Mercurial > libervia-backend
annotate src/plugins/plugin_xep_0071.py @ 2078:3a0a7e5cef49
quick frontend (chat): added Message.main_message_xhtml property to get XHTML content
author | Goffi <goffi@goffi.org> |
---|---|
date | Sun, 25 Sep 2016 16:07:44 +0200 |
parents | a2bc5089c2eb |
children | 3626b2813158 |
rev | line source |
---|---|
1934
2daf7b4c6756
use of /usr/bin/env instead of /usr/bin/python in shebang
Goffi <goffi@goffi.org>
parents:
1813
diff
changeset
|
1 #!/usr/bin/env python2 |
668 | 2 # -*- coding: utf-8 -*- |
3 | |
4 # SAT plugin for Publish-Subscribe (xep-0071) | |
1766 | 5 # Copyright (C) 2009-2016 Jérôme Poisson (goffi@goffi.org) |
668 | 6 |
7 # This program is free software: you can redistribute it and/or modify | |
8 # it under the terms of the GNU Affero General Public License as published by | |
9 # the Free Software Foundation, either version 3 of the License, or | |
10 # (at your option) any later version. | |
11 | |
12 # This program is distributed in the hope that it will be useful, | |
13 # but WITHOUT ANY WARRANTY; without even the implied warranty of | |
14 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
15 # GNU Affero General Public License for more details. | |
16 | |
17 # You should have received a copy of the GNU Affero General Public License | |
18 # along with this program. If not, see <http://www.gnu.org/licenses/>. | |
19 | |
771 | 20 from sat.core.i18n import _ |
1963
a2bc5089c2eb
backend, frontends: message refactoring (huge commit):
Goffi <goffi@goffi.org>
parents:
1955
diff
changeset
|
21 from sat.core.constants import Const as C |
832
c4b22aedb7d7
plugin groupblog, XEP-0071, XEP-0277, text_syntaxes: manage raw/rich/xhtml data for content/title:
souliane <souliane@mailoo.org>
parents:
811
diff
changeset
|
22 from sat.core import exceptions |
993
301b342c697a
core: use of the new core.log module:
Goffi <goffi@goffi.org>
parents:
922
diff
changeset
|
23 from sat.core.log import getLogger |
301b342c697a
core: use of the new core.log module:
Goffi <goffi@goffi.org>
parents:
922
diff
changeset
|
24 log = getLogger(__name__) |
668 | 25 |
1955
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
26 from twisted.internet import defer |
993
301b342c697a
core: use of the new core.log module:
Goffi <goffi@goffi.org>
parents:
922
diff
changeset
|
27 from wokkel import disco, iwokkel |
668 | 28 from zope.interface import implements |
29 # from lxml import etree | |
1542
94901070478e
plugins: added new MissingModule exceptions to plugins using third party modules
Goffi <goffi@goffi.org>
parents:
1396
diff
changeset
|
30 try: |
94901070478e
plugins: added new MissingModule exceptions to plugins using third party modules
Goffi <goffi@goffi.org>
parents:
1396
diff
changeset
|
31 from lxml import html |
94901070478e
plugins: added new MissingModule exceptions to plugins using third party modules
Goffi <goffi@goffi.org>
parents:
1396
diff
changeset
|
32 except ImportError: |
94901070478e
plugins: added new MissingModule exceptions to plugins using third party modules
Goffi <goffi@goffi.org>
parents:
1396
diff
changeset
|
33 raise exceptions.MissingModule(u"Missing module lxml, please download/install it from http://lxml.de/") |
668 | 34 try: |
35 from twisted.words.protocols.xmlstream import XMPPHandler | |
36 except ImportError: | |
37 from wokkel.subprotocols import XMPPHandler | |
38 | |
39 NS_XHTML_IM = 'http://jabber.org/protocol/xhtml-im' | |
40 NS_XHTML = 'http://www.w3.org/1999/xhtml' | |
41 | |
42 PLUGIN_INFO = { | |
43 "name": "XHTML-IM Plugin", | |
44 "import_name": "XEP-0071", | |
45 "type": "XEP", | |
46 "protocols": ["XEP-0071"], | |
47 "dependencies": ["TEXT-SYNTAXES"], | |
48 "main": "XEP_0071", | |
49 "handler": "yes", | |
50 "description": _("""Implementation of XHTML-IM""") | |
51 } | |
52 | |
53 allowed = { | |
54 "a": set(["href", "style", "type"]), | |
55 "blockquote": set(["style"]), | |
56 "body": set(["style"]), | |
57 "br": set([]), | |
58 "cite": set(["style"]), | |
59 "em": set([]), | |
60 "img": set(["alt", "height", "src", "style", "width"]), | |
61 "li": set(["style"]), | |
62 "ol": set(["style"]), | |
63 "p": set(["style"]), | |
64 "span": set(["style"]), | |
65 "strong": set([]), | |
66 "ul": set(["style"]), | |
67 } | |
68 | |
69 styles_allowed = ["background-color", "color", "font-family", "font-size", "font-style", "font-weight", "margin-left", "margin-right", "text-align", "text-decoration"] | |
70 | |
71 blacklist = ['script'] # tag that we have to kill (we don't keep content) | |
72 | |
73 | |
74 class XEP_0071(object): | |
75 SYNTAX_XHTML_IM = "XHTML-IM" | |
76 | |
77 def __init__(self, host): | |
993
301b342c697a
core: use of the new core.log module:
Goffi <goffi@goffi.org>
parents:
922
diff
changeset
|
78 log.info(_("XHTML-IM plugin initialization")) |
668 | 79 self.host = host |
1955
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
80 self._s = self.host.plugins["TEXT-SYNTAXES"] |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
81 self._s.addSyntax(self.SYNTAX_XHTML_IM, lambda xhtml: xhtml, self.XHTML2XHTML_IM, [self._s.OPT_HIDDEN]) |
668 | 82 host.trigger.add("MessageReceived", self.messageReceivedTrigger) |
1955
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
83 host.trigger.add("messageSend", self.messageSendTrigger) |
668 | 84 |
85 def getHandler(self, profile): | |
86 return XEP_0071_handler(self) | |
87 | |
1955
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
88 def _messagePostTreat(self, data, body_elts): |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
89 """Callback which manage the post treatment of the message in case of XHTML-IM found |
668 | 90 @param data: data send by MessageReceived trigger through post_treat deferred |
1955
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
91 @param body_elts: XHTML-IM body elements found |
668 | 92 @return: the data with the extra parameter updated |
93 """ | |
1955
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
94 # TODO: check if text only body is empty, then try to convert XHTML-IM to pure text and show a warning message |
1963
a2bc5089c2eb
backend, frontends: message refactoring (huge commit):
Goffi <goffi@goffi.org>
parents:
1955
diff
changeset
|
95 def converted(xhtml, lang): |
1955
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
96 if lang: |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
97 data['extra']['xhtml_{}'.format(lang)] = xhtml |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
98 else: |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
99 data['extra']['xhtml'] = xhtml |
668 | 100 |
1955
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
101 defers = [] |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
102 for body_elt in body_elts: |
1963
a2bc5089c2eb
backend, frontends: message refactoring (huge commit):
Goffi <goffi@goffi.org>
parents:
1955
diff
changeset
|
103 lang = body_elt.getAttribute((C.NS_XML, 'lang'), '') |
1955
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
104 d = self._s.convert(body_elt.toXml(), self.SYNTAX_XHTML_IM, safe=True) |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
105 d.addCallback(converted, lang) |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
106 defers.append(d) |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
107 |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
108 d_list = defer.DeferredList(defers) |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
109 d_list.addCallback(lambda dummy: data) |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
110 return d_list |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
111 |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
112 def _messageSendAddRich(self, data, client): |
702
a25db3fe3959
plugin XEP-0071: rich messages management for sendMessage
Goffi <goffi@goffi.org>
parents:
701
diff
changeset
|
113 """ Construct XHTML-IM node and add it XML element |
1955
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
114 |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
115 @param data: message data as sended by messageSend callback |
702
a25db3fe3959
plugin XEP-0071: rich messages management for sendMessage
Goffi <goffi@goffi.org>
parents:
701
diff
changeset
|
116 """ |
1955
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
117 # at this point, either ['extra']['rich'] or ['extra']['xhtml'] exists |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
118 # but both can't exist at the same time |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
119 message_elt = data['xml'] |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
120 html_elt = message_elt.addElement((NS_XHTML_IM, 'html')) |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
121 |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
122 def syntax_converted(xhtml_im, lang): |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
123 body_elt = html_elt.addElement((NS_XHTML, 'body')) |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
124 if lang: |
1963
a2bc5089c2eb
backend, frontends: message refactoring (huge commit):
Goffi <goffi@goffi.org>
parents:
1955
diff
changeset
|
125 body_elt[(C.NS_XML, 'lang')] = lang |
1955
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
126 data['extra']['xhtml_{}'.format(lang)] = xhtml_im |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
127 else: |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
128 data['extra']['xhtml'] = xhtml_im |
702
a25db3fe3959
plugin XEP-0071: rich messages management for sendMessage
Goffi <goffi@goffi.org>
parents:
701
diff
changeset
|
129 body_elt.addRawXml(xhtml_im) |
a25db3fe3959
plugin XEP-0071: rich messages management for sendMessage
Goffi <goffi@goffi.org>
parents:
701
diff
changeset
|
130 |
1955
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
131 syntax = self._s.getCurrentSyntax(client.profile) |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
132 defers = [] |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
133 try: |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
134 rich = data['extra']['rich'] |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
135 except KeyError: |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
136 # we have directly XHTML |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
137 for lang, xhtml in data['extra']['xhtml'].iteritems(): |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
138 d = self._s.convert(xhtml, self._s.SYNTAX_XHTML, self.SYNTAX_XHTML_IM) |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
139 d.addCallback(syntax_converted, lang) |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
140 defers.append(d) |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
141 else: |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
142 # we have rich syntax to convert |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
143 for lang, rich_data in rich.iteritems(): |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
144 d = self._s.convert(rich_data, syntax, self.SYNTAX_XHTML_IM) |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
145 d.addCallback(syntax_converted, lang) |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
146 defers.append(d) |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
147 d_list = defer.DeferredList(defers) |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
148 d_list.addCallback(lambda dummy: data) |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
149 return d_list |
702
a25db3fe3959
plugin XEP-0071: rich messages management for sendMessage
Goffi <goffi@goffi.org>
parents:
701
diff
changeset
|
150 |
1955
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
151 def messageReceivedTrigger(self, client, message, post_treat): |
668 | 152 """ Check presence of XHTML-IM in message |
153 """ | |
154 try: | |
155 html_elt = message.elements(NS_XHTML_IM, 'html').next() | |
156 except StopIteration: | |
157 # No XHTML-IM | |
158 pass | |
1955
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
159 else: |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
160 body_elts = html_elt.elements(NS_XHTML, 'body') |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
161 post_treat.addCallback(self._messagePostTreat, body_elts) |
668 | 162 return True |
163 | |
1955
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
164 def messageSendTrigger(self, client, data, pre_xml_treatments, post_xml_treatments): |
702
a25db3fe3959
plugin XEP-0071: rich messages management for sendMessage
Goffi <goffi@goffi.org>
parents:
701
diff
changeset
|
165 """ Check presence of rich text in extra |
a25db3fe3959
plugin XEP-0071: rich messages management for sendMessage
Goffi <goffi@goffi.org>
parents:
701
diff
changeset
|
166 """ |
1955
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
167 rich = {} |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
168 xhtml = {} |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
169 for key, value in data['extra'].iteritems(): |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
170 if key.startswith('rich'): |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
171 rich[key[5:]] = value |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
172 elif key.startswith('xhtml'): |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
173 xhtml[key[6:]] = value |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
174 if rich and xhtml: |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
175 raise exceptions.DataError(_(u"Can't have XHTML and rich content at the same time")) |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
176 if rich or xhtml: |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
177 if rich: |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
178 data['rich'] = rich |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
179 else: |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
180 data['xhtml'] = xhtml |
633b5c21aefd
backend, frontend: messages refactoring (huge commit, not finished):
Goffi <goffi@goffi.org>
parents:
1934
diff
changeset
|
181 post_xml_treatments.addCallback(self._messageSendAddRich, client) |
702
a25db3fe3959
plugin XEP-0071: rich messages management for sendMessage
Goffi <goffi@goffi.org>
parents:
701
diff
changeset
|
182 return True |
a25db3fe3959
plugin XEP-0071: rich messages management for sendMessage
Goffi <goffi@goffi.org>
parents:
701
diff
changeset
|
183 |
668 | 184 def _purgeStyle(self, styles_raw): |
185 """ Remove unauthorised styles according to the XEP-0071 | |
186 @param styles_raw: raw styles (value of the style attribute) | |
187 """ | |
188 purged = [] | |
189 | |
190 styles = [style.strip().split(':') for style in styles_raw.split(';')] | |
191 | |
192 for style_tuple in styles: | |
193 if len(style_tuple) != 2: | |
194 continue | |
195 name, value = style_tuple | |
196 name = name.strip() | |
197 if name not in styles_allowed: | |
198 continue | |
199 purged.append((name, value.strip())) | |
200 | |
201 return u'; '.join([u"%s: %s" % data for data in purged]) | |
202 | |
203 def XHTML2XHTML_IM(self, xhtml): | |
204 """ Convert XHTML document to XHTML_IM subset | |
205 @param xhtml: raw xhtml to convert | |
206 """ | |
207 # TODO: more clever tag replacement (replace forbidden tags with equivalents when possible) | |
208 | |
209 parser = html.HTMLParser(remove_comments=True, encoding='utf-8') | |
210 root = html.fromstring(xhtml, parser=parser) | |
211 body_elt = root.find('body') | |
212 if body_elt is None: | |
213 # we use the whole XML as body if no body element is found | |
214 body_elt = html.Element('body') | |
215 body_elt.append(root) | |
216 else: | |
217 body_elt.attrib.clear() | |
218 | |
219 allowed_tags = allowed.keys() | |
220 to_strip = [] | |
221 for elem in body_elt.iter(): | |
222 if elem.tag not in allowed_tags: | |
223 to_strip.append(elem) | |
224 else: | |
225 # we remove unallowed attributes | |
226 attrib = elem.attrib | |
227 att_to_remove = set(attrib).difference(allowed[elem.tag]) | |
228 for att in att_to_remove: | |
229 del(attrib[att]) | |
230 if "style" in attrib: | |
231 attrib["style"] = self._purgeStyle(attrib["style"]) | |
232 | |
233 for elem in to_strip: | |
234 if elem.tag in blacklist: | |
235 #we need to remove the element and all descendants | |
993
301b342c697a
core: use of the new core.log module:
Goffi <goffi@goffi.org>
parents:
922
diff
changeset
|
236 log.debug(u"removing black listed tag: %s" % (elem.tag)) |
668 | 237 elem.drop_tree() |
238 else: | |
239 elem.drop_tag() | |
701
98b2400e17d6
plugin XEP-0071: XHTML2XHTML_IM don't return the <body> root tag anymore.
Goffi <goffi@goffi.org>
parents:
668
diff
changeset
|
240 if len(body_elt) !=1: |
98b2400e17d6
plugin XEP-0071: XHTML2XHTML_IM don't return the <body> root tag anymore.
Goffi <goffi@goffi.org>
parents:
668
diff
changeset
|
241 root_elt = body_elt |
98b2400e17d6
plugin XEP-0071: XHTML2XHTML_IM don't return the <body> root tag anymore.
Goffi <goffi@goffi.org>
parents:
668
diff
changeset
|
242 body_elt.tag = "p" |
98b2400e17d6
plugin XEP-0071: XHTML2XHTML_IM don't return the <body> root tag anymore.
Goffi <goffi@goffi.org>
parents:
668
diff
changeset
|
243 else: |
98b2400e17d6
plugin XEP-0071: XHTML2XHTML_IM don't return the <body> root tag anymore.
Goffi <goffi@goffi.org>
parents:
668
diff
changeset
|
244 root_elt = body_elt[0] |
668 | 245 |
701
98b2400e17d6
plugin XEP-0071: XHTML2XHTML_IM don't return the <body> root tag anymore.
Goffi <goffi@goffi.org>
parents:
668
diff
changeset
|
246 return html.tostring(root_elt, encoding='unicode', method='xml') |
668 | 247 |
248 class XEP_0071_handler(XMPPHandler): | |
249 implements(iwokkel.IDisco) | |
250 | |
251 def __init__(self, plugin_parent): | |
252 self.plugin_parent = plugin_parent | |
253 self.host = plugin_parent.host | |
254 | |
255 def getDiscoInfo(self, requestor, target, nodeIdentifier=''): | |
256 return [disco.DiscoFeature(NS_XHTML_IM)] | |
257 | |
258 def getDiscoItems(self, requestor, target, nodeIdentifier=''): | |
259 return [] |