Mercurial > libervia-backend

annotate src/plugins/plugin_misc_text_syntaxes.py @ 720:3954cb4d96a7

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
jp: fixed sendFile call
author Goffi <goffi@goffi.org>
date Sat, 23 Nov 2013 13:38:05 +0100
parents 6c8a119dcc94
children 312a2842b2b8
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
665
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
1 #!/usr/bin/python
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
2 # -*- coding: utf-8 -*-
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
3
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
4 # SAT plugin for managing various text syntaxes
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
5 # Copyright (C) 2009, 2010, 2011, 2012, 2013 Jérôme Poisson (goffi@goffi.org)
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
6
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
7 # This program is free software: you can redistribute it and/or modify
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
8 # it under the terms of the GNU Affero General Public License as published by
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
9 # the Free Software Foundation, either version 3 of the License, or
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
10 # (at your option) any later version.
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
11
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
12 # This program is distributed in the hope that it will be useful,
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
13 # but WITHOUT ANY WARRANTY; without even the implied warranty of
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
14 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
15 # GNU Affero General Public License for more details.
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
16
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
17 # You should have received a copy of the GNU Affero General Public License
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
18 # along with this program. If not, see <http://www.gnu.org/licenses/>.
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
19
695
ddd2781bdd8c plugin text syntaxes: fixed missing import
Goffi <goffi@goffi.org>
parents: 692
diff changeset
20 from logging import debug, info, error, warning
665
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
21
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
22 from wokkel import disco, pubsub
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
23 from twisted.internet import defer
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
24 from twisted.internet.threads import deferToThread
705
6c8a119dcc94 plugin text syntaxes: clean_xhtml now accept lxml's HtmlElement to avoid parsing two times the same xml
Goffi <goffi@goffi.org>
parents: 702
diff changeset
25 from sat.core import exceptions
674
fb0b1100c908 plugin text_syntaxes: fixed clean_xhml (it now return XHTML instead of HTML)
Goffi <goffi@goffi.org>
parents: 665
diff changeset
26 from lxml import html
fb0b1100c908 plugin text_syntaxes: fixed clean_xhml (it now return XHTML instead of HTML)
Goffi <goffi@goffi.org>
parents: 665
diff changeset
27 from lxml.html import clean
692
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
28 import re
674
fb0b1100c908 plugin text_syntaxes: fixed clean_xhml (it now return XHTML instead of HTML)
Goffi <goffi@goffi.org>
parents: 665
diff changeset
29
665
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
30
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
31 CATEGORY = "Composition"
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
32 NAME = "Syntax"
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
33 _SYNTAX_XHTML = "XHTML"
692
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
34 # TODO: check/adapt following list
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
35 STYLES_WHITELIST = ["azimuth", "background-color", "border-bottom-color", "border-collapse", "border-color", "border-left-color", "border-right-color", "border-top-color", "clear", "color", "cursor", "direction", "display", "elevation", "float", "font", "font-family", "font-size", "font-style", "font-variant", "font-weight", "height", "letter-spacing", "line-height", "overflow", "pause", "pause-after", "pause-before", "pitch", "pitch-range", "richness", "speak", "speak-header", "speak-numeral", "speak-punctuation", "speech-rate", "stress", "text-align", "text-decoration", "text-indent", "unicode-bidi", "vertical-align", "voice-family", "volume", "white-space", "width"] # based on feedparser list (http://pythonhosted.org/feedparser/html-sanitization.html)
665
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
36
692
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
37 SAFE_ATTRS = html.defs.safe_attrs.union(('style',))
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
38 STYLES_VALUES_REGEX = r'^(' + '|'.join(['([a-z-]+)', # alphabetical names
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
39 '(#[0-9a-f]+)', # hex value
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
40 '(\d+(.\d+)? *(|%|em|ex|px|in|cm|mm|pt|pc))', # values with units (or not)
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
41 'rgb\( *((\d+(.\d+)?), *){2}(\d+(.\d+)?) *\)', # rgb function
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
42 'rgba\( *((\d+(.\d+)?), *){3}(\d+(.\d+)?) *\)', # rgba function
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
43 ]) + ') *(!important)?$' # we accept "!important" at the end
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
44 STYLES_ACCEPTED_VALUE = re.compile(STYLES_VALUES_REGEX)
665
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
45
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
46 PLUGIN_INFO = {
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
47 "name": "Text syntaxes",
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
48 "import_name": "TEXT-SYNTAXES",
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
49 "type": "MISC",
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
50 "protocols": [],
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
51 "dependencies": [],
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
52 "main": "TextSyntaxes",
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
53 "handler": "no",
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
54 "description": _("""Management of various text syntaxes (XHTML-IM, Markdown, etc)""")
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
55 }
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
56
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
57 class UnknownSyntax(Exception):
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
58 pass
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
59
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
60 class TextSyntaxes(object):
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
61 """ Text conversion class
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
62 XHTML utf-8 is used as intermediate language for conversions
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
63 """
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
64
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
65 OPT_DEFAULT = "DEFAULT"
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
66 OPT_HIDDEN = "HIDDEN"
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
67 OPT_NO_THREAD = "NO_THREAD"
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
68 SYNTAX_XHTML = _SYNTAX_XHTML
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
69 SYNTAX_MARKDOWN = "markdown"
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
70
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
71 params = """
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
72 <params>
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
73 <individual>
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
74 <category name="%(category_name)s" label="%(category_label)s">
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
75 <param name="%(name)s" label="%(label)s"
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
76 value="%(default)s" type="list" security="0">
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
77 %(options)s
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
78 </param>
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
79 </category>
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
80 </individual>
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
81 </params>
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
82 """
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
83
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
84 params_data = {
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
85 'category_name': CATEGORY,
692
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
86 'category_label': _(CATEGORY), #FIXME: check that gettext can handle this
665
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
87 'name': NAME,
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
88 'label': _(NAME),
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
89 'default': _SYNTAX_XHTML,
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
90 'syntaxes': {},
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
91 }
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
92
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
93 def __init__(self, host):
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
94 info(_("Text syntaxes plugin initialization"))
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
95 self.host = host
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
96 self.syntaxes = {}
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
97 self.addSyntax(self.SYNTAX_XHTML, lambda xhtml: defer.succeed(xhtml), lambda xhtml: defer.succeed(xhtml),
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
98 TextSyntaxes.OPT_NO_THREAD)
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
99 try:
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
100 import markdown, html2text
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
101 self.addSyntax(self.SYNTAX_MARKDOWN, markdown.markdown, html2text.html2text, [TextSyntaxes.OPT_DEFAULT])
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
102 except ImportError:
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
103 warning("markdown or html2text not found, can't use Markdown syntax")
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
104 host.bridge.addMethod("syntaxConvert", ".plugin", in_sign='sssb', out_sign='s',
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
105 async=True, method=self.convert)
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
106
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
107 def _updateParamOptions(self):
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
108 data_synt = TextSyntaxes.params_data['syntaxes']
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
109 syntaxes = []
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
110
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
111 for syntax in data_synt.keys():
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
112 flags = data_synt[syntax]["flags"]
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
113 if TextSyntaxes.OPT_HIDDEN not in flags:
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
114 syntaxes.append(syntax)
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
115
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
116 syntaxes.sort(key=unicode.lower)
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
117 options = []
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
118
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
119 for syntax in syntaxes:
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
120 options.append(u'<option value="%s" />' % syntax)
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
121
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
122 TextSyntaxes.params_data["options"] = u'\n'.join(options)
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
123 self.host.memory.updateParams(TextSyntaxes.params % TextSyntaxes.params_data)
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
124
702
a25db3fe3959 plugin XEP-0071: rich messages management for sendMessage
Goffi <goffi@goffi.org>
parents: 699
diff changeset
125 def getCurrentSyntax(self, profile):
665
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
126 """ Return the selected syntax for the given profile
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
127
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
128 @param profile: %(doc_profile)s
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
129 @return: profile selected syntax
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
130 """
699
e819630c6737 plugin text syntaxes: fixed bad parameter order for getParamA
Goffi <goffi@goffi.org>
parents: 695
diff changeset
131 return self.host.memory.getParamA(NAME, CATEGORY , profile_key=profile)
665
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
132
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
133 def clean_xhtml(self, xhtml):
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
134 """ Clean XHTML text by removing potentially dangerous/malicious parts
705
6c8a119dcc94 plugin text syntaxes: clean_xhtml now accept lxml's HtmlElement to avoid parsing two times the same xml
Goffi <goffi@goffi.org>
parents: 702
diff changeset
135 @param xhtml: raw xhtml text to clean (or lxml's HtmlElement)
665
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
136 """
674
fb0b1100c908 plugin text_syntaxes: fixed clean_xhml (it now return XHTML instead of HTML)
Goffi <goffi@goffi.org>
parents: 665
diff changeset
137 def blocking_cleaning(xhtml):
692
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
138 """ Clean XHTML and style attributes """
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
139
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
140 def clean_style(styles_raw):
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
141 """" Remove styles not in the whitelist,
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
142 or where the value doesn't match the regex """
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
143 styles = styles_raw.split(";")
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
144 cleaned_styles = []
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
145 for style in styles:
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
146 try:
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
147 key, value = style.split(':')
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
148 except ValueError:
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
149 continue
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
150 key = key.lower().strip()
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
151 if key not in STYLES_WHITELIST:
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
152 continue
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
153 value = value.lower().strip()
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
154 if not STYLES_ACCEPTED_VALUE.match(value):
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
155 continue
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
156 if value == "none":
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
157 continue
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
158 cleaned_styles.append((key, value))
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
159 return "; ".join(["%s: %s" % (key, value) for key, value in cleaned_styles])
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
160
705
6c8a119dcc94 plugin text syntaxes: clean_xhtml now accept lxml's HtmlElement to avoid parsing two times the same xml
Goffi <goffi@goffi.org>
parents: 702
diff changeset
161 if isinstance(xhtml, basestring):
6c8a119dcc94 plugin text syntaxes: clean_xhtml now accept lxml's HtmlElement to avoid parsing two times the same xml
Goffi <goffi@goffi.org>
parents: 702
diff changeset
162 xhtml_elt = html.fromstring(xhtml)
6c8a119dcc94 plugin text syntaxes: clean_xhtml now accept lxml's HtmlElement to avoid parsing two times the same xml
Goffi <goffi@goffi.org>
parents: 702
diff changeset
163 elif isinstance(xhtml, html.HtmlElement):
6c8a119dcc94 plugin text syntaxes: clean_xhtml now accept lxml's HtmlElement to avoid parsing two times the same xml
Goffi <goffi@goffi.org>
parents: 702
diff changeset
164 xhtml_elt = xhtml
6c8a119dcc94 plugin text syntaxes: clean_xhtml now accept lxml's HtmlElement to avoid parsing two times the same xml
Goffi <goffi@goffi.org>
parents: 702
diff changeset
165 else:
6c8a119dcc94 plugin text syntaxes: clean_xhtml now accept lxml's HtmlElement to avoid parsing two times the same xml
Goffi <goffi@goffi.org>
parents: 702
diff changeset
166 error("Only strings and HtmlElements can be cleaned")
6c8a119dcc94 plugin text syntaxes: clean_xhtml now accept lxml's HtmlElement to avoid parsing two times the same xml
Goffi <goffi@goffi.org>
parents: 702
diff changeset
167 raise exceptions.DataError
674
fb0b1100c908 plugin text_syntaxes: fixed clean_xhml (it now return XHTML instead of HTML)
Goffi <goffi@goffi.org>
parents: 665
diff changeset
168 cleaner = clean.Cleaner(style=False,
fb0b1100c908 plugin text_syntaxes: fixed clean_xhml (it now return XHTML instead of HTML)
Goffi <goffi@goffi.org>
parents: 665
diff changeset
169 add_nofollow=False,
692
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
170 safe_attrs=SAFE_ATTRS)
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
171 xhtml_elt = cleaner.clean_html(xhtml_elt)
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
172 for elt in xhtml_elt.xpath("//*[@style]"):
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
173 elt.set("style", clean_style(elt.get('style')))
e98db42cd78c plugin text syntaxes: styles sanitisation
Goffi <goffi@goffi.org>
parents: 674
diff changeset
174 return html.tostring(xhtml_elt, method='xml')
674
fb0b1100c908 plugin text_syntaxes: fixed clean_xhml (it now return XHTML instead of HTML)
Goffi <goffi@goffi.org>
parents: 665
diff changeset
175
705
6c8a119dcc94 plugin text syntaxes: clean_xhtml now accept lxml's HtmlElement to avoid parsing two times the same xml
Goffi <goffi@goffi.org>
parents: 702
diff changeset
176 return deferToThread(blocking_cleaning, xhtml)
665
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
177
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
178 def convert(self, text, syntax_from, syntax_to=_SYNTAX_XHTML, safe=True):
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
179 """ Convert a text between two syntaxes
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
180 @param text: text to convert
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
181 @param syntax_from: source syntax (e.g. "markdown")
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
182 @param syntax_to: dest syntax (e.g.: "XHTML")
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
183 @param safe: clean resulting XHTML to avoid malicious code if True
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
184 @return: converted text """
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
185
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
186 syntaxes = TextSyntaxes.params_data['syntaxes']
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
187 if syntax_from not in syntaxes:
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
188 raise UnknownSyntax(syntax_from)
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
189 if syntax_to not in syntaxes:
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
190 raise UnknownSyntax(syntax_to)
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
191 d = None
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
192
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
193 if TextSyntaxes.OPT_NO_THREAD in syntaxes[syntax_from]["flags"]:
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
194 d = syntaxes[syntax_from]["to"](text)
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
195 else:
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
196 d = deferToThread(syntaxes[syntax_from]["to"], text)
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
197
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
198 #TODO: keep only body element and change it to a div here ?
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
199
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
200 if safe:
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
201 d.addCallback(self.clean_xhtml)
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
202
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
203 if TextSyntaxes.OPT_NO_THREAD in syntaxes[syntax_to]["flags"]:
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
204 d.addCallback(syntaxes[syntax_to]["from"])
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
205 else:
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
206 d.addCallback(lambda xhtml: deferToThread(syntaxes[syntax_to]["from"], xhtml))
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
207
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
208 return d
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
209
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
210 def addSyntax(self, name, to_xhtml_cb, from_xhtml_cb, flags = None):
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
211 """ Add a new syntax to the manager
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
212 @param name: unique name of the syntax
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
213 @param to_xhtml_cb: callback to convert from syntax to XHTML
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
214 @param from_xhtml_cb: callback to convert from XHTML to syntax
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
215 @param flags: set of optional flags, can be:
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
216 TextSyntaxes.OPT_DEFAULT: use as the default syntax (replace former one)
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
217 TextSyntaxes.OPT_HIDDEN: do not show in parameters
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
218 TextSyntaxes.OPT_NO_THREAD: do not defer to thread when converting (the callback must then return a deferred)
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
219
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
220 """
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
221 name = unicode(name)
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
222 flags = flags or []
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
223 if TextSyntaxes.OPT_HIDDEN in flags and TextSyntaxes.OPT_DEFAULT in flags:
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
224 raise ValueError("%s and %s are mutually exclusive" % (TextSyntaxes.OPT_HIDDEN, TextSyntaxes.OPT_DEFAULT))
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
225
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
226 syntaxes = TextSyntaxes.params_data['syntaxes']
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
227 syntaxes[name] = {"to": to_xhtml_cb, "from": from_xhtml_cb, "flags": flags}
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
228 if TextSyntaxes.OPT_DEFAULT in flags:
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
229 syntaxes = TextSyntaxes.params_data['default'] = name
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
230
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
231 self._updateParamOptions()
6a64e0a759e6 plugin text syntaxes: this plugin manage rich text syntaxes conversions and cleaning.
Goffi <goffi@goffi.org>
parents:
diff changeset
232