~ubuntu-branches/ubuntu/utopic/gozerbot/utopic

« back to all changes in this revision

Viewing changes to gozerbot/utils/textutils.py

• browse files at revision 8
• compare with another revision
• download diff
• download tarball
• view history from revision 8

Show diffs side-by-side

added

removed

gozerbot/utils/textutils.py

 

1

import cgi

 

2

import re

 

3

import htmlentitydefs

 

4

 

 

5

def unescape_charref(ref): 

 

6

    name = ref[2:-1] 

 

7

    base = 10 

 

8

    if name.startswith("x"): 

 

9

        name = name[1:] 

 

10

        base = 16 

 

11

    return unichr(int(name, base)) 

 

12

def replace_entities(match): 

 

13

    ent = match.group() 

 

14

    if ent[1] == "#": 

 

15

        return unescape_charref(ent) 

 

16

    repl = htmlentitydefs.name2codepoint.get(ent[1:-1]) 

 

17

    if repl is not None: 

 

18

        repl = unichr(repl) 

 

19

    else: 

 

20

        repl = ent 

 

21

    return repl 

 

22

 

 

23

def html_unescape(data): 

 

24

    '''

 

25

    Unescape (numeric) HTML entities.

 

26

    '''

 

27

    return re.sub(r"&#?[A-Za-z0-9]+?;", replace_entities, data) 

 

28

 

 

29

def html_escape(data):

 

30

    '''

 

31

    Escape HTML entities.

 

32

    '''

 

33

    return cgi.escape(data)

 

34

 

• Older »

Loggerhead is a web-based interface for Breezy
Version: 2.0.1