tests/uuidfilt.py

   1 #!/usr/bin/env python
   2
   3 import re
   4 import sys
   5
   6
   7 def lookup_uuid(uuids, match):
   8     return "<%s>" % uuids.setdefault(match.group(0), len(uuids))
   9
  10
  11 int_re = re.compile(r'\d+')
  12
  13
  14 def sort_set(match):
  15     s = match.group(0)
  16     uuids = sorted([int(x) for x in int_re.findall(s)])
  17     return '["set",[' + ','.join('["uuid","<%s>"]' % x for x in uuids) + ']]'
  18
  19
  20 u = '[0-9a-fA-F]'
  21 uuid_re = re.compile(r'%s{8}-%s{4}-%s{4}-%s{4}-%s{12}' % ((u,) * 5))
  22 set_re = re.compile(r'(\["set",\[(,?\["uuid","<\d+>"\])+\]\])')
  23
  24
  25 def filter_uuids(src, dst):
  26     uuids = {}
  27
  28     def lf(match):
  29         return lookup_uuid(uuids, match)
  30
  31     while True:
  32         line = src.readline()
  33         if not line:
  34             break
  35         line = uuid_re.sub(lf, line)
  36
  37         # Sort sets like this:
  38         #    [["uuid","<1>"],["uuid","<0>"]]
  39         # to look like this:
  40         #    [["uuid","<0>"],["uuid","<1>"]]
  41         line = set_re.sub(sort_set, line)
  42         dst.write(line)
  43
  44
  45 if __name__ == '__main__':
  46     if len(sys.argv) > 1:
  47         for src in sys.argv[1:]:
  48             filter_uuids(open(src), sys.stdout)
  49     else:
  50         filter_uuids(sys.stdin, sys.stdout)