Demo/scripts/find-uname.py

   1 #!/usr/bin/env python
   2
   3 """
   4 For each argument on the command line, look for it in the set of all Unicode
   5 names.  Arguments are treated as case-insensitive regular expressions, e.g.:
   6
   7     % find-uname 'small letter a$' 'horizontal line'
   8     *** small letter a$ matches ***
   9     LATIN SMALL LETTER A (97)
  10     COMBINING LATIN SMALL LETTER A (867)
  11     CYRILLIC SMALL LETTER A (1072)
  12     PARENTHESIZED LATIN SMALL LETTER A (9372)
  13     CIRCLED LATIN SMALL LETTER A (9424)
  14     FULLWIDTH LATIN SMALL LETTER A (65345)
  15     *** horizontal line matches ***
  16     HORIZONTAL LINE EXTENSION (9135)
  17 """
  18
  19 import unicodedata
  20 import sys
  21 import re
  22
  23 def main(args):
  24     unicode_names = []
  25     for ix in range(sys.maxunicode+1):
  26         try:
  27             unicode_names.append((ix, unicodedata.name(unichr(ix))))
  28         except ValueError: # no name for the character
  29             pass
  30     for arg in args:
  31         pat = re.compile(arg, re.I)
  32         matches = [(y,x) for (x,y) in unicode_names
  33                    if pat.search(y) is not None]
  34         if matches:
  35             print "***", arg, "matches", "***"
  36             for match in matches:
  37                 print "%s (%d)" % match
  38
  39 if __name__ == "__main__":
  40     main(sys.argv[1:])