annotate Tools/scripts/parseentities.py @ 11239:577ea261bdf2 legacy-trunk

Find the help.txt file relative to __file__ or ".", not in sys.path. (Suggested by Moshe Zadka, but implemented differently.) Add <<python-docs>> event which, on Unix, brings up Netscape pointing to http://www.python.doc/current/ (a local copy would be nice but its location can't be predicted). Windows solution TBD.
author Guido van Rossum <guido@python.org>
date Thu, 26 Aug 1999 23:06:05 +0000
parents cf7cd7b04fa6
children 946709e82be0
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
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
11 Marc-Andre Lemburg, mal@lemburg.com, 1999.
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:
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
24 endpos = len(text)
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:
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
27 m = entityRE.search(text,pos,endpos)
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
28 if not m:
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
29 break
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
30 name,charcode,comment = m.groups()
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
31 d[name] = charcode,comment
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
32 pos = m.end()
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")
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
38 items = defs.items()
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
39 items.sort()
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
40 for name,(charcode,comment) in items:
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
41 if charcode[:2] == '&#':
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
42 code = int(charcode[2:-1])
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
43 if code < 256:
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
44 charcode = "'\%o'" % code
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
45 else:
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
46 charcode = repr(charcode)
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
47 else:
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
48 charcode = repr(charcode)
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
49 comment = TextTools.collapse(comment)
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
50 f.write(" '%s':\t%s, \t# %s\n" % (name,charcode,comment))
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
51 f.write('\n}\n')
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
52
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
53 if __name__ == '__main__':
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
54 if len(sys.argv) > 1:
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
55 infile = open(sys.argv[1])
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
56 else:
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
57 infile = sys.stdin
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
58 if len(sys.argv) > 2:
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
59 outfile = open(sys.argv[2],'w')
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
60 else:
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
61 outfile = sys.stdout
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
62 text = infile.read()
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
63 defs = parse(text)
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
64 writefile(outfile,defs)
cf7cd7b04fa6 Scrit by Marc-Andre Lemburg to generate htmlentitydefs.py.
Guido van Rossum <guido@python.org>
parents:
diff changeset
65