util/PHPUnserialize.py

   1 import types, string, re
   2
   3 """
   4 Unserialize class for the PHP serialization format.
   5
   6 @version v0.4 BETA
   7 @author Scott Hurring; scott at hurring dot com
   8 @copyright Copyright (c) 2005 Scott Hurring
   9 @license http://opensource.org/licenses/gpl-license.php GNU Public License
  10 $Id: PHPUnserialize.py,v 1.1 2006/01/08 21:53:19 shurring Exp $
  11
  12 Most recent version can be found at:
  13 http://hurring.com/code/python/phpserialize/
  14
  15 Usage:
  16 # Create an instance of the unserialize engine
  17 u = PHPUnserialize()
  18 # unserialize some string into python data
  19 data = u.unserialize(serialized_string)
  20
  21 Please see README.txt for more information.
  22 """
  23
  24 class PHPUnserialize(object):
  25         """
  26         Class to unserialize something from the PHP Serialize format.
  27
  28         Usage:
  29         u = PHPUnserialize()
  30         data = u.unserialize(serialized_string)
  31         """
  32
  33         def __init__(self):
  34                 pass
  35
  36         def session_decode(self, data):
  37                 """Thanks to Ken Restivo for suggesting the addition
  38                 of session_encode
  39                 """
  40                 session = {}
  41                 while len(data) > 0:
  42                         m = re.match('^(\w+)\|', data)
  43                         if m:
  44                                 key = m.group(1)
  45                                 offset = len(key)+1
  46                                 (dtype, dataoffset, value) = self._unserialize(data, offset)
  47                                 offset = offset + dataoffset
  48                                 data = data[offset:]
  49                                 session[key] = value
  50                         else:
  51                                 # No more stuff to decode
  52                                 return session
  53
  54                 return session
  55
  56         def unserialize(self, data):
  57                 return self._unserialize(data, 0)[2]
  58
  59         def _unserialize(self, data, offset=0):
  60                 """
  61                 Find the next token and unserialize it.
  62                 Recurse on array.
  63
  64                 offset = raw offset from start of data
  65
  66                 return (type, offset, value)
  67                 """
  68
  69                 buf = []
  70                 dtype = string.lower(data[offset:offset+1])
  71
  72                 #print "# dtype =", dtype
  73
  74                 # 't:' = 2 chars
  75                 dataoffset = offset + 2
  76                 typeconvert = lambda x : x
  77                 chars = datalength = 0
  78
  79                 # int => Integer
  80                 if dtype == 'i':
  81                         typeconvert = lambda x : int(x)
  82                         (chars, readdata) = self.read_until(data, dataoffset, ';')
  83                         # +1 for end semicolon
  84                         dataoffset += chars + 1
  85
  86                 # bool => Boolean
  87                 elif dtype == 'b':
  88                         typeconvert = lambda x : (int(x) == 1)
  89                         (chars, readdata) = self.read_until(data, dataoffset, ';')
  90                         # +1 for end semicolon
  91                         dataoffset += chars + 1
  92
  93                 # double => Floating Point
  94                 elif dtype == 'd':
  95                         typeconvert = lambda x : float(x)
  96                         (chars, readdata) = self.read_until(data, dataoffset, ';')
  97                         # +1 for end semicolon
  98                         dataoffset += chars + 1
  99
 100                 # n => None
 101                 elif dtype == 'n':
 102                         readdata = None
 103
 104                 # s => String
 105                 elif dtype == 's':
 106                         (chars, stringlength) = self.read_until(data, dataoffset, ':')
 107                         # +2 for colons around length field
 108                         dataoffset += chars + 2
 109
 110                         # +1 for start quote
 111                         (chars, readdata) = self.read_chars(data, dataoffset+1, int(stringlength))
 112                         # +2 for endquote semicolon
 113                         dataoffset += chars + 2
 114
 115                         if chars != int(stringlength) != int(readdata):
 116                                 raise Exception("String length mismatch")
 117
 118                 # array => Dict
 119                 # If you originally serialized a Tuple or List, it will
 120                 # be unserialized as a Dict.  PHP doesn't have tuples or lists,
 121                 # only arrays - so everything has to get converted into an array
 122                 # when serializing and the original type of the array is lost
 123                 elif dtype == 'a':
 124                         readdata = {}
 125
 126                         # If all dict keys are increasing ints from zero,
 127                         #       then make it a list.
 128                         # Else,
 129                         #       just return the original dict
 130                         def is_list(keys):
 131                                 # list with no elements
 132                                 if len(keys) == 0:
 133                                         return True
 134                                 # list with one element: 0
 135                                 if len(keys) == 1:
 136                                         if type(keys[0]) == int and keys[0] == 0:
 137                                                 return True
 138                                         else:
 139                                                 return False
 140
 141                                 # all other cases
 142                                 for i in range(1,len(keys)):
 143                                         x = keys[i-1]
 144                                         y = keys[i]
 145                                         if type(x) == int and type(y) == int and y-x == 1:
 146                                                 continue
 147                                         else:
 148                                                 return False
 149                                 return True
 150
 151                         # Empty {} and empty [] are ambiguous.
 152                         # A default could either be configured or assumed to be a list?
 153                         def dict_to_list(dict):
 154
 155                                 keys = dict.keys()
 156                                 keys.sort()
 157
 158                                 if is_list(keys):
 159                                         return [dict[k] for k in keys]
 160                                 else:
 161                                         return dict
 162
 163                         typeconvert = lambda x : dict_to_list(x)
 164
 165                         # How many keys does this list have?
 166                         (chars, keys) = self.read_until(data, dataoffset, ':')
 167                         # +2 for colons around length field
 168                         dataoffset += chars + 2
 169
 170                         # Loop through and fetch this number of key/value pairs
 171                         for i in range(0, int(keys)):
 172                                 # Read the key
 173                                 (ktype, kchars, key) = self._unserialize(data, dataoffset)
 174                                 dataoffset += kchars
 175                                 #print "Key(%i) = (%s, %i, %s) %i" % (i, ktype, kchars, key, dataoffset)
 176
 177                                 # Read value of the key
 178                                 (vtype, vchars, value) = self._unserialize(data, dataoffset)
 179                                 dataoffset += vchars
 180                                 #print "Value(%i) = (%s, %i, %s) %i" % (i, vtype, vchars, value, dataoffset)
 181
 182                                 # Set the list element
 183                                 readdata[key] = value
 184
 185                                 # +1 for end semicolon
 186                         dataoffset += 1
 187                         #chars = int(dataoffset) - start
 188
 189                 # I don't know how to unserialize this
 190                 else:
 191                         raise Exception("Unknown / Unhandled data type (%s)!" % dtype)
 192
 193
 194                 return (dtype, dataoffset-offset, typeconvert(readdata))
 195
 196         def read_until(self, data, offset, stopchar):
 197                 """
 198                 Read from data[offset] until you encounter some char 'stopchar'.
 199                 """
 200                 buf = []
 201                 char = data[offset:offset+1]
 202                 i = 2
 203                 while char != stopchar:
 204                         # Consumed all the characters and havent found ';'
 205                         if i+offset > len(data):
 206                                 raise Exception("Invalid")
 207                         buf.append(char)
 208                         char = data[offset+(i-1):offset+i]
 209                         i += 1
 210
 211                 # (chars_read, data)
 212                 return (len(buf), "".join(buf))
 213
 214         def read_chars(self, data, offset, length):
 215                 """
 216                 Read 'length' number of chars from data[offset].
 217                 """
 218                 buf = []
 219                 # Account for the starting quote char
 220                 #offset += 1
 221                 for i in range(0, length):
 222                         char = data[offset+(i-1):offset+i]
 223                         buf.append(char)
 224
 225                 # (chars_read, data)
 226                 return (len(buf), "".join(buf))
 227
 228