Mercurial > libervia-backend

annotate frontends/src/tools/strings.py @ 1809:821c77574ad9

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
test, setup: requires lxml >= 3.1.0 (fixes html cleaning issue)
author souliane <souliane@mailoo.org>
date Thu, 21 Aug 2014 20:45:24 +0200
parents 07390a9d1c09
children b43ee22eac98
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
690
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
1 #!/usr/bin/python
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
2 # -*- coding: utf-8 -*-
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
3
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
4 # SAT helpers methods for plugins
1766
d17772b0fe22 copyright update
Goffi <goffi@goffi.org>
parents: 1396
diff changeset
5 # Copyright (C) 2013-2016 Adrien Cossa (souliane@mailoo.org)
690
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
6
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
7 # This program is free software: you can redistribute it and/or modify
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
8 # it under the terms of the GNU Affero General Public License as published by
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
9 # the Free Software Foundation, either version 3 of the License, or
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
10 # (at your option) any later version.
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
11
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
12 # This program is distributed in the hope that it will be useful,
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
13 # but WITHOUT ANY WARRANTY; without even the implied warranty of
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
14 # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
15 # GNU Affero General Public License for more details.
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
16
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
17 # You should have received a copy of the GNU Affero General Public License
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
18 # along with this program. If not, see <http://www.gnu.org/licenses/>.
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
19
769
8b6137f7b4e1 tools: addURLToText moved from libervia to sat_frontends/tools/strings
souliane <souliane@mailoo.org>
parents: 719
diff changeset
20 import re
1801
07390a9d1c09 tools (strings): add method fixXHTMLLinks to add a scheme if missing, and force opening in new tab
souliane <souliane@mailoo.org>
parents: 1790
diff changeset
21 from feed.atom import Content
769
8b6137f7b4e1 tools: addURLToText moved from libervia to sat_frontends/tools/strings
souliane <souliane@mailoo.org>
parents: 719
diff changeset
22
690
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
23
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
24 def getURLParams(url):
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
25 """This comes from pyjamas.Location.makeUrlDict with a small change
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
26 to also parse full URLs, and parameters with no value specified
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
27 (in that case the default value "" is used).
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
28 @param url: any URL with or without parameters
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
29 @return: a dictionary of the parameters, if any was given, or {}
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
30 """
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
31 dict_ = {}
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
32 if "/" in url:
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
33 # keep the part after the last "/"
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
34 url = url[url.rindex("/") + 1:]
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
35 if url.startswith("?"):
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
36 # remove the first "?"
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
37 url = url[1:]
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
38 pairs = url.split("&")
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
39 for pair in pairs:
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
40 if len(pair) < 3:
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
41 continue
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
42 kv = pair.split("=", 1)
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
43 dict_[kv[0]] = kv[1] if len(kv) > 1 else ""
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
44 return dict_
d8e7a58eaa00 misc: added a file for frontend's string operations:
souliane <souliane@mailoo.org>
parents:
diff changeset
45
769
8b6137f7b4e1 tools: addURLToText moved from libervia to sat_frontends/tools/strings
souliane <souliane@mailoo.org>
parents: 719
diff changeset
46
1790
bf7e468fe440 tools (strings): add a parameter "new_target" to addURLToText
souliane <souliane@mailoo.org>
parents: 1766
diff changeset
47 def addURLToText(string, new_target=True):
bf7e468fe440 tools (strings): add a parameter "new_target" to addURLToText
souliane <souliane@mailoo.org>
parents: 1766
diff changeset
48 """Check a text for what looks like an URL and make it clickable.
769
8b6137f7b4e1 tools: addURLToText moved from libervia to sat_frontends/tools/strings
souliane <souliane@mailoo.org>
parents: 719
diff changeset
49
1790
bf7e468fe440 tools (strings): add a parameter "new_target" to addURLToText
souliane <souliane@mailoo.org>
parents: 1766
diff changeset
50 @param string (unicode): text to process
bf7e468fe440 tools (strings): add a parameter "new_target" to addURLToText
souliane <souliane@mailoo.org>
parents: 1766
diff changeset
51 @param new_target (bool): if True, make the link open in a new window
bf7e468fe440 tools (strings): add a parameter "new_target" to addURLToText
souliane <souliane@mailoo.org>
parents: 1766
diff changeset
52 """
bf7e468fe440 tools (strings): add a parameter "new_target" to addURLToText
souliane <souliane@mailoo.org>
parents: 1766
diff changeset
53 # XXX: report any change to libervia.browser.strings.addURLToText
bf7e468fe440 tools (strings): add a parameter "new_target" to addURLToText
souliane <souliane@mailoo.org>
parents: 1766
diff changeset
54 # Regexp from http://daringfireball.net/2010/07/improved_regex_for_matching_urls
769
8b6137f7b4e1 tools: addURLToText moved from libervia to sat_frontends/tools/strings
souliane <souliane@mailoo.org>
parents: 719
diff changeset
55 def repl(match):
8b6137f7b4e1 tools: addURLToText moved from libervia to sat_frontends/tools/strings
souliane <souliane@mailoo.org>
parents: 719
diff changeset
56 url = match.group(0)
8b6137f7b4e1 tools: addURLToText moved from libervia to sat_frontends/tools/strings
souliane <souliane@mailoo.org>
parents: 719
diff changeset
57 if not re.match(r"""[a-z]{3,}://|mailto:|xmpp:""", url):
8b6137f7b4e1 tools: addURLToText moved from libervia to sat_frontends/tools/strings
souliane <souliane@mailoo.org>
parents: 719
diff changeset
58 url = "http://" + url
1790
bf7e468fe440 tools (strings): add a parameter "new_target" to addURLToText
souliane <souliane@mailoo.org>
parents: 1766
diff changeset
59 target = ' target="_blank"' if new_target else ''
bf7e468fe440 tools (strings): add a parameter "new_target" to addURLToText
souliane <souliane@mailoo.org>
parents: 1766
diff changeset
60 return '<a href="%s"%s class="url">%s</a>' % (url, target, match.group(0))
769
8b6137f7b4e1 tools: addURLToText moved from libervia to sat_frontends/tools/strings
souliane <souliane@mailoo.org>
parents: 719
diff changeset
61 pattern = r"""(?i)\b((?:[a-z]{3,}://|(www|ftp)\d{0,3}[.]|[a-z0-9.\-]+[.][a-z]{2,4}/|mailto:|xmpp:)(?:[^\s()<>]+|\(([^\s()<>]+|(\([^\s()<>]+\)))*\))+(?:\(([^\s()<>]+|(\([^\s()<>]+\)))*\)|[^\s`!()\[\]{};:'".,<>?]))"""
8b6137f7b4e1 tools: addURLToText moved from libervia to sat_frontends/tools/strings
souliane <souliane@mailoo.org>
parents: 719
diff changeset
62 return re.sub(pattern, repl, string)
842
429c6a0ef73d frontends (tools): addURLToImage makes an image clickable
souliane <souliane@mailoo.org>
parents: 811
diff changeset
63
429c6a0ef73d frontends (tools): addURLToImage makes an image clickable
souliane <souliane@mailoo.org>
parents: 811
diff changeset
64
429c6a0ef73d frontends (tools): addURLToImage makes an image clickable
souliane <souliane@mailoo.org>
parents: 811
diff changeset
65 def addURLToImage(string):
1790
bf7e468fe440 tools (strings): add a parameter "new_target" to addURLToText
souliane <souliane@mailoo.org>
parents: 1766
diff changeset
66 """Check a XHTML text for what looks like an imageURL and make it clickable.
bf7e468fe440 tools (strings): add a parameter "new_target" to addURLToText
souliane <souliane@mailoo.org>
parents: 1766
diff changeset
67
bf7e468fe440 tools (strings): add a parameter "new_target" to addURLToText
souliane <souliane@mailoo.org>
parents: 1766
diff changeset
68 @param string (unicode): text to process
bf7e468fe440 tools (strings): add a parameter "new_target" to addURLToText
souliane <souliane@mailoo.org>
parents: 1766
diff changeset
69 """
bf7e468fe440 tools (strings): add a parameter "new_target" to addURLToText
souliane <souliane@mailoo.org>
parents: 1766
diff changeset
70 # XXX: report any change to libervia.browser.strings.addURLToImage
842
429c6a0ef73d frontends (tools): addURLToImage makes an image clickable
souliane <souliane@mailoo.org>
parents: 811
diff changeset
71 def repl(match):
429c6a0ef73d frontends (tools): addURLToImage makes an image clickable
souliane <souliane@mailoo.org>
parents: 811
diff changeset
72 url = match.group(1)
429c6a0ef73d frontends (tools): addURLToImage makes an image clickable
souliane <souliane@mailoo.org>
parents: 811
diff changeset
73 return '<a href="%s" target="_blank">%s</a>' % (url, match.group(0))
429c6a0ef73d frontends (tools): addURLToImage makes an image clickable
souliane <souliane@mailoo.org>
parents: 811
diff changeset
74 pattern = r"""<img[^>]* src="([^"]+)"[^>]*>"""
429c6a0ef73d frontends (tools): addURLToImage makes an image clickable
souliane <souliane@mailoo.org>
parents: 811
diff changeset
75 return re.sub(pattern, repl, string)
1801
07390a9d1c09 tools (strings): add method fixXHTMLLinks to add a scheme if missing, and force opening in new tab
souliane <souliane@mailoo.org>
parents: 1790
diff changeset
76
07390a9d1c09 tools (strings): add method fixXHTMLLinks to add a scheme if missing, and force opening in new tab
souliane <souliane@mailoo.org>
parents: 1790
diff changeset
77
07390a9d1c09 tools (strings): add method fixXHTMLLinks to add a scheme if missing, and force opening in new tab
souliane <souliane@mailoo.org>
parents: 1790
diff changeset
78 def fixXHTMLLinks(xhtml):
07390a9d1c09 tools (strings): add method fixXHTMLLinks to add a scheme if missing, and force opening in new tab
souliane <souliane@mailoo.org>
parents: 1790
diff changeset
79 """Add http:// if the scheme is missing and force opening in a new window.
07390a9d1c09 tools (strings): add method fixXHTMLLinks to add a scheme if missing, and force opening in new tab
souliane <souliane@mailoo.org>
parents: 1790
diff changeset
80
07390a9d1c09 tools (strings): add method fixXHTMLLinks to add a scheme if missing, and force opening in new tab
souliane <souliane@mailoo.org>
parents: 1790
diff changeset
81 @param string (unicode): XHTML Content
07390a9d1c09 tools (strings): add method fixXHTMLLinks to add a scheme if missing, and force opening in new tab
souliane <souliane@mailoo.org>
parents: 1790
diff changeset
82 """
07390a9d1c09 tools (strings): add method fixXHTMLLinks to add a scheme if missing, and force opening in new tab
souliane <souliane@mailoo.org>
parents: 1790
diff changeset
83 subs = []
07390a9d1c09 tools (strings): add method fixXHTMLLinks to add a scheme if missing, and force opening in new tab
souliane <souliane@mailoo.org>
parents: 1790
diff changeset
84 for match in re.finditer(r'<a( \w+="[^"]*")* ?/?>', xhtml):
07390a9d1c09 tools (strings): add method fixXHTMLLinks to add a scheme if missing, and force opening in new tab
souliane <souliane@mailoo.org>
parents: 1790
diff changeset
85 tag = match.group(0)
07390a9d1c09 tools (strings): add method fixXHTMLLinks to add a scheme if missing, and force opening in new tab
souliane <souliane@mailoo.org>
parents: 1790
diff changeset
86 url = re.search(r'href="([^"]*)"', tag)
07390a9d1c09 tools (strings): add method fixXHTMLLinks to add a scheme if missing, and force opening in new tab
souliane <souliane@mailoo.org>
parents: 1790
diff changeset
87 if url:
07390a9d1c09 tools (strings): add method fixXHTMLLinks to add a scheme if missing, and force opening in new tab
souliane <souliane@mailoo.org>
parents: 1790
diff changeset
88 if not re.search(r'target="([^"]*)"', tag): # no target
07390a9d1c09 tools (strings): add method fixXHTMLLinks to add a scheme if missing, and force opening in new tab
souliane <souliane@mailoo.org>
parents: 1790
diff changeset
89 subs.append((tag, '<a target="_blank"%s' % tag[2:]))
07390a9d1c09 tools (strings): add method fixXHTMLLinks to add a scheme if missing, and force opening in new tab
souliane <souliane@mailoo.org>
parents: 1790
diff changeset
90 if not re.match(r"^\w+://", url.group(1)): # no scheme
07390a9d1c09 tools (strings): add method fixXHTMLLinks to add a scheme if missing, and force opening in new tab
souliane <souliane@mailoo.org>
parents: 1790
diff changeset
91 subs.append((url.group(0), 'href="http://%s"' % url.group(1)))
07390a9d1c09 tools (strings): add method fixXHTMLLinks to add a scheme if missing, and force opening in new tab
souliane <souliane@mailoo.org>
parents: 1790
diff changeset
92
07390a9d1c09 tools (strings): add method fixXHTMLLinks to add a scheme if missing, and force opening in new tab
souliane <souliane@mailoo.org>
parents: 1790
diff changeset
93 for url, new_url in subs:
07390a9d1c09 tools (strings): add method fixXHTMLLinks to add a scheme if missing, and force opening in new tab
souliane <souliane@mailoo.org>
parents: 1790
diff changeset
94 xhtml = xhtml.replace(url, new_url)
07390a9d1c09 tools (strings): add method fixXHTMLLinks to add a scheme if missing, and force opening in new tab
souliane <souliane@mailoo.org>
parents: 1790
diff changeset
95 return xhtml
07390a9d1c09 tools (strings): add method fixXHTMLLinks to add a scheme if missing, and force opening in new tab
souliane <souliane@mailoo.org>
parents: 1790
diff changeset
96