user-tools/url-parts

   1 #!/usr/bin/env python
   2
   3 """
   4 =pod
   5
   6 =head1 NAME
   7
   8 url-parts - Extract specified parts from URLs given in input stream
   9
  10 =head1 SYNOPSIS
  11
  12 echo <URL> | url-parts <PART> [<PART> [<PART> [...]]]
  13
  14 =head1 DESCRIPTION
  15
  16 Supported parts:
  17 fragment, hostname, netloc, password, path, port, query, scheme, username,
  18 and query.I<NAME> for the query parameter I<NAME>,
  19 and query.I<NAME>.I<N> for I<N>th element of the array parameter I<NAME>.
  20
  21 Run C<url-parts --help> for the definitive list of URL part names
  22 supported by the python urlparse module installed on your system.
  23
  24 =cut
  25
  26 """
  27
  28 from __future__ import print_function
  29 import sys
  30 import re
  31 import urlparse
  32
  33 delimiter = '\t'
  34
  35 if '--help' in sys.argv:
  36         sample = urlparse.urlsplit('')
  37         print("Usage: echo URL | url-parts [PART-1 [PART-2 [...]]]")
  38         print("PART-N is one of:", ' '.join([attr for attr in dir(sample)
  39                 if not attr.startswith('_') and not hasattr(getattr(sample, attr), '__call__')]))
  40         print("or 'query.PARAM' to access individual query string parameters,")
  41         print("or 'query.PARAM.N' where N is an integer to access array-like parameters.")
  42         sys.exit(0)
  43
  44 while True:
  45         line = sys.stdin.readline()
  46         if line == '':
  47                 break
  48         line = line.strip()
  49         url = urlparse.urlsplit(line)
  50         was_output = False
  51         for param_name, param_values in urlparse.parse_qs(url.query).iteritems():
  52                 field_name = 'query.%s' % param_name
  53                 if len(param_values) == 1:
  54                         setattr(url, field_name, param_values[0])
  55                 for index, param_value in enumerate(param_values):
  56                         setattr(url, '%s.%d' % (field_name, index), param_value)
  57         for arg in sys.argv[1:]:
  58                 if hasattr(url, arg):
  59                         value = getattr(url, arg)
  60                 else:
  61                         value = ''
  62                 if was_output:
  63                         print(delimiter, end='')
  64                 print(value, end='')
  65                 was_output = True
  66         print('')