Mercurial > genshi > genshi-test
annotate genshi/input.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 | fbe34d12acde |
children |
rev | line source |
---|---|
1 | 1 # -*- coding: utf-8 -*- |
2 # | |
854
0d9e87c6cf6e
More work on reducing the size of the diff produced by 2to3.
cmlenz
parents:
853
diff
changeset
|
3 # Copyright (C) 2006-2009 Edgewall Software |
1 | 4 # All rights reserved. |
5 # | |
6 # This software is licensed as described in the file COPYING, which | |
7 # you should have received as part of this distribution. The terms | |
230 | 8 # are also available at http://genshi.edgewall.org/wiki/License. |
1 | 9 # |
10 # This software consists of voluntary contributions made by many | |
11 # individuals. For the exact contribution history, see the revision | |
230 | 12 # history and logs, available at http://genshi.edgewall.org/log/. |
1 | 13 |
425
5b248708bbed
Try to use proper reStructuredText for docstrings throughout.
cmlenz
parents:
423
diff
changeset
|
14 """Support for constructing markup streams from files, strings, or other |
5b248708bbed
Try to use proper reStructuredText for docstrings throughout.
cmlenz
parents:
423
diff
changeset
|
15 sources. |
5b248708bbed
Try to use proper reStructuredText for docstrings throughout.
cmlenz
parents:
423
diff
changeset
|
16 """ |
5b248708bbed
Try to use proper reStructuredText for docstrings throughout.
cmlenz
parents:
423
diff
changeset
|
17 |
144
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
18 from itertools import chain |
859 | 19 import htmlentitydefs as entities |
20 import HTMLParser as html | |
1 | 21 from xml.parsers import expat |
22 | |
293
38adb4aa7df5
Fix a bug in the XML parser, where attributes containing HTML entity references would get pulled out of the attribute value, and instead added as a text node just before the associated start tag. Thanks to Hamish Lawson for [http://groups.google.com/group/genshi/browse_thread/thread/c64eb48676b0ff96/0e6ce786e8820f3d pointing out the problem].
cmlenz
parents:
290
diff
changeset
|
23 from genshi.core import Attrs, QName, Stream, stripentities |
859 | 24 from genshi.core import START, END, XML_DECL, DOCTYPE, TEXT, START_NS, \ |
25 END_NS, START_CDATA, END_CDATA, PI, COMMENT | |
932 | 26 from genshi.compat import StringIO, BytesIO |
27 | |
144
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
28 |
290
a6738047c85e
Move the ElementTree ''element-to-stream'' adaptation function `ET()` into the `genshi.input` module.
cmlenz
parents:
230
diff
changeset
|
29 __all__ = ['ET', 'ParseError', 'XMLParser', 'XML', 'HTMLParser', 'HTML'] |
425
5b248708bbed
Try to use proper reStructuredText for docstrings throughout.
cmlenz
parents:
423
diff
changeset
|
30 __docformat__ = 'restructuredtext en' |
290
a6738047c85e
Move the ElementTree ''element-to-stream'' adaptation function `ET()` into the `genshi.input` module.
cmlenz
parents:
230
diff
changeset
|
31 |
859 | 32 |
290
a6738047c85e
Move the ElementTree ''element-to-stream'' adaptation function `ET()` into the `genshi.input` module.
cmlenz
parents:
230
diff
changeset
|
33 def ET(element): |
433 | 34 """Convert a given ElementTree element to a markup stream. |
35 | |
36 :param element: an ElementTree element | |
37 :return: a markup stream | |
38 """ | |
290
a6738047c85e
Move the ElementTree ''element-to-stream'' adaptation function `ET()` into the `genshi.input` module.
cmlenz
parents:
230
diff
changeset
|
39 tag_name = QName(element.tag.lstrip('{')) |
458
160f787cc818
The `ET()` function now correctly handles attributes with a namespace.
cmlenz
parents:
434
diff
changeset
|
40 attrs = Attrs([(QName(attr.lstrip('{')), value) |
160f787cc818
The `ET()` function now correctly handles attributes with a namespace.
cmlenz
parents:
434
diff
changeset
|
41 for attr, value in element.items()]) |
290
a6738047c85e
Move the ElementTree ''element-to-stream'' adaptation function `ET()` into the `genshi.input` module.
cmlenz
parents:
230
diff
changeset
|
42 |
a6738047c85e
Move the ElementTree ''element-to-stream'' adaptation function `ET()` into the `genshi.input` module.
cmlenz
parents:
230
diff
changeset
|
43 yield START, (tag_name, attrs), (None, -1, -1) |
a6738047c85e
Move the ElementTree ''element-to-stream'' adaptation function `ET()` into the `genshi.input` module.
cmlenz
parents:
230
diff
changeset
|
44 if element.text: |
a6738047c85e
Move the ElementTree ''element-to-stream'' adaptation function `ET()` into the `genshi.input` module.
cmlenz
parents:
230
diff
changeset
|
45 yield TEXT, element.text, (None, -1, -1) |
a6738047c85e
Move the ElementTree ''element-to-stream'' adaptation function `ET()` into the `genshi.input` module.
cmlenz
parents:
230
diff
changeset
|
46 for child in element.getchildren(): |
a6738047c85e
Move the ElementTree ''element-to-stream'' adaptation function `ET()` into the `genshi.input` module.
cmlenz
parents:
230
diff
changeset
|
47 for item in ET(child): |
a6738047c85e
Move the ElementTree ''element-to-stream'' adaptation function `ET()` into the `genshi.input` module.
cmlenz
parents:
230
diff
changeset
|
48 yield item |
a6738047c85e
Move the ElementTree ''element-to-stream'' adaptation function `ET()` into the `genshi.input` module.
cmlenz
parents:
230
diff
changeset
|
49 yield END, tag_name, (None, -1, -1) |
a6738047c85e
Move the ElementTree ''element-to-stream'' adaptation function `ET()` into the `genshi.input` module.
cmlenz
parents:
230
diff
changeset
|
50 if element.tail: |
a6738047c85e
Move the ElementTree ''element-to-stream'' adaptation function `ET()` into the `genshi.input` module.
cmlenz
parents:
230
diff
changeset
|
51 yield TEXT, element.tail, (None, -1, -1) |
1 | 52 |
53 | |
21
eca77129518a
* Include paths are now interpreted relative to the path of the including template. Closes #3.
cmlenz
parents:
1
diff
changeset
|
54 class ParseError(Exception): |
eca77129518a
* Include paths are now interpreted relative to the path of the including template. Closes #3.
cmlenz
parents:
1
diff
changeset
|
55 """Exception raised when fatal syntax errors are found in the input being |
433 | 56 parsed. |
57 """ | |
21
eca77129518a
* Include paths are now interpreted relative to the path of the including template. Closes #3.
cmlenz
parents:
1
diff
changeset
|
58 |
422
95089b6e37ca
More work to include absolute file paths in exceptions.
cmlenz
parents:
419
diff
changeset
|
59 def __init__(self, message, filename=None, lineno=-1, offset=-1): |
433 | 60 """Exception initializer. |
61 | |
62 :param message: the error message from the parser | |
63 :param filename: the path to the file that was parsed | |
64 :param lineno: the number of the line on which the error was encountered | |
65 :param offset: the column number where the error was encountered | |
66 """ | |
422
95089b6e37ca
More work to include absolute file paths in exceptions.
cmlenz
parents:
419
diff
changeset
|
67 self.msg = message |
95089b6e37ca
More work to include absolute file paths in exceptions.
cmlenz
parents:
419
diff
changeset
|
68 if filename: |
434
e065d7906b68
* Better method to propogate the full path to the template file on parse errors. Supersedes r513.
cmlenz
parents:
433
diff
changeset
|
69 message += ', in ' + filename |
21
eca77129518a
* Include paths are now interpreted relative to the path of the including template. Closes #3.
cmlenz
parents:
1
diff
changeset
|
70 Exception.__init__(self, message) |
422
95089b6e37ca
More work to include absolute file paths in exceptions.
cmlenz
parents:
419
diff
changeset
|
71 self.filename = filename or '<string>' |
21
eca77129518a
* Include paths are now interpreted relative to the path of the including template. Closes #3.
cmlenz
parents:
1
diff
changeset
|
72 self.lineno = lineno |
eca77129518a
* Include paths are now interpreted relative to the path of the including template. Closes #3.
cmlenz
parents:
1
diff
changeset
|
73 self.offset = offset |
eca77129518a
* Include paths are now interpreted relative to the path of the including template. Closes #3.
cmlenz
parents:
1
diff
changeset
|
74 |
eca77129518a
* Include paths are now interpreted relative to the path of the including template. Closes #3.
cmlenz
parents:
1
diff
changeset
|
75 |
1 | 76 class XMLParser(object): |
77 """Generator-based XML parser based on roughly equivalent code in | |
26
039fc5b87405
* Split out the XPath tests into a separate `unittest`-based file.
cmlenz
parents:
21
diff
changeset
|
78 Kid/ElementTree. |
039fc5b87405
* Split out the XPath tests into a separate `unittest`-based file.
cmlenz
parents:
21
diff
changeset
|
79 |
039fc5b87405
* Split out the XPath tests into a separate `unittest`-based file.
cmlenz
parents:
21
diff
changeset
|
80 The parsing is initiated by iterating over the parser object: |
039fc5b87405
* Split out the XPath tests into a separate `unittest`-based file.
cmlenz
parents:
21
diff
changeset
|
81 |
039fc5b87405
* Split out the XPath tests into a separate `unittest`-based file.
cmlenz
parents:
21
diff
changeset
|
82 >>> parser = XMLParser(StringIO('<root id="2"><child>Foo</child></root>')) |
039fc5b87405
* Split out the XPath tests into a separate `unittest`-based file.
cmlenz
parents:
21
diff
changeset
|
83 >>> for kind, data, pos in parser: |
853
4376010bb97e
Convert a bunch of print statements to py3k compatible syntax.
cmlenz
parents:
852
diff
changeset
|
84 ... print('%s %s' % (kind, data)) |
857
24733a5854d9
Avoid unicode literals in `repr`s of `QName` and `Namespace` when not necessary.
cmlenz
parents:
856
diff
changeset
|
85 START (QName('root'), Attrs([(QName('id'), u'2')])) |
24733a5854d9
Avoid unicode literals in `repr`s of `QName` and `Namespace` when not necessary.
cmlenz
parents:
856
diff
changeset
|
86 START (QName('child'), Attrs()) |
26
039fc5b87405
* Split out the XPath tests into a separate `unittest`-based file.
cmlenz
parents:
21
diff
changeset
|
87 TEXT Foo |
039fc5b87405
* Split out the XPath tests into a separate `unittest`-based file.
cmlenz
parents:
21
diff
changeset
|
88 END child |
039fc5b87405
* Split out the XPath tests into a separate `unittest`-based file.
cmlenz
parents:
21
diff
changeset
|
89 END root |
039fc5b87405
* Split out the XPath tests into a separate `unittest`-based file.
cmlenz
parents:
21
diff
changeset
|
90 """ |
1 | 91 |
293
38adb4aa7df5
Fix a bug in the XML parser, where attributes containing HTML entity references would get pulled out of the attribute value, and instead added as a text node just before the associated start tag. Thanks to Hamish Lawson for [http://groups.google.com/group/genshi/browse_thread/thread/c64eb48676b0ff96/0e6ce786e8820f3d pointing out the problem].
cmlenz
parents:
290
diff
changeset
|
92 _entitydefs = ['<!ENTITY %s "&#%d;">' % (name, value) for name, value in |
856 | 93 entities.name2codepoint.items()] |
932 | 94 _external_dtd = u'\n'.join(_entitydefs).encode('utf-8') |
293
38adb4aa7df5
Fix a bug in the XML parser, where attributes containing HTML entity references would get pulled out of the attribute value, and instead added as a text node just before the associated start tag. Thanks to Hamish Lawson for [http://groups.google.com/group/genshi/browse_thread/thread/c64eb48676b0ff96/0e6ce786e8820f3d pointing out the problem].
cmlenz
parents:
290
diff
changeset
|
95 |
316
4ab9edf5e83b
Configurable encoding of template files, closing #65.
cmlenz
parents:
312
diff
changeset
|
96 def __init__(self, source, filename=None, encoding=None): |
311
01e2c48f6dfb
* The `HTMLParser` class and the `HTML` function now accept an `encoding` parameter to properly deal with bytestring input (defaults to UTF-8).
cmlenz
parents:
293
diff
changeset
|
97 """Initialize the parser for the given XML input. |
26
039fc5b87405
* Split out the XPath tests into a separate `unittest`-based file.
cmlenz
parents:
21
diff
changeset
|
98 |
425
5b248708bbed
Try to use proper reStructuredText for docstrings throughout.
cmlenz
parents:
423
diff
changeset
|
99 :param source: the XML text as a file-like object |
5b248708bbed
Try to use proper reStructuredText for docstrings throughout.
cmlenz
parents:
423
diff
changeset
|
100 :param filename: the name of the file, if appropriate |
5b248708bbed
Try to use proper reStructuredText for docstrings throughout.
cmlenz
parents:
423
diff
changeset
|
101 :param encoding: the encoding of the file; if not specified, the |
5b248708bbed
Try to use proper reStructuredText for docstrings throughout.
cmlenz
parents:
423
diff
changeset
|
102 encoding is assumed to be ASCII, UTF-8, or UTF-16, or |
5b248708bbed
Try to use proper reStructuredText for docstrings throughout.
cmlenz
parents:
423
diff
changeset
|
103 whatever the encoding specified in the XML declaration |
5b248708bbed
Try to use proper reStructuredText for docstrings throughout.
cmlenz
parents:
423
diff
changeset
|
104 (if any) |
26
039fc5b87405
* Split out the XPath tests into a separate `unittest`-based file.
cmlenz
parents:
21
diff
changeset
|
105 """ |
1 | 106 self.source = source |
21
eca77129518a
* Include paths are now interpreted relative to the path of the including template. Closes #3.
cmlenz
parents:
1
diff
changeset
|
107 self.filename = filename |
1 | 108 |
109 # Setup the Expat parser | |
316
4ab9edf5e83b
Configurable encoding of template files, closing #65.
cmlenz
parents:
312
diff
changeset
|
110 parser = expat.ParserCreate(encoding, '}') |
1 | 111 parser.buffer_text = True |
932 | 112 # Python 3 does not have returns_unicode |
113 if hasattr(parser, 'returns_unicode'): | |
114 parser.returns_unicode = True | |
160 | 115 parser.ordered_attributes = True |
116 | |
1 | 117 parser.StartElementHandler = self._handle_start |
118 parser.EndElementHandler = self._handle_end | |
119 parser.CharacterDataHandler = self._handle_data | |
120 parser.StartDoctypeDeclHandler = self._handle_doctype | |
121 parser.StartNamespaceDeclHandler = self._handle_start_ns | |
122 parser.EndNamespaceDeclHandler = self._handle_end_ns | |
143
ef761afcedff
CDATA sections in XML input now appear as CDATA sections in the output. This should address the problem with escaping the contents of `<style>` and `<script>` elements, which would only get interpreted correctly if the output was served as `application/xhtml+xml`. Closes #24.
cmlenz
parents:
140
diff
changeset
|
123 parser.StartCdataSectionHandler = self._handle_start_cdata |
ef761afcedff
CDATA sections in XML input now appear as CDATA sections in the output. This should address the problem with escaping the contents of `<style>` and `<script>` elements, which would only get interpreted correctly if the output was served as `application/xhtml+xml`. Closes #24.
cmlenz
parents:
140
diff
changeset
|
124 parser.EndCdataSectionHandler = self._handle_end_cdata |
1 | 125 parser.ProcessingInstructionHandler = self._handle_pi |
460
6b5544bb5a99
Apply patch by Alec Thomas for processing XML declarations (#111). Thanks!
cmlenz
parents:
458
diff
changeset
|
126 parser.XmlDeclHandler = self._handle_xml_decl |
1 | 127 parser.CommentHandler = self._handle_comment |
209
5b422db07359
* Fix bug in handling of undefined entities. Thanks to Arnar for reporting the issue on IRC.
cmlenz
parents:
207
diff
changeset
|
128 |
5b422db07359
* Fix bug in handling of undefined entities. Thanks to Arnar for reporting the issue on IRC.
cmlenz
parents:
207
diff
changeset
|
129 # Tell Expat that we'll handle non-XML entities ourselves |
5b422db07359
* Fix bug in handling of undefined entities. Thanks to Arnar for reporting the issue on IRC.
cmlenz
parents:
207
diff
changeset
|
130 # (in _handle_other) |
1 | 131 parser.DefaultHandler = self._handle_other |
293
38adb4aa7df5
Fix a bug in the XML parser, where attributes containing HTML entity references would get pulled out of the attribute value, and instead added as a text node just before the associated start tag. Thanks to Hamish Lawson for [http://groups.google.com/group/genshi/browse_thread/thread/c64eb48676b0ff96/0e6ce786e8820f3d pointing out the problem].
cmlenz
parents:
290
diff
changeset
|
132 parser.SetParamEntityParsing(expat.XML_PARAM_ENTITY_PARSING_ALWAYS) |
209
5b422db07359
* Fix bug in handling of undefined entities. Thanks to Arnar for reporting the issue on IRC.
cmlenz
parents:
207
diff
changeset
|
133 parser.UseForeignDTD() |
293
38adb4aa7df5
Fix a bug in the XML parser, where attributes containing HTML entity references would get pulled out of the attribute value, and instead added as a text node just before the associated start tag. Thanks to Hamish Lawson for [http://groups.google.com/group/genshi/browse_thread/thread/c64eb48676b0ff96/0e6ce786e8820f3d pointing out the problem].
cmlenz
parents:
290
diff
changeset
|
134 parser.ExternalEntityRefHandler = self._build_foreign |
1 | 135 |
136 self.expat = parser | |
21
eca77129518a
* Include paths are now interpreted relative to the path of the including template. Closes #3.
cmlenz
parents:
1
diff
changeset
|
137 self._queue = [] |
1 | 138 |
144
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
139 def parse(self): |
433 | 140 """Generator that parses the XML source, yielding markup events. |
141 | |
142 :return: a markup event stream | |
143 :raises ParseError: if the XML text is not well formed | |
144 """ | |
144
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
145 def _generate(): |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
146 try: |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
147 bufsize = 4 * 1024 # 4K |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
148 done = False |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
149 while 1: |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
150 while not done and len(self._queue) == 0: |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
151 data = self.source.read(bufsize) |
932 | 152 if not data: # end of data |
144
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
153 if hasattr(self, 'expat'): |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
154 self.expat.Parse('', True) |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
155 del self.expat # get rid of circular references |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
156 done = True |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
157 else: |
207
0619a27f5e67
The `XMLParser` now correctly handles unicode input. Closes #43.
cmlenz
parents:
182
diff
changeset
|
158 if isinstance(data, unicode): |
0619a27f5e67
The `XMLParser` now correctly handles unicode input. Closes #43.
cmlenz
parents:
182
diff
changeset
|
159 data = data.encode('utf-8') |
144
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
160 self.expat.Parse(data, False) |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
161 for event in self._queue: |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
162 yield event |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
163 self._queue = [] |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
164 if done: |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
165 break |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
166 except expat.ExpatError, e: |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
167 msg = str(e) |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
168 raise ParseError(msg, self.filename, e.lineno, e.offset) |
146 | 169 return Stream(_generate()).filter(_coalesce) |
144
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
170 |
1 | 171 def __iter__(self): |
144
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
172 return iter(self.parse()) |
1 | 173 |
293
38adb4aa7df5
Fix a bug in the XML parser, where attributes containing HTML entity references would get pulled out of the attribute value, and instead added as a text node just before the associated start tag. Thanks to Hamish Lawson for [http://groups.google.com/group/genshi/browse_thread/thread/c64eb48676b0ff96/0e6ce786e8820f3d pointing out the problem].
cmlenz
parents:
290
diff
changeset
|
174 def _build_foreign(self, context, base, sysid, pubid): |
38adb4aa7df5
Fix a bug in the XML parser, where attributes containing HTML entity references would get pulled out of the attribute value, and instead added as a text node just before the associated start tag. Thanks to Hamish Lawson for [http://groups.google.com/group/genshi/browse_thread/thread/c64eb48676b0ff96/0e6ce786e8820f3d pointing out the problem].
cmlenz
parents:
290
diff
changeset
|
175 parser = self.expat.ExternalEntityParserCreate(context) |
932 | 176 parser.ParseFile(BytesIO(self._external_dtd)) |
293
38adb4aa7df5
Fix a bug in the XML parser, where attributes containing HTML entity references would get pulled out of the attribute value, and instead added as a text node just before the associated start tag. Thanks to Hamish Lawson for [http://groups.google.com/group/genshi/browse_thread/thread/c64eb48676b0ff96/0e6ce786e8820f3d pointing out the problem].
cmlenz
parents:
290
diff
changeset
|
177 return 1 |
38adb4aa7df5
Fix a bug in the XML parser, where attributes containing HTML entity references would get pulled out of the attribute value, and instead added as a text node just before the associated start tag. Thanks to Hamish Lawson for [http://groups.google.com/group/genshi/browse_thread/thread/c64eb48676b0ff96/0e6ce786e8820f3d pointing out the problem].
cmlenz
parents:
290
diff
changeset
|
178 |
143
ef761afcedff
CDATA sections in XML input now appear as CDATA sections in the output. This should address the problem with escaping the contents of `<style>` and `<script>` elements, which would only get interpreted correctly if the output was served as `application/xhtml+xml`. Closes #24.
cmlenz
parents:
140
diff
changeset
|
179 def _enqueue(self, kind, data=None, pos=None): |
26
039fc5b87405
* Split out the XPath tests into a separate `unittest`-based file.
cmlenz
parents:
21
diff
changeset
|
180 if pos is None: |
039fc5b87405
* Split out the XPath tests into a separate `unittest`-based file.
cmlenz
parents:
21
diff
changeset
|
181 pos = self._getpos() |
144
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
182 if kind is TEXT: |
134
df44110ca91d
* Improve the accuracy of line numbers for text nodes, so that reported errors about syntax or evaluation errors in expressions point to the right line (not quite perfect yet, though).
cmlenz
parents:
69
diff
changeset
|
183 # Expat reports the *end* of the text event as current position. We |
df44110ca91d
* Improve the accuracy of line numbers for text nodes, so that reported errors about syntax or evaluation errors in expressions point to the right line (not quite perfect yet, though).
cmlenz
parents:
69
diff
changeset
|
184 # try to fix that up here as much as possible. Unfortunately, the |
df44110ca91d
* Improve the accuracy of line numbers for text nodes, so that reported errors about syntax or evaluation errors in expressions point to the right line (not quite perfect yet, though).
cmlenz
parents:
69
diff
changeset
|
185 # offset is only valid for single-line text. For multi-line text, |
df44110ca91d
* Improve the accuracy of line numbers for text nodes, so that reported errors about syntax or evaluation errors in expressions point to the right line (not quite perfect yet, though).
cmlenz
parents:
69
diff
changeset
|
186 # it is apparently not possible to determine at what offset it |
df44110ca91d
* Improve the accuracy of line numbers for text nodes, so that reported errors about syntax or evaluation errors in expressions point to the right line (not quite perfect yet, though).
cmlenz
parents:
69
diff
changeset
|
187 # started |
df44110ca91d
* Improve the accuracy of line numbers for text nodes, so that reported errors about syntax or evaluation errors in expressions point to the right line (not quite perfect yet, though).
cmlenz
parents:
69
diff
changeset
|
188 if '\n' in data: |
df44110ca91d
* Improve the accuracy of line numbers for text nodes, so that reported errors about syntax or evaluation errors in expressions point to the right line (not quite perfect yet, though).
cmlenz
parents:
69
diff
changeset
|
189 lines = data.splitlines() |
df44110ca91d
* Improve the accuracy of line numbers for text nodes, so that reported errors about syntax or evaluation errors in expressions point to the right line (not quite perfect yet, though).
cmlenz
parents:
69
diff
changeset
|
190 lineno = pos[1] - len(lines) + 1 |
df44110ca91d
* Improve the accuracy of line numbers for text nodes, so that reported errors about syntax or evaluation errors in expressions point to the right line (not quite perfect yet, though).
cmlenz
parents:
69
diff
changeset
|
191 offset = -1 |
df44110ca91d
* Improve the accuracy of line numbers for text nodes, so that reported errors about syntax or evaluation errors in expressions point to the right line (not quite perfect yet, though).
cmlenz
parents:
69
diff
changeset
|
192 else: |
df44110ca91d
* Improve the accuracy of line numbers for text nodes, so that reported errors about syntax or evaluation errors in expressions point to the right line (not quite perfect yet, though).
cmlenz
parents:
69
diff
changeset
|
193 lineno = pos[1] |
df44110ca91d
* Improve the accuracy of line numbers for text nodes, so that reported errors about syntax or evaluation errors in expressions point to the right line (not quite perfect yet, though).
cmlenz
parents:
69
diff
changeset
|
194 offset = pos[2] - len(data) |
df44110ca91d
* Improve the accuracy of line numbers for text nodes, so that reported errors about syntax or evaluation errors in expressions point to the right line (not quite perfect yet, though).
cmlenz
parents:
69
diff
changeset
|
195 pos = (pos[0], lineno, offset) |
26
039fc5b87405
* Split out the XPath tests into a separate `unittest`-based file.
cmlenz
parents:
21
diff
changeset
|
196 self._queue.append((kind, data, pos)) |
039fc5b87405
* Split out the XPath tests into a separate `unittest`-based file.
cmlenz
parents:
21
diff
changeset
|
197 |
1 | 198 def _getpos_unknown(self): |
134
df44110ca91d
* Improve the accuracy of line numbers for text nodes, so that reported errors about syntax or evaluation errors in expressions point to the right line (not quite perfect yet, though).
cmlenz
parents:
69
diff
changeset
|
199 return (self.filename, -1, -1) |
1 | 200 |
21
eca77129518a
* Include paths are now interpreted relative to the path of the including template. Closes #3.
cmlenz
parents:
1
diff
changeset
|
201 def _getpos(self): |
134
df44110ca91d
* Improve the accuracy of line numbers for text nodes, so that reported errors about syntax or evaluation errors in expressions point to the right line (not quite perfect yet, though).
cmlenz
parents:
69
diff
changeset
|
202 return (self.filename, self.expat.CurrentLineNumber, |
21
eca77129518a
* Include paths are now interpreted relative to the path of the including template. Closes #3.
cmlenz
parents:
1
diff
changeset
|
203 self.expat.CurrentColumnNumber) |
1 | 204 |
205 def _handle_start(self, tag, attrib): | |
403
32b283e1d310
Remove some magic/overhead from `Attrs` creation and manipulation by not automatically wrapping attribute names in `QName`.
cmlenz
parents:
378
diff
changeset
|
206 attrs = Attrs([(QName(name), value) for name, value in |
32b283e1d310
Remove some magic/overhead from `Attrs` creation and manipulation by not automatically wrapping attribute names in `QName`.
cmlenz
parents:
378
diff
changeset
|
207 zip(*[iter(attrib)] * 2)]) |
32b283e1d310
Remove some magic/overhead from `Attrs` creation and manipulation by not automatically wrapping attribute names in `QName`.
cmlenz
parents:
378
diff
changeset
|
208 self._enqueue(START, (QName(tag), attrs)) |
1 | 209 |
210 def _handle_end(self, tag): | |
144
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
211 self._enqueue(END, QName(tag)) |
1 | 212 |
213 def _handle_data(self, text): | |
144
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
214 self._enqueue(TEXT, text) |
1 | 215 |
460
6b5544bb5a99
Apply patch by Alec Thomas for processing XML declarations (#111). Thanks!
cmlenz
parents:
458
diff
changeset
|
216 def _handle_xml_decl(self, version, encoding, standalone): |
6b5544bb5a99
Apply patch by Alec Thomas for processing XML declarations (#111). Thanks!
cmlenz
parents:
458
diff
changeset
|
217 self._enqueue(XML_DECL, (version, encoding, standalone)) |
6b5544bb5a99
Apply patch by Alec Thomas for processing XML declarations (#111). Thanks!
cmlenz
parents:
458
diff
changeset
|
218 |
1 | 219 def _handle_doctype(self, name, sysid, pubid, has_internal_subset): |
144
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
220 self._enqueue(DOCTYPE, (name, pubid, sysid)) |
1 | 221 |
222 def _handle_start_ns(self, prefix, uri): | |
144
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
223 self._enqueue(START_NS, (prefix or '', uri)) |
1 | 224 |
225 def _handle_end_ns(self, prefix): | |
144
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
226 self._enqueue(END_NS, prefix or '') |
1 | 227 |
143
ef761afcedff
CDATA sections in XML input now appear as CDATA sections in the output. This should address the problem with escaping the contents of `<style>` and `<script>` elements, which would only get interpreted correctly if the output was served as `application/xhtml+xml`. Closes #24.
cmlenz
parents:
140
diff
changeset
|
228 def _handle_start_cdata(self): |
144
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
229 self._enqueue(START_CDATA) |
143
ef761afcedff
CDATA sections in XML input now appear as CDATA sections in the output. This should address the problem with escaping the contents of `<style>` and `<script>` elements, which would only get interpreted correctly if the output was served as `application/xhtml+xml`. Closes #24.
cmlenz
parents:
140
diff
changeset
|
230 |
ef761afcedff
CDATA sections in XML input now appear as CDATA sections in the output. This should address the problem with escaping the contents of `<style>` and `<script>` elements, which would only get interpreted correctly if the output was served as `application/xhtml+xml`. Closes #24.
cmlenz
parents:
140
diff
changeset
|
231 def _handle_end_cdata(self): |
144
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
232 self._enqueue(END_CDATA) |
143
ef761afcedff
CDATA sections in XML input now appear as CDATA sections in the output. This should address the problem with escaping the contents of `<style>` and `<script>` elements, which would only get interpreted correctly if the output was served as `application/xhtml+xml`. Closes #24.
cmlenz
parents:
140
diff
changeset
|
233 |
1 | 234 def _handle_pi(self, target, data): |
144
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
235 self._enqueue(PI, (target, data)) |
1 | 236 |
237 def _handle_comment(self, text): | |
144
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
238 self._enqueue(COMMENT, text) |
1 | 239 |
240 def _handle_other(self, text): | |
241 if text.startswith('&'): | |
242 # deal with undefined entities | |
243 try: | |
856 | 244 text = unichr(entities.name2codepoint[text[1:-1]]) |
144
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
245 self._enqueue(TEXT, text) |
1 | 246 except KeyError: |
209
5b422db07359
* Fix bug in handling of undefined entities. Thanks to Arnar for reporting the issue on IRC.
cmlenz
parents:
207
diff
changeset
|
247 filename, lineno, offset = self._getpos() |
5b422db07359
* Fix bug in handling of undefined entities. Thanks to Arnar for reporting the issue on IRC.
cmlenz
parents:
207
diff
changeset
|
248 error = expat.error('undefined entity "%s": line %d, column %d' |
5b422db07359
* Fix bug in handling of undefined entities. Thanks to Arnar for reporting the issue on IRC.
cmlenz
parents:
207
diff
changeset
|
249 % (text, lineno, offset)) |
5b422db07359
* Fix bug in handling of undefined entities. Thanks to Arnar for reporting the issue on IRC.
cmlenz
parents:
207
diff
changeset
|
250 error.code = expat.errors.XML_ERROR_UNDEFINED_ENTITY |
5b422db07359
* Fix bug in handling of undefined entities. Thanks to Arnar for reporting the issue on IRC.
cmlenz
parents:
207
diff
changeset
|
251 error.lineno = lineno |
5b422db07359
* Fix bug in handling of undefined entities. Thanks to Arnar for reporting the issue on IRC.
cmlenz
parents:
207
diff
changeset
|
252 error.offset = offset |
5b422db07359
* Fix bug in handling of undefined entities. Thanks to Arnar for reporting the issue on IRC.
cmlenz
parents:
207
diff
changeset
|
253 raise error |
1 | 254 |
255 | |
256 def XML(text): | |
433 | 257 """Parse the given XML source and return a markup stream. |
258 | |
259 Unlike with `XMLParser`, the returned stream is reusable, meaning it can be | |
260 iterated over multiple times: | |
261 | |
262 >>> xml = XML('<doc><elem>Foo</elem><elem>Bar</elem></doc>') | |
853
4376010bb97e
Convert a bunch of print statements to py3k compatible syntax.
cmlenz
parents:
852
diff
changeset
|
263 >>> print(xml) |
433 | 264 <doc><elem>Foo</elem><elem>Bar</elem></doc> |
853
4376010bb97e
Convert a bunch of print statements to py3k compatible syntax.
cmlenz
parents:
852
diff
changeset
|
265 >>> print(xml.select('elem')) |
433 | 266 <elem>Foo</elem><elem>Bar</elem> |
853
4376010bb97e
Convert a bunch of print statements to py3k compatible syntax.
cmlenz
parents:
852
diff
changeset
|
267 >>> print(xml.select('elem/text()')) |
433 | 268 FooBar |
269 | |
270 :param text: the XML source | |
271 :return: the parsed XML event stream | |
272 :raises ParseError: if the XML text is not well-formed | |
273 """ | |
859 | 274 return Stream(list(XMLParser(StringIO(text)))) |
1 | 275 |
276 | |
21
eca77129518a
* Include paths are now interpreted relative to the path of the including template. Closes #3.
cmlenz
parents:
1
diff
changeset
|
277 class HTMLParser(html.HTMLParser, object): |
1 | 278 """Parser for HTML input based on the Python `HTMLParser` module. |
279 | |
280 This class provides the same interface for generating stream events as | |
281 `XMLParser`, and attempts to automatically balance tags. | |
26
039fc5b87405
* Split out the XPath tests into a separate `unittest`-based file.
cmlenz
parents:
21
diff
changeset
|
282 |
039fc5b87405
* Split out the XPath tests into a separate `unittest`-based file.
cmlenz
parents:
21
diff
changeset
|
283 The parsing is initiated by iterating over the parser object: |
039fc5b87405
* Split out the XPath tests into a separate `unittest`-based file.
cmlenz
parents:
21
diff
changeset
|
284 |
932 | 285 >>> parser = HTMLParser(BytesIO(u'<UL compact><LI>Foo</UL>'.encode('utf-8')), encoding='utf-8') |
26
039fc5b87405
* Split out the XPath tests into a separate `unittest`-based file.
cmlenz
parents:
21
diff
changeset
|
286 >>> for kind, data, pos in parser: |
853
4376010bb97e
Convert a bunch of print statements to py3k compatible syntax.
cmlenz
parents:
852
diff
changeset
|
287 ... print('%s %s' % (kind, data)) |
857
24733a5854d9
Avoid unicode literals in `repr`s of `QName` and `Namespace` when not necessary.
cmlenz
parents:
856
diff
changeset
|
288 START (QName('ul'), Attrs([(QName('compact'), u'compact')])) |
24733a5854d9
Avoid unicode literals in `repr`s of `QName` and `Namespace` when not necessary.
cmlenz
parents:
856
diff
changeset
|
289 START (QName('li'), Attrs()) |
26
039fc5b87405
* Split out the XPath tests into a separate `unittest`-based file.
cmlenz
parents:
21
diff
changeset
|
290 TEXT Foo |
039fc5b87405
* Split out the XPath tests into a separate `unittest`-based file.
cmlenz
parents:
21
diff
changeset
|
291 END li |
039fc5b87405
* Split out the XPath tests into a separate `unittest`-based file.
cmlenz
parents:
21
diff
changeset
|
292 END ul |
1 | 293 """ |
294 | |
295 _EMPTY_ELEMS = frozenset(['area', 'base', 'basefont', 'br', 'col', 'frame', | |
296 'hr', 'img', 'input', 'isindex', 'link', 'meta', | |
297 'param']) | |
298 | |
932 | 299 def __init__(self, source, filename=None, encoding=None): |
311
01e2c48f6dfb
* The `HTMLParser` class and the `HTML` function now accept an `encoding` parameter to properly deal with bytestring input (defaults to UTF-8).
cmlenz
parents:
293
diff
changeset
|
300 """Initialize the parser for the given HTML input. |
01e2c48f6dfb
* The `HTMLParser` class and the `HTML` function now accept an `encoding` parameter to properly deal with bytestring input (defaults to UTF-8).
cmlenz
parents:
293
diff
changeset
|
301 |
425
5b248708bbed
Try to use proper reStructuredText for docstrings throughout.
cmlenz
parents:
423
diff
changeset
|
302 :param source: the HTML text as a file-like object |
5b248708bbed
Try to use proper reStructuredText for docstrings throughout.
cmlenz
parents:
423
diff
changeset
|
303 :param filename: the name of the file, if known |
5b248708bbed
Try to use proper reStructuredText for docstrings throughout.
cmlenz
parents:
423
diff
changeset
|
304 :param filename: encoding of the file; ignored if the input is unicode |
311
01e2c48f6dfb
* The `HTMLParser` class and the `HTML` function now accept an `encoding` parameter to properly deal with bytestring input (defaults to UTF-8).
cmlenz
parents:
293
diff
changeset
|
305 """ |
1 | 306 html.HTMLParser.__init__(self) |
307 self.source = source | |
21
eca77129518a
* Include paths are now interpreted relative to the path of the including template. Closes #3.
cmlenz
parents:
1
diff
changeset
|
308 self.filename = filename |
311
01e2c48f6dfb
* The `HTMLParser` class and the `HTML` function now accept an `encoding` parameter to properly deal with bytestring input (defaults to UTF-8).
cmlenz
parents:
293
diff
changeset
|
309 self.encoding = encoding |
21
eca77129518a
* Include paths are now interpreted relative to the path of the including template. Closes #3.
cmlenz
parents:
1
diff
changeset
|
310 self._queue = [] |
1 | 311 self._open_tags = [] |
312 | |
144
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
313 def parse(self): |
433 | 314 """Generator that parses the HTML source, yielding markup events. |
315 | |
316 :return: a markup event stream | |
317 :raises ParseError: if the HTML text is not well formed | |
318 """ | |
144
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
319 def _generate(): |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
320 try: |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
321 bufsize = 4 * 1024 # 4K |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
322 done = False |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
323 while 1: |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
324 while not done and len(self._queue) == 0: |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
325 data = self.source.read(bufsize) |
932 | 326 if not data: # end of data |
144
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
327 self.close() |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
328 done = True |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
329 else: |
932 | 330 if not isinstance(data, unicode): |
331 # bytes | |
332 if self.encoding: | |
333 data = data.decode(self.encoding) | |
334 else: | |
335 raise UnicodeError("source returned bytes, but no encoding specified") | |
144
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
336 self.feed(data) |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
337 for kind, data, pos in self._queue: |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
338 yield kind, data, pos |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
339 self._queue = [] |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
340 if done: |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
341 open_tags = self._open_tags |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
342 open_tags.reverse() |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
343 for tag in open_tags: |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
344 yield END, QName(tag), pos |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
345 break |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
346 except html.HTMLParseError, e: |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
347 msg = '%s: line %d, column %d' % (e.msg, e.lineno, e.offset) |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
348 raise ParseError(msg, self.filename, e.lineno, e.offset) |
146 | 349 return Stream(_generate()).filter(_coalesce) |
144
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
350 |
1 | 351 def __iter__(self): |
144
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
352 return iter(self.parse()) |
21
eca77129518a
* Include paths are now interpreted relative to the path of the including template. Closes #3.
cmlenz
parents:
1
diff
changeset
|
353 |
26
039fc5b87405
* Split out the XPath tests into a separate `unittest`-based file.
cmlenz
parents:
21
diff
changeset
|
354 def _enqueue(self, kind, data, pos=None): |
039fc5b87405
* Split out the XPath tests into a separate `unittest`-based file.
cmlenz
parents:
21
diff
changeset
|
355 if pos is None: |
039fc5b87405
* Split out the XPath tests into a separate `unittest`-based file.
cmlenz
parents:
21
diff
changeset
|
356 pos = self._getpos() |
039fc5b87405
* Split out the XPath tests into a separate `unittest`-based file.
cmlenz
parents:
21
diff
changeset
|
357 self._queue.append((kind, data, pos)) |
039fc5b87405
* Split out the XPath tests into a separate `unittest`-based file.
cmlenz
parents:
21
diff
changeset
|
358 |
21
eca77129518a
* Include paths are now interpreted relative to the path of the including template. Closes #3.
cmlenz
parents:
1
diff
changeset
|
359 def _getpos(self): |
eca77129518a
* Include paths are now interpreted relative to the path of the including template. Closes #3.
cmlenz
parents:
1
diff
changeset
|
360 lineno, column = self.getpos() |
eca77129518a
* Include paths are now interpreted relative to the path of the including template. Closes #3.
cmlenz
parents:
1
diff
changeset
|
361 return (self.filename, lineno, column) |
1 | 362 |
363 def handle_starttag(self, tag, attrib): | |
26
039fc5b87405
* Split out the XPath tests into a separate `unittest`-based file.
cmlenz
parents:
21
diff
changeset
|
364 fixed_attrib = [] |
039fc5b87405
* Split out the XPath tests into a separate `unittest`-based file.
cmlenz
parents:
21
diff
changeset
|
365 for name, value in attrib: # Fixup minimized attributes |
039fc5b87405
* Split out the XPath tests into a separate `unittest`-based file.
cmlenz
parents:
21
diff
changeset
|
366 if value is None: |
312
7e743338a799
Follow-up to [385]: also decode attribute values in the `HTMLParser`.
cmlenz
parents:
311
diff
changeset
|
367 value = unicode(name) |
7e743338a799
Follow-up to [385]: also decode attribute values in the `HTMLParser`.
cmlenz
parents:
311
diff
changeset
|
368 elif not isinstance(value, unicode): |
7e743338a799
Follow-up to [385]: also decode attribute values in the `HTMLParser`.
cmlenz
parents:
311
diff
changeset
|
369 value = value.decode(self.encoding, 'replace') |
403
32b283e1d310
Remove some magic/overhead from `Attrs` creation and manipulation by not automatically wrapping attribute names in `QName`.
cmlenz
parents:
378
diff
changeset
|
370 fixed_attrib.append((QName(name), stripentities(value))) |
26
039fc5b87405
* Split out the XPath tests into a separate `unittest`-based file.
cmlenz
parents:
21
diff
changeset
|
371 |
182
41db0260ebb1
Renamed `Attributes` to `Attrs` to reduce the verbosity.
cmlenz
parents:
160
diff
changeset
|
372 self._enqueue(START, (QName(tag), Attrs(fixed_attrib))) |
1 | 373 if tag in self._EMPTY_ELEMS: |
144
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
374 self._enqueue(END, QName(tag)) |
1 | 375 else: |
376 self._open_tags.append(tag) | |
377 | |
378 def handle_endtag(self, tag): | |
379 if tag not in self._EMPTY_ELEMS: | |
380 while self._open_tags: | |
381 open_tag = self._open_tags.pop() | |
378
fff4a81ffc56
Improve handling of incorrectly nested tags in the HTML parser.
cmlenz
parents:
376
diff
changeset
|
382 self._enqueue(END, QName(open_tag)) |
1 | 383 if open_tag.lower() == tag.lower(): |
384 break | |
385 | |
386 def handle_data(self, text): | |
311
01e2c48f6dfb
* The `HTMLParser` class and the `HTML` function now accept an `encoding` parameter to properly deal with bytestring input (defaults to UTF-8).
cmlenz
parents:
293
diff
changeset
|
387 if not isinstance(text, unicode): |
01e2c48f6dfb
* The `HTMLParser` class and the `HTML` function now accept an `encoding` parameter to properly deal with bytestring input (defaults to UTF-8).
cmlenz
parents:
293
diff
changeset
|
388 text = text.decode(self.encoding, 'replace') |
144
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
389 self._enqueue(TEXT, text) |
1 | 390 |
391 def handle_charref(self, name): | |
423
7589a0e51001
Applied patch for #106 (handling of hex charrefs in HTML parser).
cmlenz
parents:
422
diff
changeset
|
392 if name.lower().startswith('x'): |
7589a0e51001
Applied patch for #106 (handling of hex charrefs in HTML parser).
cmlenz
parents:
422
diff
changeset
|
393 text = unichr(int(name[1:], 16)) |
7589a0e51001
Applied patch for #106 (handling of hex charrefs in HTML parser).
cmlenz
parents:
422
diff
changeset
|
394 else: |
7589a0e51001
Applied patch for #106 (handling of hex charrefs in HTML parser).
cmlenz
parents:
422
diff
changeset
|
395 text = unichr(int(name)) |
144
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
396 self._enqueue(TEXT, text) |
1 | 397 |
398 def handle_entityref(self, name): | |
144
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
399 try: |
856 | 400 text = unichr(entities.name2codepoint[name]) |
144
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
401 except KeyError: |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
402 text = '&%s;' % name |
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
403 self._enqueue(TEXT, text) |
1 | 404 |
405 def handle_pi(self, data): | |
376
74b6bf92f0cd
Fix parsing of processing instructions in HTML input.
cmlenz
parents:
326
diff
changeset
|
406 target, data = data.split(None, 1) |
74b6bf92f0cd
Fix parsing of processing instructions in HTML input.
cmlenz
parents:
326
diff
changeset
|
407 if data.endswith('?'): |
74b6bf92f0cd
Fix parsing of processing instructions in HTML input.
cmlenz
parents:
326
diff
changeset
|
408 data = data[:-1] |
144
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
409 self._enqueue(PI, (target.strip(), data.strip())) |
1 | 410 |
411 def handle_comment(self, text): | |
144
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
412 self._enqueue(COMMENT, text) |
1 | 413 |
414 | |
932 | 415 def HTML(text, encoding=None): |
433 | 416 """Parse the given HTML source and return a markup stream. |
417 | |
418 Unlike with `HTMLParser`, the returned stream is reusable, meaning it can be | |
419 iterated over multiple times: | |
420 | |
932 | 421 >>> html = HTML('<body><h1>Foo</h1></body>', encoding='utf-8') |
853
4376010bb97e
Convert a bunch of print statements to py3k compatible syntax.
cmlenz
parents:
852
diff
changeset
|
422 >>> print(html) |
433 | 423 <body><h1>Foo</h1></body> |
853
4376010bb97e
Convert a bunch of print statements to py3k compatible syntax.
cmlenz
parents:
852
diff
changeset
|
424 >>> print(html.select('h1')) |
433 | 425 <h1>Foo</h1> |
853
4376010bb97e
Convert a bunch of print statements to py3k compatible syntax.
cmlenz
parents:
852
diff
changeset
|
426 >>> print(html.select('h1/text()')) |
433 | 427 Foo |
428 | |
429 :param text: the HTML source | |
430 :return: the parsed XML event stream | |
431 :raises ParseError: if the HTML text is not well-formed, and error recovery | |
432 fails | |
433 """ | |
932 | 434 if isinstance(text, unicode): |
435 return Stream(list(HTMLParser(StringIO(text), encoding=encoding))) | |
436 return Stream(list(HTMLParser(BytesIO(text), encoding=encoding))) | |
859 | 437 |
144
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
438 |
146 | 439 def _coalesce(stream): |
144
28b56f09a7e1
* Coalesce adjacent text events that the parsers would produce when text crossed the buffer boundaries. Fixes #26.
cmlenz
parents:
143
diff
changeset
|
440 """Coalesces adjacent TEXT events into a single event.""" |
146 | 441 textbuf = [] |
442 textpos = None | |
443 for kind, data, pos in chain(stream, [(None, None, None)]): | |
444 if kind is TEXT: | |
445 textbuf.append(data) | |
446 if textpos is None: | |
447 textpos = pos | |
448 else: | |
449 if textbuf: | |
852
04945cd67dad
Remove usage of unicode literals in a couple of places where they were not strictly necessary.
cmlenz
parents:
750
diff
changeset
|
450 yield TEXT, ''.join(textbuf), textpos |
146 | 451 del textbuf[:] |
452 textpos = None | |
453 if kind: | |
454 yield kind, data, pos |