new bitcoind backend (experimental)
[electrum-server.git] / backends / bitcoind / deserialize.py
diff --git a/backends/bitcoind/deserialize.py b/backends/bitcoind/deserialize.py
new file mode 100644 (file)
index 0000000..d92a2b5
--- /dev/null
@@ -0,0 +1,390 @@
+# this code comes from ABE. it can probably be simplified
+#
+#
+
+#from bitcoin import public_key_to_bc_address, hash_160_to_bc_address, hash_encode
+#import socket
+import time, hashlib
+import struct
+addrtype = 0
+
+
+Hash = lambda x: hashlib.sha256(hashlib.sha256(x).digest()).digest()
+hash_encode = lambda x: x[::-1].encode('hex')
+hash_decode = lambda x: x.decode('hex')[::-1]
+
+def hash_160(public_key):
+    md = hashlib.new('ripemd160')
+    md.update(hashlib.sha256(public_key).digest())
+    return md.digest()
+
+def public_key_to_bc_address(public_key):
+    h160 = hash_160(public_key)
+    return hash_160_to_bc_address(h160)
+
+def hash_160_to_bc_address(h160):
+    vh160 = chr(addrtype) + h160
+    h = Hash(vh160)
+    addr = vh160 + h[0:4]
+    return b58encode(addr)
+
+__b58chars = '123456789ABCDEFGHJKLMNPQRSTUVWXYZabcdefghijkmnopqrstuvwxyz'
+__b58base = len(__b58chars)
+
+def b58encode(v):
+    """ encode v, which is a string of bytes, to base58."""
+
+    long_value = 0L
+    for (i, c) in enumerate(v[::-1]):
+        long_value += (256**i) * ord(c)
+
+    result = ''
+    while long_value >= __b58base:
+        div, mod = divmod(long_value, __b58base)
+        result = __b58chars[mod] + result
+        long_value = div
+    result = __b58chars[long_value] + result
+
+    # Bitcoin does a little leading-zero-compression:
+    # leading 0-bytes in the input become leading-1s
+    nPad = 0
+    for c in v:
+        if c == '\0': nPad += 1
+        else: break
+
+    return (__b58chars[0]*nPad) + result
+
+def b58decode(v, length):
+    """ decode v into a string of len bytes."""
+    long_value = 0L
+    for (i, c) in enumerate(v[::-1]):
+        long_value += __b58chars.find(c) * (__b58base**i)
+
+    result = ''
+    while long_value >= 256:
+        div, mod = divmod(long_value, 256)
+        result = chr(mod) + result
+        long_value = div
+    result = chr(long_value) + result
+
+    nPad = 0
+    for c in v:
+        if c == __b58chars[0]: nPad += 1
+        else: break
+
+    result = chr(0)*nPad + result
+    if length is not None and len(result) != length:
+        return None
+
+    return result
+
+
+#
+# Workalike python implementation of Bitcoin's CDataStream class.
+#
+import struct
+import StringIO
+import mmap
+
+class SerializationError(Exception):
+  """ Thrown when there's a problem deserializing or serializing """
+
+class BCDataStream(object):
+  def __init__(self):
+    self.input = None
+    self.read_cursor = 0
+
+  def clear(self):
+    self.input = None
+    self.read_cursor = 0
+
+  def write(self, bytes):  # Initialize with string of bytes
+    if self.input is None:
+      self.input = bytes
+    else:
+      self.input += bytes
+
+  def map_file(self, file, start):  # Initialize with bytes from file
+    self.input = mmap.mmap(file.fileno(), 0, access=mmap.ACCESS_READ)
+    self.read_cursor = start
+  def seek_file(self, position):
+    self.read_cursor = position
+  def close_file(self):
+    self.input.close()
+
+  def read_string(self):
+    # Strings are encoded depending on length:
+    # 0 to 252 :  1-byte-length followed by bytes (if any)
+    # 253 to 65,535 : byte'253' 2-byte-length followed by bytes
+    # 65,536 to 4,294,967,295 : byte '254' 4-byte-length followed by bytes
+    # ... and the Bitcoin client is coded to understand:
+    # greater than 4,294,967,295 : byte '255' 8-byte-length followed by bytes of string
+    # ... but I don't think it actually handles any strings that big.
+    if self.input is None:
+      raise SerializationError("call write(bytes) before trying to deserialize")
+
+    try:
+      length = self.read_compact_size()
+    except IndexError:
+      raise SerializationError("attempt to read past end of buffer")
+
+    return self.read_bytes(length)
+
+  def write_string(self, string):
+    # Length-encoded as with read-string
+    self.write_compact_size(len(string))
+    self.write(string)
+
+  def read_bytes(self, length):
+    try:
+      result = self.input[self.read_cursor:self.read_cursor+length]
+      self.read_cursor += length
+      return result
+    except IndexError:
+      raise SerializationError("attempt to read past end of buffer")
+
+    return ''
+
+  def read_boolean(self): return self.read_bytes(1)[0] != chr(0)
+  def read_int16(self): return self._read_num('<h')
+  def read_uint16(self): return self._read_num('<H')
+  def read_int32(self): return self._read_num('<i')
+  def read_uint32(self): return self._read_num('<I')
+  def read_int64(self): return self._read_num('<q')
+  def read_uint64(self): return self._read_num('<Q')
+
+  def write_boolean(self, val): return self.write(chr(1) if val else chr(0))
+  def write_int16(self, val): return self._write_num('<h', val)
+  def write_uint16(self, val): return self._write_num('<H', val)
+  def write_int32(self, val): return self._write_num('<i', val)
+  def write_uint32(self, val): return self._write_num('<I', val)
+  def write_int64(self, val): return self._write_num('<q', val)
+  def write_uint64(self, val): return self._write_num('<Q', val)
+
+  def read_compact_size(self):
+    size = ord(self.input[self.read_cursor])
+    self.read_cursor += 1
+    if size == 253:
+      size = self._read_num('<H')
+    elif size == 254:
+      size = self._read_num('<I')
+    elif size == 255:
+      size = self._read_num('<Q')
+    return size
+
+  def write_compact_size(self, size):
+    if size < 0:
+      raise SerializationError("attempt to write size < 0")
+    elif size < 253:
+       self.write(chr(size))
+    elif size < 2**16:
+      self.write('\xfd')
+      self._write_num('<H', size)
+    elif size < 2**32:
+      self.write('\xfe')
+      self._write_num('<I', size)
+    elif size < 2**64:
+      self.write('\xff')
+      self._write_num('<Q', size)
+
+  def _read_num(self, format):
+    (i,) = struct.unpack_from(format, self.input, self.read_cursor)
+    self.read_cursor += struct.calcsize(format)
+    return i
+
+  def _write_num(self, format, num):
+    s = struct.pack(format, num)
+    self.write(s)
+
+#
+# enum-like type
+# From the Python Cookbook, downloaded from http://code.activestate.com/recipes/67107/
+#
+import types, string, exceptions
+
+class EnumException(exceptions.Exception):
+    pass
+
+class Enumeration:
+    def __init__(self, name, enumList):
+        self.__doc__ = name
+        lookup = { }
+        reverseLookup = { }
+        i = 0
+        uniqueNames = [ ]
+        uniqueValues = [ ]
+        for x in enumList:
+            if type(x) == types.TupleType:
+                x, i = x
+            if type(x) != types.StringType:
+                raise EnumException, "enum name is not a string: " + x
+            if type(i) != types.IntType:
+                raise EnumException, "enum value is not an integer: " + i
+            if x in uniqueNames:
+                raise EnumException, "enum name is not unique: " + x
+            if i in uniqueValues:
+                raise EnumException, "enum value is not unique for " + x
+            uniqueNames.append(x)
+            uniqueValues.append(i)
+            lookup[x] = i
+            reverseLookup[i] = x
+            i = i + 1
+        self.lookup = lookup
+        self.reverseLookup = reverseLookup
+    def __getattr__(self, attr):
+        if not self.lookup.has_key(attr):
+            raise AttributeError
+        return self.lookup[attr]
+    def whatis(self, value):
+        return self.reverseLookup[value]
+
+
+# This function comes from bitcointools, bct-LICENSE.txt.
+def long_hex(bytes):
+    return bytes.encode('hex_codec')
+
+# This function comes from bitcointools, bct-LICENSE.txt.
+def short_hex(bytes):
+    t = bytes.encode('hex_codec')
+    if len(t) < 11:
+        return t
+    return t[0:4]+"..."+t[-4:]
+
+
+
+def parse_TxIn(vds):
+  d = {}
+  d['prevout_hash'] = hash_encode(vds.read_bytes(32))
+  d['prevout_n'] = vds.read_uint32()
+  scriptSig = vds.read_bytes(vds.read_compact_size())
+  d['sequence'] = vds.read_uint32()
+  d['address'] = extract_public_key(scriptSig)
+  #d['script'] = decode_script(scriptSig)
+  return d
+
+
+def parse_TxOut(vds, i):
+  d = {}
+  d['value'] = vds.read_int64()
+  scriptPubKey = vds.read_bytes(vds.read_compact_size())
+  d['address'] = extract_public_key(scriptPubKey)
+  #d['script'] = decode_script(scriptPubKey)
+  d['raw_output_script'] = scriptPubKey.encode('hex')
+  d['index'] = i
+  return d
+
+
+def parse_Transaction(vds):
+  d = {}
+  start = vds.read_cursor
+  d['version'] = vds.read_int32()
+  n_vin = vds.read_compact_size()
+  d['inputs'] = []
+  for i in xrange(n_vin):
+    d['inputs'].append(parse_TxIn(vds))
+  n_vout = vds.read_compact_size()
+  d['outputs'] = []
+  for i in xrange(n_vout):
+    d['outputs'].append(parse_TxOut(vds, i))
+  d['lockTime'] = vds.read_uint32()
+  return d
+
+
+
+
+opcodes = Enumeration("Opcodes", [
+    ("OP_0", 0), ("OP_PUSHDATA1",76), "OP_PUSHDATA2", "OP_PUSHDATA4", "OP_1NEGATE", "OP_RESERVED",
+    "OP_1", "OP_2", "OP_3", "OP_4", "OP_5", "OP_6", "OP_7",
+    "OP_8", "OP_9", "OP_10", "OP_11", "OP_12", "OP_13", "OP_14", "OP_15", "OP_16",
+    "OP_NOP", "OP_VER", "OP_IF", "OP_NOTIF", "OP_VERIF", "OP_VERNOTIF", "OP_ELSE", "OP_ENDIF", "OP_VERIFY",
+    "OP_RETURN", "OP_TOALTSTACK", "OP_FROMALTSTACK", "OP_2DROP", "OP_2DUP", "OP_3DUP", "OP_2OVER", "OP_2ROT", "OP_2SWAP",
+    "OP_IFDUP", "OP_DEPTH", "OP_DROP", "OP_DUP", "OP_NIP", "OP_OVER", "OP_PICK", "OP_ROLL", "OP_ROT",
+    "OP_SWAP", "OP_TUCK", "OP_CAT", "OP_SUBSTR", "OP_LEFT", "OP_RIGHT", "OP_SIZE", "OP_INVERT", "OP_AND",
+    "OP_OR", "OP_XOR", "OP_EQUAL", "OP_EQUALVERIFY", "OP_RESERVED1", "OP_RESERVED2", "OP_1ADD", "OP_1SUB", "OP_2MUL",
+    "OP_2DIV", "OP_NEGATE", "OP_ABS", "OP_NOT", "OP_0NOTEQUAL", "OP_ADD", "OP_SUB", "OP_MUL", "OP_DIV",
+    "OP_MOD", "OP_LSHIFT", "OP_RSHIFT", "OP_BOOLAND", "OP_BOOLOR",
+    "OP_NUMEQUAL", "OP_NUMEQUALVERIFY", "OP_NUMNOTEQUAL", "OP_LESSTHAN",
+    "OP_GREATERTHAN", "OP_LESSTHANOREQUAL", "OP_GREATERTHANOREQUAL", "OP_MIN", "OP_MAX",
+    "OP_WITHIN", "OP_RIPEMD160", "OP_SHA1", "OP_SHA256", "OP_HASH160",
+    "OP_HASH256", "OP_CODESEPARATOR", "OP_CHECKSIG", "OP_CHECKSIGVERIFY", "OP_CHECKMULTISIG",
+    "OP_CHECKMULTISIGVERIFY",
+    ("OP_SINGLEBYTE_END", 0xF0),
+    ("OP_DOUBLEBYTE_BEGIN", 0xF000),
+    "OP_PUBKEY", "OP_PUBKEYHASH",
+    ("OP_INVALIDOPCODE", 0xFFFF),
+])
+
+def script_GetOp(bytes):
+  i = 0
+  while i < len(bytes):
+    vch = None
+    opcode = ord(bytes[i])
+    i += 1
+    if opcode >= opcodes.OP_SINGLEBYTE_END:
+      opcode <<= 8
+      opcode |= ord(bytes[i])
+      i += 1
+
+    if opcode <= opcodes.OP_PUSHDATA4:
+      nSize = opcode
+      if opcode == opcodes.OP_PUSHDATA1:
+        nSize = ord(bytes[i])
+        i += 1
+      elif opcode == opcodes.OP_PUSHDATA2:
+        (nSize,) = struct.unpack_from('<H', bytes, i)
+        i += 2
+      elif opcode == opcodes.OP_PUSHDATA4:
+        (nSize,) = struct.unpack_from('<I', bytes, i)
+        i += 4
+      vch = bytes[i:i+nSize]
+      i += nSize
+
+    yield (opcode, vch, i)
+
+def script_GetOpName(opcode):
+  return (opcodes.whatis(opcode)).replace("OP_", "")
+
+def decode_script(bytes):
+  result = ''
+  for (opcode, vch, i) in script_GetOp(bytes):
+    if len(result) > 0: result += " "
+    if opcode <= opcodes.OP_PUSHDATA4:
+      result += "%d:"%(opcode,)
+      result += short_hex(vch)
+    else:
+      result += script_GetOpName(opcode)
+  return result
+
+def match_decoded(decoded, to_match):
+  if len(decoded) != len(to_match):
+    return False;
+  for i in range(len(decoded)):
+    if to_match[i] == opcodes.OP_PUSHDATA4 and decoded[i][0] <= opcodes.OP_PUSHDATA4:
+      continue  # Opcodes below OP_PUSHDATA4 all just push data onto stack, and are equivalent.
+    if to_match[i] != decoded[i][0]:
+      return False
+  return True
+
+def extract_public_key(bytes):
+  decoded = [ x for x in script_GetOp(bytes) ]
+
+  # non-generated TxIn transactions push a signature
+  # (seventy-something bytes) and then their public key
+  # (65 bytes) onto the stack:
+  match = [ opcodes.OP_PUSHDATA4, opcodes.OP_PUSHDATA4 ]
+  if match_decoded(decoded, match):
+    return public_key_to_bc_address(decoded[1][1])
+
+  # The Genesis Block, self-payments, and pay-by-IP-address payments look like:
+  # 65 BYTES:... CHECKSIG
+  match = [ opcodes.OP_PUSHDATA4, opcodes.OP_CHECKSIG ]
+  if match_decoded(decoded, match):
+    return public_key_to_bc_address(decoded[0][1])
+
+  # Pay-by-Bitcoin-address TxOuts look like:
+  # DUP HASH160 20 BYTES:... EQUALVERIFY CHECKSIG
+  match = [ opcodes.OP_DUP, opcodes.OP_HASH160, opcodes.OP_PUSHDATA4, opcodes.OP_EQUALVERIFY, opcodes.OP_CHECKSIG ]
+  if match_decoded(decoded, match):
+    return hash_160_to_bc_address(decoded[2][1])
+
+  return "(None)"