annotate Tools/scripts/parseentities.py @ 55729:d91aff206dfd 3.1

Merged revisions 74551 via svnmerge from svn+ssh://pythondev@svn.python.org/python/branches/py3k ........ r74551 | benjamin.peterson | 2009-08-26 09:35:45 +0200 (Wed, 26 Aug 2009) | 1 line the old buffer object doesn't exist in Python 3 #6659 ........
author Benjamin Peterson <benjamin@python.org>
date Wed, 26 Aug 2009 07:38:00 +0000
parents 893a4dc17a91
children fcfe324663a6
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
11210
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
1 #!/usr/local/bin/python
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
2 """ Utility for parsing HTML entity definitions available from:
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
3
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
4 http://www.w3.org/ as e.g.
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
5 http://www.w3.org/TR/REC-html40/HTMLlat1.ent
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
6
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
7 Input is read from stdin, output is written to stdout in form of a
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
8 Python snippet defining a dictionary "entitydefs" mapping literal
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
9 entity name to character or numeric entity.
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
10
15930
14a27acc755a Whitespace normalization.
Tim Peters <tim.peters@gmail.com>
parents: 14657
diff changeset
11 Marc-Andre Lemburg, mal@lemburg.com, 1999.
11210
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
12 Use as you like. NO WARRANTIES.
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
13
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
14 """
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
15 import re,sys
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
16 import TextTools
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
17
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
18 entityRE = re.compile('<!ENTITY +(\w+) +CDATA +"([^"]+)" +-- +((?:.|\n)+?) *-->')
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
19
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
20 def parse(text,pos=0,endpos=None):
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
21
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
22 pos = 0
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
23 if endpos is None:
14657
946709e82be0 satisfy the tabnanny
Jeremy Hylton <jeremy@alum.mit.edu>
parents: 11210
diff changeset
24 endpos = len(text)
11210
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
25 d = {}
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
26 while 1:
14657
946709e82be0 satisfy the tabnanny
Jeremy Hylton <jeremy@alum.mit.edu>
parents: 11210
diff changeset
27 m = entityRE.search(text,pos,endpos)
946709e82be0 satisfy the tabnanny
Jeremy Hylton <jeremy@alum.mit.edu>
parents: 11210
diff changeset
28 if not m:
946709e82be0 satisfy the tabnanny
Jeremy Hylton <jeremy@alum.mit.edu>
parents: 11210
diff changeset
29 break
946709e82be0 satisfy the tabnanny
Jeremy Hylton <jeremy@alum.mit.edu>
parents: 11210
diff changeset
30 name,charcode,comment = m.groups()
946709e82be0 satisfy the tabnanny
Jeremy Hylton <jeremy@alum.mit.edu>
parents: 11210
diff changeset
31 d[name] = charcode,comment
946709e82be0 satisfy the tabnanny
Jeremy Hylton <jeremy@alum.mit.edu>
parents: 11210
diff changeset
32 pos = m.end()
11210
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
33 return d
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
34
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
35 def writefile(f,defs):
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
36
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
37 f.write("entitydefs = {\n")
47100
893a4dc17a91 Ran 2to3 over scripts directory.
Georg Brandl <georg@python.org>
parents: 15930
diff changeset
38 items = sorted(defs.items())
893a4dc17a91 Ran 2to3 over scripts directory.
Georg Brandl <georg@python.org>
parents: 15930
diff changeset
39 for name, (charcode,comment) in items:
14657
946709e82be0 satisfy the tabnanny
Jeremy Hylton <jeremy@alum.mit.edu>
parents: 11210
diff changeset
40 if charcode[:2] == '&#':
946709e82be0 satisfy the tabnanny
Jeremy Hylton <jeremy@alum.mit.edu>
parents: 11210
diff changeset
41 code = int(charcode[2:-1])
946709e82be0 satisfy the tabnanny
Jeremy Hylton <jeremy@alum.mit.edu>
parents: 11210
diff changeset
42 if code < 256:
946709e82be0 satisfy the tabnanny
Jeremy Hylton <jeremy@alum.mit.edu>
parents: 11210
diff changeset
43 charcode = "'\%o'" % code
946709e82be0 satisfy the tabnanny
Jeremy Hylton <jeremy@alum.mit.edu>
parents: 11210
diff changeset
44 else:
946709e82be0 satisfy the tabnanny
Jeremy Hylton <jeremy@alum.mit.edu>
parents: 11210
diff changeset
45 charcode = repr(charcode)
946709e82be0 satisfy the tabnanny
Jeremy Hylton <jeremy@alum.mit.edu>
parents: 11210
diff changeset
46 else:
946709e82be0 satisfy the tabnanny
Jeremy Hylton <jeremy@alum.mit.edu>
parents: 11210
diff changeset
47 charcode = repr(charcode)
946709e82be0 satisfy the tabnanny
Jeremy Hylton <jeremy@alum.mit.edu>
parents: 11210
diff changeset
48 comment = TextTools.collapse(comment)
946709e82be0 satisfy the tabnanny
Jeremy Hylton <jeremy@alum.mit.edu>
parents: 11210
diff changeset
49 f.write(" '%s':\t%s, \t# %s\n" % (name,charcode,comment))
11210
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
50 f.write('\n}\n')
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
51
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
52 if __name__ == '__main__':
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
53 if len(sys.argv) > 1:
14657
946709e82be0 satisfy the tabnanny
Jeremy Hylton <jeremy@alum.mit.edu>
parents: 11210
diff changeset
54 infile = open(sys.argv[1])
11210
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
55 else:
14657
946709e82be0 satisfy the tabnanny
Jeremy Hylton <jeremy@alum.mit.edu>
parents: 11210
diff changeset
56 infile = sys.stdin
11210
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
57 if len(sys.argv) > 2:
14657
946709e82be0 satisfy the tabnanny
Jeremy Hylton <jeremy@alum.mit.edu>
parents: 11210
diff changeset
58 outfile = open(sys.argv[2],'w')
11210
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
59 else:
14657
946709e82be0 satisfy the tabnanny
Jeremy Hylton <jeremy@alum.mit.edu>
parents: 11210
diff changeset
60 outfile = sys.stdout
11210
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
61 text = infile.read()
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
62 defs = parse(text)
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
63 writefile(outfile,defs)