lib/deserialize.py

   1 # this code comes from ABE. it can probably be simplified
   2 #
   3 #
   4
   5 from bitcoin import public_key_to_bc_address, hash_160_to_bc_address, hash_encode, hash_160
   6 #import socket
   7 import time
   8 import struct
   9
  10 #
  11 # Workalike python implementation of Bitcoin's CDataStream class.
  12 #
  13 import struct
  14 import StringIO
  15 import mmap
  16
  17 class SerializationError(Exception):
  18     """ Thrown when there's a problem deserializing or serializing """
  19
  20 class BCDataStream(object):
  21     def __init__(self):
  22         self.input = None
  23         self.read_cursor = 0
  24
  25     def clear(self):
  26         self.input = None
  27         self.read_cursor = 0
  28
  29     def write(self, bytes):  # Initialize with string of bytes
  30         if self.input is None:
  31             self.input = bytes
  32         else:
  33             self.input += bytes
  34
  35     def map_file(self, file, start):  # Initialize with bytes from file
  36         self.input = mmap.mmap(file.fileno(), 0, access=mmap.ACCESS_READ)
  37         self.read_cursor = start
  38
  39     def seek_file(self, position):
  40         self.read_cursor = position
  41
  42     def close_file(self):
  43         self.input.close()
  44
  45     def read_string(self):
  46         # Strings are encoded depending on length:
  47         # 0 to 252 :  1-byte-length followed by bytes (if any)
  48         # 253 to 65,535 : byte'253' 2-byte-length followed by bytes
  49         # 65,536 to 4,294,967,295 : byte '254' 4-byte-length followed by bytes
  50         # ... and the Bitcoin client is coded to understand:
  51         # greater than 4,294,967,295 : byte '255' 8-byte-length followed by bytes of string
  52         # ... but I don't think it actually handles any strings that big.
  53         if self.input is None:
  54             raise SerializationError("call write(bytes) before trying to deserialize")
  55
  56         try:
  57             length = self.read_compact_size()
  58         except IndexError:
  59             raise SerializationError("attempt to read past end of buffer")
  60
  61         return self.read_bytes(length)
  62
  63     def write_string(self, string):
  64         # Length-encoded as with read-string
  65         self.write_compact_size(len(string))
  66         self.write(string)
  67
  68     def read_bytes(self, length):
  69         try:
  70             result = self.input[self.read_cursor:self.read_cursor+length]
  71             self.read_cursor += length
  72             return result
  73         except IndexError:
  74             raise SerializationError("attempt to read past end of buffer")
  75
  76         return ''
  77
  78     def read_boolean(self): return self.read_bytes(1)[0] != chr(0)
  79     def read_int16(self): return self._read_num('<h')
  80     def read_uint16(self): return self._read_num('<H')
  81     def read_int32(self): return self._read_num('<i')
  82     def read_uint32(self): return self._read_num('<I')
  83     def read_int64(self): return self._read_num('<q')
  84     def read_uint64(self): return self._read_num('<Q')
  85
  86     def write_boolean(self, val): return self.write(chr(1) if val else chr(0))
  87     def write_int16(self, val): return self._write_num('<h', val)
  88     def write_uint16(self, val): return self._write_num('<H', val)
  89     def write_int32(self, val): return self._write_num('<i', val)
  90     def write_uint32(self, val): return self._write_num('<I', val)
  91     def write_int64(self, val): return self._write_num('<q', val)
  92     def write_uint64(self, val): return self._write_num('<Q', val)
  93
  94     def read_compact_size(self):
  95         size = ord(self.input[self.read_cursor])
  96         self.read_cursor += 1
  97         if size == 253:
  98             size = self._read_num('<H')
  99         elif size == 254:
 100             size = self._read_num('<I')
 101         elif size == 255:
 102             size = self._read_num('<Q')
 103         return size
 104
 105     def write_compact_size(self, size):
 106         if size < 0:
 107             raise SerializationError("attempt to write size < 0")
 108         elif size < 253:
 109             self.write(chr(size))
 110         elif size < 2**16:
 111             self.write('\xfd')
 112             self._write_num('<H', size)
 113         elif size < 2**32:
 114             self.write('\xfe')
 115             self._write_num('<I', size)
 116         elif size < 2**64:
 117             self.write('\xff')
 118             self._write_num('<Q', size)
 119
 120     def _read_num(self, format):
 121         (i,) = struct.unpack_from(format, self.input, self.read_cursor)
 122         self.read_cursor += struct.calcsize(format)
 123         return i
 124
 125     def _write_num(self, format, num):
 126         s = struct.pack(format, num)
 127         self.write(s)
 128
 129 #
 130 # enum-like type
 131 # From the Python Cookbook, downloaded from http://code.activestate.com/recipes/67107/
 132 #
 133 import types, string, exceptions
 134
 135 class EnumException(exceptions.Exception):
 136     pass
 137
 138 class Enumeration:
 139     def __init__(self, name, enumList):
 140         self.__doc__ = name
 141         lookup = { }
 142         reverseLookup = { }
 143         i = 0
 144         uniqueNames = [ ]
 145         uniqueValues = [ ]
 146         for x in enumList:
 147             if type(x) == types.TupleType:
 148                 x, i = x
 149             if type(x) != types.StringType:
 150                 raise EnumException, "enum name is not a string: " + x
 151             if type(i) != types.IntType:
 152                 raise EnumException, "enum value is not an integer: " + i
 153             if x in uniqueNames:
 154                 raise EnumException, "enum name is not unique: " + x
 155             if i in uniqueValues:
 156                 raise EnumException, "enum value is not unique for " + x
 157             uniqueNames.append(x)
 158             uniqueValues.append(i)
 159             lookup[x] = i
 160             reverseLookup[i] = x
 161             i = i + 1
 162         self.lookup = lookup
 163         self.reverseLookup = reverseLookup
 164     def __getattr__(self, attr):
 165         if not self.lookup.has_key(attr):
 166             raise AttributeError
 167         return self.lookup[attr]
 168     def whatis(self, value):
 169         return self.reverseLookup[value]
 170
 171
 172 # This function comes from bitcointools, bct-LICENSE.txt.
 173 def long_hex(bytes):
 174     return bytes.encode('hex_codec')
 175
 176 # This function comes from bitcointools, bct-LICENSE.txt.
 177 def short_hex(bytes):
 178     t = bytes.encode('hex_codec')
 179     if len(t) < 11:
 180         return t
 181     return t[0:4]+"..."+t[-4:]
 182
 183
 184
 185 def parse_TxIn(vds):
 186     d = {}
 187     d['prevout_hash'] = hash_encode(vds.read_bytes(32))
 188     d['prevout_n'] = vds.read_uint32()
 189     scriptSig = vds.read_bytes(vds.read_compact_size())
 190     d['sequence'] = vds.read_uint32()
 191
 192     if scriptSig:
 193         pubkeys, signatures, address = get_address_from_input_script(scriptSig)
 194     else:
 195         pubkeys = []
 196         signatures = []
 197         address = None
 198
 199     d['address'] = address
 200     d['signatures'] = signatures
 201
 202     return d
 203
 204
 205 def parse_TxOut(vds, i):
 206     d = {}
 207     d['value'] = vds.read_int64()
 208     scriptPubKey = vds.read_bytes(vds.read_compact_size())
 209     d['address'] = get_address_from_output_script(scriptPubKey)
 210     d['raw_output_script'] = scriptPubKey.encode('hex')
 211     d['index'] = i
 212     return d
 213
 214
 215 def parse_Transaction(vds):
 216     d = {}
 217     start = vds.read_cursor
 218     d['version'] = vds.read_int32()
 219     n_vin = vds.read_compact_size()
 220     d['inputs'] = []
 221     for i in xrange(n_vin):
 222         d['inputs'].append(parse_TxIn(vds))
 223     n_vout = vds.read_compact_size()
 224     d['outputs'] = []
 225     for i in xrange(n_vout):
 226         d['outputs'].append(parse_TxOut(vds, i))
 227     d['lockTime'] = vds.read_uint32()
 228     return d
 229
 230 def parse_redeemScript(bytes):
 231     dec = [ x for x in script_GetOp(bytes.decode('hex')) ]
 232
 233     # 2 of 2
 234     match = [ opcodes.OP_2, opcodes.OP_PUSHDATA4, opcodes.OP_PUSHDATA4, opcodes.OP_2, opcodes.OP_CHECKMULTISIG ]
 235     if match_decoded(dec, match):
 236         pubkeys = [ dec[1][1].encode('hex'), dec[2][1].encode('hex') ]
 237         return 2, pubkeys
 238
 239     # 2 of 3
 240     match = [ opcodes.OP_2, opcodes.OP_PUSHDATA4, opcodes.OP_PUSHDATA4, opcodes.OP_PUSHDATA4, opcodes.OP_3, opcodes.OP_CHECKMULTISIG ]
 241     if match_decoded(dec, match):
 242         pubkeys = [ dec[1][1].encode('hex'), dec[2][1].encode('hex'), dec[3][1].encode('hex') ]
 243         return 3, pubkeys
 244
 245
 246
 247 opcodes = Enumeration("Opcodes", [
 248     ("OP_0", 0), ("OP_PUSHDATA1",76), "OP_PUSHDATA2", "OP_PUSHDATA4", "OP_1NEGATE", "OP_RESERVED",
 249     "OP_1", "OP_2", "OP_3", "OP_4", "OP_5", "OP_6", "OP_7",
 250     "OP_8", "OP_9", "OP_10", "OP_11", "OP_12", "OP_13", "OP_14", "OP_15", "OP_16",
 251     "OP_NOP", "OP_VER", "OP_IF", "OP_NOTIF", "OP_VERIF", "OP_VERNOTIF", "OP_ELSE", "OP_ENDIF", "OP_VERIFY",
 252     "OP_RETURN", "OP_TOALTSTACK", "OP_FROMALTSTACK", "OP_2DROP", "OP_2DUP", "OP_3DUP", "OP_2OVER", "OP_2ROT", "OP_2SWAP",
 253     "OP_IFDUP", "OP_DEPTH", "OP_DROP", "OP_DUP", "OP_NIP", "OP_OVER", "OP_PICK", "OP_ROLL", "OP_ROT",
 254     "OP_SWAP", "OP_TUCK", "OP_CAT", "OP_SUBSTR", "OP_LEFT", "OP_RIGHT", "OP_SIZE", "OP_INVERT", "OP_AND",
 255     "OP_OR", "OP_XOR", "OP_EQUAL", "OP_EQUALVERIFY", "OP_RESERVED1", "OP_RESERVED2", "OP_1ADD", "OP_1SUB", "OP_2MUL",
 256     "OP_2DIV", "OP_NEGATE", "OP_ABS", "OP_NOT", "OP_0NOTEQUAL", "OP_ADD", "OP_SUB", "OP_MUL", "OP_DIV",
 257     "OP_MOD", "OP_LSHIFT", "OP_RSHIFT", "OP_BOOLAND", "OP_BOOLOR",
 258     "OP_NUMEQUAL", "OP_NUMEQUALVERIFY", "OP_NUMNOTEQUAL", "OP_LESSTHAN",
 259     "OP_GREATERTHAN", "OP_LESSTHANOREQUAL", "OP_GREATERTHANOREQUAL", "OP_MIN", "OP_MAX",
 260     "OP_WITHIN", "OP_RIPEMD160", "OP_SHA1", "OP_SHA256", "OP_HASH160",
 261     "OP_HASH256", "OP_CODESEPARATOR", "OP_CHECKSIG", "OP_CHECKSIGVERIFY", "OP_CHECKMULTISIG",
 262     "OP_CHECKMULTISIGVERIFY",
 263     ("OP_SINGLEBYTE_END", 0xF0),
 264     ("OP_DOUBLEBYTE_BEGIN", 0xF000),
 265     "OP_PUBKEY", "OP_PUBKEYHASH",
 266     ("OP_INVALIDOPCODE", 0xFFFF),
 267 ])
 268
 269
 270 def script_GetOp(bytes):
 271     i = 0
 272     while i < len(bytes):
 273         vch = None
 274         opcode = ord(bytes[i])
 275         i += 1
 276         if opcode >= opcodes.OP_SINGLEBYTE_END:
 277             opcode <<= 8
 278             opcode |= ord(bytes[i])
 279             i += 1
 280
 281         if opcode <= opcodes.OP_PUSHDATA4:
 282             nSize = opcode
 283             if opcode == opcodes.OP_PUSHDATA1:
 284                 nSize = ord(bytes[i])
 285                 i += 1
 286             elif opcode == opcodes.OP_PUSHDATA2:
 287                 (nSize,) = struct.unpack_from('<H', bytes, i)
 288                 i += 2
 289             elif opcode == opcodes.OP_PUSHDATA4:
 290                 (nSize,) = struct.unpack_from('<I', bytes, i)
 291                 i += 4
 292             vch = bytes[i:i+nSize]
 293             i += nSize
 294
 295         yield (opcode, vch, i)
 296
 297
 298 def script_GetOpName(opcode):
 299     return (opcodes.whatis(opcode)).replace("OP_", "")
 300
 301
 302 def decode_script(bytes):
 303     result = ''
 304     for (opcode, vch, i) in script_GetOp(bytes):
 305         if len(result) > 0: result += " "
 306         if opcode <= opcodes.OP_PUSHDATA4:
 307             result += "%d:"%(opcode,)
 308             result += short_hex(vch)
 309         else:
 310             result += script_GetOpName(opcode)
 311     return result
 312
 313
 314 def match_decoded(decoded, to_match):
 315     if len(decoded) != len(to_match):
 316         return False;
 317     for i in range(len(decoded)):
 318         if to_match[i] == opcodes.OP_PUSHDATA4 and decoded[i][0] <= opcodes.OP_PUSHDATA4 and decoded[i][0]>0:
 319             continue  # Opcodes below OP_PUSHDATA4 all just push data onto stack, and are equivalent.
 320         if to_match[i] != decoded[i][0]:
 321             return False
 322     return True
 323
 324 def get_address_from_input_script(bytes):
 325     decoded = [ x for x in script_GetOp(bytes) ]
 326
 327     # non-generated TxIn transactions push a signature
 328     # (seventy-something bytes) and then their public key
 329     # (65 bytes) onto the stack:
 330     match = [ opcodes.OP_PUSHDATA4, opcodes.OP_PUSHDATA4 ]
 331     if match_decoded(decoded, match):
 332         return None, None, public_key_to_bc_address(decoded[1][1])
 333
 334     # p2sh transaction, 2 of n
 335     match = [ opcodes.OP_0 ]
 336     while len(match) < len(decoded):
 337         match.append(opcodes.OP_PUSHDATA4)
 338
 339     if match_decoded(decoded, match):
 340
 341         redeemScript = decoded[-1][1]
 342         num = len(match) - 2
 343         signatures = map(lambda x:x[1].encode('hex'), decoded[1:-1])
 344
 345         dec2 = [ x for x in script_GetOp(redeemScript) ]
 346
 347         # 2 of 2
 348         match2 = [ opcodes.OP_2, opcodes.OP_PUSHDATA4, opcodes.OP_PUSHDATA4, opcodes.OP_2, opcodes.OP_CHECKMULTISIG ]
 349         if match_decoded(dec2, match2):
 350             pubkeys = [ dec2[1][1].encode('hex'), dec2[2][1].encode('hex') ]
 351             return pubkeys, signatures, hash_160_to_bc_address(hash_160(redeemScript), 5)
 352
 353         # 2 of 3
 354         match2 = [ opcodes.OP_2, opcodes.OP_PUSHDATA4, opcodes.OP_PUSHDATA4, opcodes.OP_PUSHDATA4, opcodes.OP_3, opcodes.OP_CHECKMULTISIG ]
 355         if match_decoded(dec2, match2):
 356             pubkeys = [ dec2[1][1].encode('hex'), dec2[2][1].encode('hex'), dec2[3][1].encode('hex') ]
 357             s = multisig_script(pubkeys)
 358             return pubkeys, signatures, hash_160_to_bc_address(hash_160(redeemScript), 5)
 359
 360     raise BaseException("no match for scriptsig")
 361
 362
 363
 364 def get_address_from_output_script(bytes):
 365     decoded = [ x for x in script_GetOp(bytes) ]
 366
 367     # The Genesis Block, self-payments, and pay-by-IP-address payments look like:
 368     # 65 BYTES:... CHECKSIG
 369     match = [ opcodes.OP_PUSHDATA4, opcodes.OP_CHECKSIG ]
 370     if match_decoded(decoded, match):
 371         return public_key_to_bc_address(decoded[0][1])
 372
 373     # Pay-by-Bitcoin-address TxOuts look like:
 374     # DUP HASH160 20 BYTES:... EQUALVERIFY CHECKSIG
 375     match = [ opcodes.OP_DUP, opcodes.OP_HASH160, opcodes.OP_PUSHDATA4, opcodes.OP_EQUALVERIFY, opcodes.OP_CHECKSIG ]
 376     if match_decoded(decoded, match):
 377         return hash_160_to_bc_address(decoded[2][1])
 378
 379     # p2sh
 380     match = [ opcodes.OP_HASH160, opcodes.OP_PUSHDATA4, opcodes.OP_EQUAL ]
 381     if match_decoded(decoded, match):
 382         return hash_160_to_bc_address(decoded[1][1],5)
 383
 384     return "(None)"
 385
 386