rencode.py

"""
rencode -- Web safe object pickling/unpickling.

Public domain, Connelly Barnes 2006-2007.

The rencode module is a modified version of bencode from the
BitTorrent project.  For complex, heterogeneous data structures with
many small elements, r-encodings take up significantly less space than
b-encodings:

 >>> len(rencode.dumps({'a':0, 'b':[1,2], 'c':99}))
 13
 >>> len(bencode.bencode({'a':0, 'b':[1,2], 'c':99}))
 26

The rencode format is not standardized, and may change with different
rencode module versions, so you should check that you are using the
same rencode version throughout your project.
"""

__version__ = '1.0.2'
__all__ = ['dumps', 'loads']

# Original bencode module by Petru Paler, et al.
#
# Modifications by Connelly Barnes:
#
#  - Added support for floats (sent as 32-bit or 64-bit in network
#    order), bools, None.
#  - Allowed dict keys to be of any serializable type.
#  - Lists/tuples are always decoded as tuples (thus, tuples can be
#    used as dict keys).
#  - Embedded extra information in the 'typecodes' to save some space.
#  - Added a restriction on integer length, so that malicious hosts
#    cannot pass us large integers which take a long time to decode.
#
# Licensed by Bram Cohen under the "MIT license":
#
#  "Copyright (C) 2001-2002 Bram Cohen
#
#  Permission is hereby granted, free of charge, to any person
#  obtaining a copy of this software and associated documentation files
#  (the "Software"), to deal in the Software without restriction,
#  including without limitation the rights to use, copy, modify, merge,
#  publish, distribute, sublicense, and/or sell copies of the Software,
#  and to permit persons to whom the Software is furnished to do so,
#  subject to the following conditions:
#
#  The above copyright notice and this permission notice shall be
#  included in all copies or substantial portions of the Software.
#
#  The Software is provided "AS IS", without warranty of any kind,
#  express or implied, including but not limited to the warranties of
#  merchantability,  fitness for a particular purpose and
#  noninfringement. In no event shall the  authors or copyright holders
#  be liable for any claim, damages or other liability, whether in an
#  action of contract, tort or otherwise, arising from, out of or in
#  connection with the Software or the use or other dealings in the
#  Software."
#
# (The rencode module is licensed under the above license as well).
#

import sys

py3 = False
if sys.version_info.major >= 3:
    py3 = True
    long = int
    unicode = str

def int2byte(c):
    if py3:
        return bytes([c])
    else:
        return chr(c)

import struct
from threading import Lock

# Default number of bits for serialized floats, either 32 or 64 (also a parameter for dumps()).
DEFAULT_FLOAT_BITS = 32

# Maximum length of integer when written as base 10 string.
MAX_INT_LENGTH = 64

# The bencode 'typecodes' such as i, d, etc have been extended and
# relocated on the base-256 character set.
CHR_LIST    = int2byte(59)
CHR_DICT    = int2byte(60)
CHR_INT     = int2byte(61)
CHR_INT1    = int2byte(62)
CHR_INT2    = int2byte(63)
CHR_INT4    = int2byte(64)
CHR_INT8    = int2byte(65)
CHR_FLOAT32 = int2byte(66)
CHR_FLOAT64 = int2byte(44)
CHR_TRUE    = int2byte(67)
CHR_FALSE   = int2byte(68)
CHR_NONE    = int2byte(69)
CHR_TERM    = int2byte(127)

# Positive integers with value embedded in typecode.
INT_POS_FIXED_START = 0
INT_POS_FIXED_COUNT = 44

# Dictionaries with length embedded in typecode.
DICT_FIXED_START = 102
DICT_FIXED_COUNT = 25

# Negative integers with value embedded in typecode.
INT_NEG_FIXED_START = 70
INT_NEG_FIXED_COUNT = 32

# Strings with length embedded in typecode.
STR_FIXED_START = 128
STR_FIXED_COUNT = 64

# Lists with length embedded in typecode.
LIST_FIXED_START = STR_FIXED_START+STR_FIXED_COUNT
LIST_FIXED_COUNT = 64

# Whether strings should be decoded when loading
_decode_utf8 = False

def decode_int(x, f):
    f += 1
    newf = x.index(CHR_TERM, f)
    if newf - f >= MAX_INT_LENGTH:
        raise ValueError('overflow')
    try:
        n = int(x[f:newf])
    except (OverflowError, ValueError):
        n = long(x[f:newf])
    if x[f:f+1] == '-':
        if x[f + 1:f + 2] == '0':
            raise ValueError
    elif x[f:f+1] == '0' and newf != f+1:
        raise ValueError
    return (n, newf+1)

def decode_intb(x, f):
    f += 1
    return (struct.unpack('!b', x[f:f+1])[0], f+1)

def decode_inth(x, f):
    f += 1
    return (struct.unpack('!h', x[f:f+2])[0], f+2)

def decode_intl(x, f):
    f += 1
    x = x + "    "
    print "INTL: X: " + x
    print "INTL: LEN(X): " + str(len(x))
    print "INTL: F: " + str(f)
    return (struct.unpack('!l', x[f:f+4])[0], f+4)

def decode_intq(x, f):
    f += 1
    return (struct.unpack('!q', x[f:f+8])[0], f+8)

def decode_float32(x, f):
    f += 1
    n = struct.unpack('!f', x[f:f+4])[0]
    return (n, f+4)

def decode_float64(x, f):
    f += 1
    n = struct.unpack('!d', x[f:f+8])[0]
    return (n, f+8)

def decode_string(x, f):
    colon = x.index(b':', f)
    try:
        n = int(x[f:colon])
    except (OverflowError, ValueError):
        n = long(x[f:colon])
    if x[f] == '0' and colon != f+1:
        raise ValueError
    colon += 1
    s = x[colon:colon+n]
    if _decode_utf8:
        s = s.decode('utf8')
    return (s, colon+n)

def decode_list(x, f):
    r, f = [], f+1
    while x[f:f+1] != CHR_TERM:
        v, f = decode_func[x[f:f+1]](x, f)
        r.append(v)
    return (tuple(r), f + 1)

def decode_dict(x, f):
    r, f = {}, f+1
    while x[f:f+1] != CHR_TERM:
        k, f = decode_func[x[f:f+1]](x, f)
        print "K: " + k
        print "F: " + str(f)
        print "X: " + x
        print "R: "
        print r
        r[k], f = decode_func[x[f:f+1]](x, f)
    return (r, f + 1)

def decode_true(x, f):
  return (True, f+1)

def decode_false(x, f):
  return (False, f+1)

def decode_none(x, f):
  return (None, f+1)

decode_func = {}
decode_func[b'0'] = decode_string
decode_func[b'1'] = decode_string
decode_func[b'2'] = decode_string
decode_func[b'3'] = decode_string
decode_func[b'4'] = decode_string
decode_func[b'5'] = decode_string
decode_func[b'6'] = decode_string
decode_func[b'7'] = decode_string
decode_func[b'8'] = decode_string
decode_func[b'9'] = decode_string
decode_func[CHR_LIST   ] = decode_list
decode_func[CHR_DICT   ] = decode_dict
decode_func[CHR_INT    ] = decode_int
decode_func[CHR_INT1   ] = decode_intb
decode_func[CHR_INT2   ] = decode_inth
decode_func[CHR_INT4   ] = decode_intl
decode_func[CHR_INT8   ] = decode_intq
decode_func[CHR_FLOAT32] = decode_float32
decode_func[CHR_FLOAT64] = decode_float64
decode_func[CHR_TRUE   ] = decode_true
decode_func[CHR_FALSE  ] = decode_false
decode_func[CHR_NONE   ] = decode_none

def make_fixed_length_string_decoders():
    def make_decoder(slen):
        def f(x, f):
            s = x[f+1:f+1+slen]
            if _decode_utf8:
                s = s.decode("utf8")
            return (s, f+1+slen)
        return f
    for i in range(STR_FIXED_COUNT):
        decode_func[int2byte(STR_FIXED_START+i)] = make_decoder(i)

make_fixed_length_string_decoders()

def make_fixed_length_list_decoders():
    def make_decoder(slen):
        def f(x, f):
            r, f = [], f+1
            for i in range(slen):
                v, f = decode_func[x[f:f+1]](x, f)
                r.append(v)
            return (tuple(r), f)
        return f
    for i in range(LIST_FIXED_COUNT):
        decode_func[int2byte(LIST_FIXED_START+i)] = make_decoder(i)

make_fixed_length_list_decoders()

def make_fixed_length_int_decoders():
    def make_decoder(j):
        def f(x, f):
            return (j, f+1)
        return f
    for i in range(INT_POS_FIXED_COUNT):
        decode_func[int2byte(INT_POS_FIXED_START+i)] = make_decoder(i)
    for i in range(INT_NEG_FIXED_COUNT):
        decode_func[int2byte(INT_NEG_FIXED_START+i)] = make_decoder(-1-i)

make_fixed_length_int_decoders()

def make_fixed_length_dict_decoders():
    def make_decoder(slen):
        def f(x, f):
            r, f = {}, f+1
            for j in range(slen):
                k, f = decode_func[x[f:f+1]](x, f)
                r[k], f = decode_func[x[f:f+1]](x, f)
            return (r, f)
        return f
    for i in range(DICT_FIXED_COUNT):
        decode_func[int2byte(DICT_FIXED_START+i)] = make_decoder(i)

make_fixed_length_dict_decoders()

def loads(x, decode_utf8=True):
    global _decode_utf8
    _decode_utf8 = decode_utf8
    print x
    print x[0:1]
    try:
        r, l = decode_func[x[0:1]](x, 0)
    except (IndexError, KeyError):
        raise ValueError
    if l != len(x):
        raise ValueError
    return r

def encode_int(x, r):
    if 0 <= x < INT_POS_FIXED_COUNT:
        r.append(int2byte(INT_POS_FIXED_START+x))
    elif -INT_NEG_FIXED_COUNT <= x < 0:
        r.append(int2byte(INT_NEG_FIXED_START-1-x))
    elif -128 <= x < 128:
        r.extend((CHR_INT1, struct.pack('!b', x)))
    elif -32768 <= x < 32768:
        r.extend((CHR_INT2, struct.pack('!h', x)))
    elif -2147483648 <= x < 2147483648:
        r.extend((CHR_INT4, struct.pack('!l', x)))
    elif -9223372036854775808 <= x < 9223372036854775808:
        r.extend((CHR_INT8, struct.pack('!q', x)))
    else:
        s = str(x)
        if py3:
            s = bytes(s, "ascii")

        if len(s) >= MAX_INT_LENGTH:
            raise ValueError('overflow')
        r.extend((CHR_INT, s, CHR_TERM))

def encode_float32(x, r):
    r.extend((CHR_FLOAT32, struct.pack('!f', x)))

def encode_float64(x, r):
    r.extend((CHR_FLOAT64, struct.pack('!d', x)))

def encode_bool(x, r):
    r.append({False: CHR_FALSE, True: CHR_TRUE}[bool(x)])

def encode_none(x, r):
    r.append(CHR_NONE)

def encode_string(x, r):
    if len(x) < STR_FIXED_COUNT:
        r.extend((int2byte(STR_FIXED_START + len(x)), x))
    else:
        s = str(len(x))
        if py3:
            s = bytes(s, "ascii")
        r.extend((s, b':', x))

def encode_unicode(x, r):
    encode_string(x.encode("utf8"), r)

def encode_list(x, r):
    if len(x) < LIST_FIXED_COUNT:
        r.append(int2byte(LIST_FIXED_START + len(x)))
        for i in x:
            encode_func[type(i)](i, r)
    else:
        r.append(CHR_LIST)
        for i in x:
            encode_func[type(i)](i, r)
        r.append(CHR_TERM)

def encode_dict(x,r):
    if len(x) < DICT_FIXED_COUNT:
        r.append(int2byte(DICT_FIXED_START + len(x)))
        for k, v in x.items():
            encode_func[type(k)](k, r)
            encode_func[type(v)](v, r)
    else:
        r.append(CHR_DICT)
        for k, v in x.items():
            encode_func[type(k)](k, r)
            encode_func[type(v)](v, r)
        r.append(CHR_TERM)

encode_func = {}
encode_func[int] = encode_int
encode_func[long] = encode_int
encode_func[bytes] = encode_string
encode_func[list] = encode_list
encode_func[tuple] = encode_list
encode_func[dict] = encode_dict
encode_func[type(None)] = encode_none
encode_func[unicode] = encode_unicode
encode_func[bool] = encode_bool

lock = Lock()

def dumps(x, float_bits=DEFAULT_FLOAT_BITS):
    """
    Dump data structure to str.

    Here float_bits is either 32 or 64.
    """
    lock.acquire()
    try:
        if float_bits == 32:
            encode_func[float] = encode_float32
        elif float_bits == 64:
            encode_func[float] = encode_float64
        else:
            raise ValueError('Float bits (%d) is not 32 or 64' % float_bits)
        r = []
        encode_func[type(x)](x, r)
    finally:
        lock.release()

    return b''.join(r)

def test():
    f1 = struct.unpack('!f', struct.pack('!f', 25.5))[0]
    f2 = struct.unpack('!f', struct.pack('!f', 29.3))[0]
    f3 = struct.unpack('!f', struct.pack('!f', -0.6))[0]
    L = (({b'a':15, b'bb':f1, b'ccc':f2, b'':(f3,(),False,True,b'')},(b'a',10**20),tuple(range(-100000,100000)),b'b'*31,b'b'*62,b'b'*64,2**30,2**33,2**62,2**64,2**30,2**33,2**62,2**64,False,False, True, -1, 2, 0),)
    assert loads(dumps(L)) == L
    d = dict(zip(range(-100000,100000),range(-100000,100000)))
    d.update({b'a':20, 20:40, 40:41, f1:f2, f2:f3, f3:False, False:True, True:False})
    L = (d, {}, {5:6}, {7:7,True:8}, {9:10, 22:39, 49:50, 44: b''})
    assert loads(dumps(L)) == L
    L = (b'', b'a'*10, b'a'*100, b'a'*1000, b'a'*10000, b'a'*100000, b'a'*1000000, b'a'*10000000)
    assert loads(dumps(L)) == L
    L = tuple([dict(zip(range(n),range(n))) for n in range(100)]) + (b'b',)
    assert loads(dumps(L)) == L
    L = tuple([dict(zip(range(n),range(-n,0))) for n in range(100)]) + (b'b',)
    assert loads(dumps(L)) == L
    L = tuple([tuple(range(n)) for n in range(100)]) + (b'b',)
    assert loads(dumps(L)) == L
    L = tuple([b'a'*n for n in range(1000)]) + (b'b',)
    assert loads(dumps(L)) == L
    L = tuple([b'a'*n for n in range(1000)]) + (None,True,None)
    assert loads(dumps(L)) == L
    assert loads(dumps(None)) == None
    assert loads(dumps({None:None})) == {None:None}
    assert 1e-10<abs(loads(dumps(1.1))-1.1)<1e-6
    assert 1e-10<abs(loads(dumps(1.1,32))-1.1)<1e-6
    assert abs(loads(dumps(1.1,64))-1.1)<1e-12
    assert loads(dumps("Hello World!!"), decode_utf8=True)
try:
    import psyco
    psyco.bind(dumps)
    psyco.bind(loads)
except ImportError:
    pass


if __name__ == '__main__':
  test()