "text", or a custom serializer
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
235 class; if `None`, the default serialization method of
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
236 the stream is used
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
237 :return: an iterator over the serialization results (`Markup` or
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
238 `unicode` objects, depending on the serialization method)
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
239 :rtype: ``iterator``
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
240 :see: XMLSerializer, XHTMLSerializer, HTMLSerializer, TextSerializer
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
241 """
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
242 from genshi.output import get_serializer
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
243 if method is None:
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
244 method = self.serializer or 'xml'
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
245 return get_serializer(method, **kwargs)(_ensure(self))
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
246
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
247 def __str__(self):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
248 return self.render()
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
249
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
250 def __unicode__(self):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
251 return self.render(encoding=None)
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
252
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
253 def __html__(self):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
254 return self
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
255
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
256
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
257 START = Stream.START
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
258 END = Stream.END
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
259 TEXT = Stream.TEXT
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
260 XML_DECL = Stream.XML_DECL
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
261 DOCTYPE = Stream.DOCTYPE
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
262 START_NS = Stream.START_NS
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
263 END_NS = Stream.END_NS
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
264 START_CDATA = Stream.START_CDATA
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
265 END_CDATA = Stream.END_CDATA
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
266 PI = Stream.PI
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
267 COMMENT = Stream.COMMENT
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
268
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
269
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
270 def _ensure(stream):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
271 """Ensure that every item on the stream is actually a markup event."""
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
272 stream = iter(stream)
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
273 event = stream.next()
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
274
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
275 # Check whether the iterable is a real markup event stream by examining the
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
276 # first item it yields; if it's not we'll need to do some conversion
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
277 if type(event) is not tuple or len(event) != 3:
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
278 for event in chain([event], stream):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
279 if hasattr(event, 'totuple'):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
280 event = event.totuple()
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
281 else:
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
282 event = TEXT, unicode(event), (None, -1, -1)
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
283 yield event
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
284 return
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
285
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
286 # This looks like a markup event stream, so we'll just pass it through
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
287 # unchanged
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
288 yield event
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
289 for event in stream:
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
290 yield event
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
291
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
292
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
293 class Attrs(tuple):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
294 """Immutable sequence type that stores the attributes of an element.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
295
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
296 Ordering of the attributes is preserved, while access by name is also
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
297 supported.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
298
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
299 >>> attrs = Attrs([('href', '#'), ('title', 'Foo')])
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
300 >>> attrs
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
301 Attrs([('href', '#'), ('title', 'Foo')])
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
302
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
303 >>> 'href' in attrs
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
304 True
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
305 >>> 'tabindex' in attrs
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
306 False
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
307 >>> attrs.get('title')
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
308 'Foo'
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
309
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
310 Instances may not be manipulated directly. Instead, the operators ``|`` and
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
311 ``-`` can be used to produce new instances that have specific attributes
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
312 added, replaced or removed.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
313
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
314 To remove an attribute, use the ``-`` operator. The right hand side can be
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
315 either a string or a set/sequence of strings, identifying the name(s) of
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
316 the attribute(s) to remove:
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
317
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
318 >>> attrs - 'title'
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
319 Attrs([('href', '#')])
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
320 >>> attrs - ('title', 'href')
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
321 Attrs()
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
322
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
323 The original instance is not modified, but the operator can of course be
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
324 used with an assignment:
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
325
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
326 >>> attrs
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
327 Attrs([('href', '#'), ('title', 'Foo')])
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
328 >>> attrs -= 'title'
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
329 >>> attrs
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
330 Attrs([('href', '#')])
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
331
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
332 To add a new attribute, use the ``|`` operator, where the right hand value
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
333 is a sequence of ``(name, value)`` tuples (which includes `Attrs`
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
334 instances):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
335
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
336 >>> attrs | [('title', 'Bar')]
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
337 Attrs([('href', '#'), ('title', 'Bar')])
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
338
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
339 If the attributes already contain an attribute with a given name, the value
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
340 of that attribute is replaced:
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
341
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
342 >>> attrs | [('href', 'http://example.org/')]
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
343 Attrs([('href', 'http://example.org/')])
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
344 """
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
345 __slots__ = []
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
346
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
347 def __contains__(self, name):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
348 """Return whether the list includes an attribute with the specified
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
349 name.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
350
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
351 :return: `True` if the list includes the attribute
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
352 :rtype: `bool`
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
353 """
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
354 for attr, _ in self:
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
355 if attr == name:
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
356 return True
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
357
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
358 def __getitem__(self, i):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
359 """Return an item or slice of the attributes list.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
360
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
361 >>> attrs = Attrs([('href', '#'), ('title', 'Foo')])
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
362 >>> attrs[1]
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
363 ('title', 'Foo')
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
364 >>> attrs[1:]
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
365 Attrs([('title', 'Foo')])
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
366 """
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
367 items = tuple.__getitem__(self, i)
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
368 if type(i) is slice:
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
369 return Attrs(items)
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
370 return items
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
371
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
372 def __getslice__(self, i, j):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
373 """Return a slice of the attributes list.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
374
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
375 >>> attrs = Attrs([('href', '#'), ('title', 'Foo')])
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
376 >>> attrs[1:]
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
377 Attrs([('title', 'Foo')])
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
378 """
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
379 return Attrs(tuple.__getslice__(self, i, j))
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
380
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
381 def __or__(self, attrs):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
382 """Return a new instance that contains the attributes in `attrs` in
904
2772fa7c10a3 The `|` operator on `Attrs` objects now removes any existing attributes if the value is `None`.
cmlenz
parents: 862
diff changeset
383 addition to any already existing attributes. Any attributes in the new
2772fa7c10a3 The `|` operator on `Attrs` objects now removes any existing attributes if the value is `None`.
cmlenz
parents: 862
diff changeset
384 set that have a value of `None` are removed.
860
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
385
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
386 :return: a new instance with the merged attributes
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
387 :rtype: `Attrs`
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
388 """
904
2772fa7c10a3 The `|` operator on `Attrs` objects now removes any existing attributes if the value is `None`.
cmlenz
parents: 862
diff changeset
389 remove = set([an for an, av in attrs if av is None])
2772fa7c10a3 The `|` operator on `Attrs` objects now removes any existing attributes if the value is `None`.
cmlenz
parents: 862
diff changeset
390 replace = dict([(an, av) for an, av in attrs
2772fa7c10a3 The `|` operator on `Attrs` objects now removes any existing attributes if the value is `None`.
cmlenz
parents: 862
diff changeset
391 if an in self and av is not None])
2772fa7c10a3 The `|` operator on `Attrs` objects now removes any existing attributes if the value is `None`.
cmlenz
parents: 862
diff changeset
392 return Attrs([(sn, replace.get(sn, sv)) for sn, sv in self
2772fa7c10a3 The `|` operator on `Attrs` objects now removes any existing attributes if the value is `None`.
cmlenz
parents: 862
diff changeset
393 if sn not in remove] +
2772fa7c10a3 The `|` operator on `Attrs` objects now removes any existing attributes if the value is `None`.
cmlenz
parents: 862
diff changeset
394 [(an, av) for an, av in attrs
2772fa7c10a3 The `|` operator on `Attrs` objects now removes any existing attributes if the value is `None`.
cmlenz
parents: 862
diff changeset
395 if an not in self and an not in remove])
860
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
396
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
397 def __repr__(self):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
398 if not self:
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
399 return 'Attrs()'
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
400 return 'Attrs([%s])' % ', '.join([repr(item) for item in self])
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
401
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
402 def __sub__(self, names):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
403 """Return a new instance with all attributes with a name in `names` are
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
404 removed.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
405
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
406 :param names: the names of the attributes to remove
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
407 :return: a new instance with the attribute removed
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
408 :rtype: `Attrs`
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
409 """
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
410 if isinstance(names, basestring):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
411 names = (names,)
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
412 return Attrs([(name, val) for name, val in self if name not in names])
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
413
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
414 def get(self, name, default=None):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
415 """Return the value of the attribute with the specified name, or the
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
416 value of the `default` parameter if no such attribute is found.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
417
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
418 :param name: the name of the attribute
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
419 :param default: the value to return when the attribute does not exist
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
420 :return: the attribute value, or the `default` value if that attribute
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
421 does not exist
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
422 :rtype: `object`
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
423 """
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
424 for attr, value in self:
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
425 if attr == name:
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
426 return value
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
427 return default
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
428
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
429 def totuple(self):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
430 """Return the attributes as a markup event.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
431
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
432 The returned event is a `TEXT` event, the data is the value of all
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
433 attributes joined together.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
434
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
435 >>> Attrs([('href', '#'), ('title', 'Foo')]).totuple()
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
436 ('TEXT', '#Foo', (None, -1, -1))
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
437
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
438 :return: a `TEXT` event
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
439 :rtype: `tuple`
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
440 """
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
441 return TEXT, ''.join([x[1] for x in self]), (None, -1, -1)
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
442
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
443
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
444 class Markup(unicode):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
445 """Marks a string as being safe for inclusion in HTML/XML output without
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
446 needing to be escaped.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
447 """
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
448 __slots__ = []
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
449
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
450 def __add__(self, other):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
451 return Markup(unicode.__add__(self, escape(other)))
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
452
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
453 def __radd__(self, other):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
454 return Markup(unicode.__add__(escape(other), self))
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
455
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
456 def __mod__(self, args):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
457 if isinstance(args, dict):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
458 args = dict(zip(args.keys(), map(escape, args.values())))
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
459 elif isinstance(args, (list, tuple)):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
460 args = tuple(map(escape, args))
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
461 else:
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
462 args = escape(args)
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
463 return Markup(unicode.__mod__(self, args))
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
464
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
465 def __mul__(self, num):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
466 return Markup(unicode.__mul__(self, num))
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
467 __rmul__ = __mul__
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
468
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
469 def __repr__(self):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
470 return "<%s %s>" % (type(self).__name__, unicode.__repr__(self))
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
471
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
472 def join(self, seq, escape_quotes=True):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
473 """Return a `Markup` object which is the concatenation of the strings
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
474 in the given sequence, where this `Markup` object is the separator
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
475 between the joined elements.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
476
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
477 Any element in the sequence that is not a `Markup` instance is
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
478 automatically escaped.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
479
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
480 :param seq: the sequence of strings to join
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
481 :param escape_quotes: whether double quote characters in the elements
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
482 should be escaped
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
483 :return: the joined `Markup` object
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
484 :rtype: `Markup`
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
485 :see: `escape`
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
486 """
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
487 return Markup(unicode.join(self, [escape(item, quotes=escape_quotes)
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
488 for item in seq]))
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
489
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
490 @classmethod
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
491 def escape(cls, text, quotes=True):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
492 """Create a Markup instance from a string and escape special characters
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
493 it may contain (<, >, & and \").
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
494
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
495 >>> escape('"1 < 2"')
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
496 <Markup u'&#34;1 &lt; 2&#34;'>
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
497
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
498 If the `quotes` parameter is set to `False`, the \" character is left
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
499 as is. Escaping quotes is generally only required for strings that are
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
500 to be used in attribute values.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
501
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
502 >>> escape('"1 < 2"', quotes=False)
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
503 <Markup u'"1 &lt; 2"'>
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
504
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
505 :param text: the text to escape
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
506 :param quotes: if ``True``, double quote characters are escaped in
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
507 addition to the other special characters
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
508 :return: the escaped `Markup` string
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
509 :rtype: `Markup`
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
510 """
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
511 if not text:
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
512 return cls()
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
513 if type(text) is cls:
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
514 return text
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
515 if hasattr(text, '__html__'):
904
2772fa7c10a3 The `|` operator on `Attrs` objects now removes any existing attributes if the value is `None`.
cmlenz
parents: 862
diff changeset
516 return cls(text.__html__())
860
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
517
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
518 text = text.replace('&', '&amp;') \
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
519 .replace('<', '&lt;') \
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
520 .replace('>', '&gt;')
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
521 if quotes:
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
522 text = text.replace('"', '&#34;')
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
523 return cls(text)
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
524
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
525 def unescape(self):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
526 """Reverse-escapes &, <, >, and \" and returns a `unicode` object.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
527
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
528 >>> Markup('1 &lt; 2').unescape()
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
529 u'1 < 2'
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
530
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
531 :return: the unescaped string
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
532 :rtype: `unicode`
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
533 :see: `genshi.core.unescape`
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
534 """
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
535 if not self:
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
536 return ''
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
537 return unicode(self).replace('&#34;', '"') \
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
538 .replace('&gt;', '>') \
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
539 .replace('&lt;', '<') \
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
540 .replace('&amp;', '&')
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
541
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
542 def stripentities(self, keepxmlentities=False):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
543 """Return a copy of the text with any character or numeric entities
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
544 replaced by the equivalent UTF-8 characters.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
545
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
546 If the `keepxmlentities` parameter is provided and evaluates to `True`,
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
547 the core XML entities (``&amp;``, ``&apos;``, ``&gt;``, ``&lt;`` and
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
548 ``&quot;``) are not stripped.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
549
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
550 :return: a `Markup` instance with entities removed
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
551 :rtype: `Markup`
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
552 :see: `genshi.util.stripentities`
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
553 """
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
554 return Markup(stripentities(self, keepxmlentities=keepxmlentities))
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
555
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
556 def striptags(self):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
557 """Return a copy of the text with all XML/HTML tags removed.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
558
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
559 :return: a `Markup` instance with all tags removed
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
560 :rtype: `Markup`
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
561 :see: `genshi.util.striptags`
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
562 """
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
563 return Markup(striptags(self))
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
564
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
565
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
566 try:
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
567 from genshi._speedups import Markup
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
568 except ImportError:
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
569 pass # just use the Python implementation
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
570
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
571
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
572 escape = Markup.escape
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
573
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
574
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
575 def unescape(text):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
576 """Reverse-escapes &, <, >, and \" and returns a `unicode` object.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
577
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
578 >>> unescape(Markup('1 &lt; 2'))
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
579 u'1 < 2'
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
580
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
581 If the provided `text` object is not a `Markup` instance, it is returned
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
582 unchanged.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
583
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
584 >>> unescape('1 &lt; 2')
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
585 '1 &lt; 2'
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
586
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
587 :param text: the text to unescape
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
588 :return: the unescsaped string
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
589 :rtype: `unicode`
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
590 """
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
591 if not isinstance(text, Markup):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
592 return text
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
593 return text.unescape()
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
594
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
595
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
596 class Namespace(object):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
597 """Utility class creating and testing elements with a namespace.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
598
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
599 Internally, namespace URIs are encoded in the `QName` of any element or
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
600 attribute, the namespace URI being enclosed in curly braces. This class
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
601 helps create and test these strings.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
602
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
603 A `Namespace` object is instantiated with the namespace URI.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
604
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
605 >>> html = Namespace('http://www.w3.org/1999/xhtml')
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
606 >>> html
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
607 Namespace('http://www.w3.org/1999/xhtml')
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
608 >>> html.uri
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
609 u'http://www.w3.org/1999/xhtml'
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
610
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
611 The `Namespace` object can than be used to generate `QName` objects with
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
612 that namespace:
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
613
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
614 >>> html.body
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
615 QName('http://www.w3.org/1999/xhtml}body')
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
616 >>> html.body.localname
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
617 u'body'
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
618 >>> html.body.namespace
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
619 u'http://www.w3.org/1999/xhtml'
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
620
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
621 The same works using item access notation, which is useful for element or
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
622 attribute names that are not valid Python identifiers:
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
623
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
624 >>> html['body']
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
625 QName('http://www.w3.org/1999/xhtml}body')
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
626
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
627 A `Namespace` object can also be used to test whether a specific `QName`
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
628 belongs to that namespace using the ``in`` operator:
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
629
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
630 >>> qname = html.body
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
631 >>> qname in html
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
632 True
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
633 >>> qname in Namespace('http://www.w3.org/2002/06/xhtml2')
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
634 False
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
635 """
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
636 def __new__(cls, uri):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
637 if type(uri) is cls:
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
638 return uri
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
639 return object.__new__(cls)
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
640
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
641 def __getnewargs__(self):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
642 return (self.uri,)
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
643
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
644 def __getstate__(self):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
645 return self.uri
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
646
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
647 def __setstate__(self, uri):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
648 self.uri = uri
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
649
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
650 def __init__(self, uri):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
651 self.uri = unicode(uri)
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
652
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
653 def __contains__(self, qname):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
654 return qname.namespace == self.uri
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
655
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
656 def __ne__(self, other):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
657 return not self == other
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
658
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
659 def __eq__(self, other):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
660 if isinstance(other, Namespace):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
661 return self.uri == other.uri
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
662 return self.uri == other
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
663
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
664 def __getitem__(self, name):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
665 return QName(self.uri + '}' + name)
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
666 __getattr__ = __getitem__
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
667
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
668 def __hash__(self):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
669 return hash(self.uri)
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
670
932
e53161c2773c Merge r1140 from py3k:
hodgestar
parents: 923
diff changeset
671 if sys.version_info[0] == 2:
e53161c2773c Merge r1140 from py3k:
hodgestar
parents: 923
diff changeset
672 # Only use stringrepr in python 2
e53161c2773c Merge r1140 from py3k:
hodgestar
parents: 923
diff changeset
673 def __repr__(self):
e53161c2773c Merge r1140 from py3k:
hodgestar
parents: 923
diff changeset
674 return '%s(%s)' % (type(self).__name__, stringrepr(self.uri))
e53161c2773c Merge r1140 from py3k:
hodgestar
parents: 923
diff changeset
675 else:
e53161c2773c Merge r1140 from py3k:
hodgestar
parents: 923
diff changeset
676 def __repr__(self):
e53161c2773c Merge r1140 from py3k:
hodgestar
parents: 923
diff changeset
677 return '%s(%r)' % (type(self).__name__, self.uri)
860
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
678
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
679 def __str__(self):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
680 return self.uri.encode('utf-8')
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
681
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
682 def __unicode__(self):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
683 return self.uri
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
684
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
685
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
686 # The namespace used by attributes such as xml:lang and xml:space
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
687 XML_NAMESPACE = Namespace('http://www.w3.org/XML/1998/namespace')
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
688
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
689
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
690 class QName(unicode):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
691 """A qualified element or attribute name.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
692
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
693 The unicode value of instances of this class contains the qualified name of
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
694 the element or attribute, in the form ``{namespace-uri}local-name``. The
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
695 namespace URI can be obtained through the additional `namespace` attribute,
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
696 while the local name can be accessed through the `localname` attribute.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
697
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
698 >>> qname = QName('foo')
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
699 >>> qname
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
700 QName('foo')
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
701 >>> qname.localname
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
702 u'foo'
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
703 >>> qname.namespace
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
704
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
705 >>> qname = QName('http://www.w3.org/1999/xhtml}body')
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
706 >>> qname
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
707 QName('http://www.w3.org/1999/xhtml}body')
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
708 >>> qname.localname
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
709 u'body'
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
710 >>> qname.namespace
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
711 u'http://www.w3.org/1999/xhtml'
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
712 """
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
713 __slots__ = ['namespace', 'localname']
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
714
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
715 def __new__(cls, qname):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
716 """Create the `QName` instance.
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
717
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
718 :param qname: the qualified name as a string of the form
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
719 ``{namespace-uri}local-name``, where the leading curly
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
720 brace is optional
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
721 """
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
722 if type(qname) is cls:
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
723 return qname
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
724
923
8cef75b02ac1 Pull up r1145 to trunk.
jruigrok
parents: 904
diff changeset
725 qname = qname.lstrip('{')
8cef75b02ac1 Pull up r1145 to trunk.
jruigrok
parents: 904
diff changeset
726 parts = qname.split('}', 1)
860
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
727 if len(parts) > 1:
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
728 self = unicode.__new__(cls, '{%s' % qname)
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
729 self.namespace, self.localname = map(unicode, parts)
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
730 else:
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
731 self = unicode.__new__(cls, qname)
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
732 self.namespace, self.localname = None, unicode(qname)
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
733 return self
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
734
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
735 def __getnewargs__(self):
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
736 return (self.lstrip('{'),)
61d37796da98 A bit of cleanup of the `Markup` Python implementation.
cmlenz
parents: 857
diff changeset
737
932
e53161c2773c Merge r1140 from py3k:
hodgestar
parents: 923
diff changeset
738 if sys.version_info[0] == 2:
e53161c2773c Merge r1140 from py3k:
hodgestar
parents: 923
diff changeset
739 # Only use stringrepr in python 2
e53161c2773c Merge r1140 from py3k:
hodgestar
parents: 923
diff changeset
740 def __repr__(self):
e53161c2773c Merge r1140 from py3k:
hodgestar
parents: 923
diff changeset
741 return '%s(%s)' % (type(self).__name__, stringrepr(self.lstrip('{')))
e53161c2773c Merge r1140 from py3k:
hodgestar
parents: 923
diff changeset
742 else:
e53161c2773c Merge r1140 from py3k:
hodgestar
parents: 923
diff changeset
743 def __repr__(self):
e53161c2773c Merge r1140 from py3k:
hodgestar
parents: 923
diff changeset
744 return '%s(%r)' % (type(self).__name__, self.lstrip('{'))
Copyright (C) 2012-2017 Edgewall Software