GitHub Viewer
"""Fallback pure Python implementation of msgpack"""
import struct
import sys
from datetime import datetime as _DateTime
if hasattr(sys, "pypy_version_info"):
from __pypy__ import newlist_hint
from __pypy__.builders import BytesBuilder
_USING_STRINGBUILDER = True
class BytesIO:
def __init__(self, s=b""):
if s:
self.builder = BytesBuilder(len(s))
self.builder.append(s)
else:
self.builder = BytesBuilder()
def write(self, s):
if isinstance(s, memoryview):
s = s.tobytes()
elif isinstance(s, bytearray):
s = bytes(s)
self.builder.append(s)
def getvalue(self):
return self.builder.build()
else:
from io import BytesIO
_USING_STRINGBUILDER = False
def newlist_hint(size):
return []
from .exceptions import BufferFull, ExtraData, FormatError, OutOfData, StackError
from .ext import ExtType, Timestamp
EX_SKIP = 0
EX_CONSTRUCT = 1
EX_READ_ARRAY_HEADER = 2
EX_READ_MAP_HEADER = 3
TYPE_IMMEDIATE = 0
TYPE_ARRAY = 1
TYPE_MAP = 2
TYPE_RAW = 3
TYPE_BIN = 4
TYPE_EXT = 5
DEFAULT_RECURSE_LIMIT = 1024
def _check_type_strict(obj, t, type=type, tuple=tuple):
if type(t) is tuple:
return type(obj) in t
else:
return type(obj) is t
def _get_data_from_buffer(obj):
view = memoryview(obj)
if view.itemsize != 1:
raise ValueError("cannot unpack from multi-byte object")
return view
def unpackb(packed, **kwargs):
"""
Unpack an object from `packed`.
Raises ``ExtraData`` when *packed* contains extra bytes.
Raises ``ValueError`` when *packed* is incomplete.
Raises ``FormatError`` when *packed* is not valid msgpack.
Raises ``StackError`` when *packed* contains too nested.
Other exceptions can be raised during unpacking.
See :class:`Unpacker` for options.
"""
unpacker = Unpacker(None, max_buffer_size=len(packed), **kwargs)
unpacker.feed(packed)
try:
ret = unpacker._unpack()
except OutOfData:
raise ValueError("Unpack failed: incomplete input")
except RecursionError:
raise StackError
if unpacker._got_extradata():
raise ExtraData(ret, unpacker._get_extradata())
return ret
_NO_FORMAT_USED = ""
_MSGPACK_HEADERS = {
0xC4: (1, _NO_FORMAT_USED, TYPE_BIN),
0xC5: (2, ">H", TYPE_BIN),
0xC6: (4, ">I", TYPE_BIN),
0xC7: (2, "Bb", TYPE_EXT),
0xC8: (3, ">Hb", TYPE_EXT),
0xC9: (5, ">Ib", TYPE_EXT),
0xCA: (4, ">f"),
0xCB: (8, ">d"),
0xCC: (1, _NO_FORMAT_USED),
0xCD: (2, ">H"),
0xCE: (4, ">I"),
0xCF: (8, ">Q"),
0xD0: (1, "b"),
0xD1: (2, ">h"),
0xD2: (4, ">i"),
0xD3: (8, ">q"),
0xD4: (1, "b1s", TYPE_EXT),
0xD5: (2, "b2s", TYPE_EXT),
0xD6: (4, "b4s", TYPE_EXT),
0xD7: (8, "b8s", TYPE_EXT),
0xD8: (16, "b16s", TYPE_EXT),
0xD9: (1, _NO_FORMAT_USED, TYPE_RAW),
0xDA: (2, ">H", TYPE_RAW),
0xDB: (4, ">I", TYPE_RAW),
0xDC: (2, ">H", TYPE_ARRAY),
0xDD: (4, ">I", TYPE_ARRAY),
0xDE: (2, ">H", TYPE_MAP),
0xDF: (4, ">I", TYPE_MAP),
}
class Unpacker:
"""Streaming unpacker.
Arguments:
:param file_like:
File-like object having `.read(n)` method.
If specified, unpacker reads serialized data from it and `.feed()` is not usable.
:param int read_size:
Used as `file_like.read(read_size)`. Must be equal to or smaller than *max_buffer_size*.
:param bool use_list:
If true, unpack msgpack array to Python list.
Otherwise, unpack to Python tuple. (default: True)
:param bool raw:
If true, unpack msgpack raw to Python bytes.
Otherwise, unpack to Python str by decoding with UTF-8 encoding (default).
:param int timestamp:
Control how timestamp type is unpacked:
0 - Timestamp
1 - float (Seconds from the EPOCH)
2 - int (Nanoseconds from the EPOCH)
3 - datetime.datetime (UTC).
:param bool strict_map_key:
If true (default), only str or bytes are accepted for map (dict) keys.
:param object_hook:
When specified, it should be callable.
Unpacker calls it with a dict argument after unpacking msgpack map.
(See also simplejson)
:param object_pairs_hook:
When specified, it should be callable.
Unpacker calls it with a list of key-value pairs after unpacking msgpack map.
(See also simplejson)
:param str unicode_errors:
The error handler for decoding unicode. (default: 'strict')
This option should be used only when you have msgpack data which
contains invalid UTF-8 string.
:param int max_buffer_size:
Limits size of data waiting unpacked. 0 means 2**32-1.
The default value is 100*1024*1024 (100MiB).
Raises `BufferFull` exception when it is insufficient.
You should set this parameter when unpacking data from untrusted source.
:param int max_str_len:
Deprecated, use *max_buffer_size* instead.
Limits max length of str. (default: max_buffer_size)
:param int max_bin_len:
Deprecated, use *max_buffer_size* instead.
Limits max length of bin. (default: max_buffer_size)
:param int max_array_len:
Limits max length of array.
(default: max_buffer_size)
:param int max_map_len:
Limits max length of map.
(default: max_buffer_size//2)
:param int max_ext_len:
Deprecated, use *max_buffer_size* instead.
Limits max size of ext type. (default: max_buffer_size)
Example of streaming deserialize from file-like object::
unpacker = Unpacker(file_like)
for o in unpacker:
process(o)
Example of streaming deserialize from socket::
unpacker = Unpacker()
while True:
buf = sock.recv(1024**2)
if not buf:
break
unpacker.feed(buf)
for o in unpacker:
process(o)
Raises ``ExtraData`` when *packed* contains extra bytes.
Raises ``OutOfData`` when *packed* is incomplete.
Raises ``FormatError`` when *packed* is not valid msgpack.
Raises ``StackError`` when *packed* contains too nested.
Other exceptions can be raised during unpacking.
"""
def __init__(
self,
file_like=None,
*,
read_size=0,
use_list=True,
raw=False,
timestamp=0,
strict_map_key=True,
object_hook=None,
object_pairs_hook=None,
list_hook=None,
unicode_errors=None,
max_buffer_size=100 * 1024 * 1024,
ext_hook=ExtType,
max_str_len=-1,
max_bin_len=-1,
max_array_len=-1,
max_map_len=-1,
max_ext_len=-1,
):
if unicode_errors is None:
unicode_errors = "strict"
if file_like is None:
self._feeding = True
else:
if not callable(file_like.read):
raise TypeError("`file_like.read` must be callable")
self.file_like = file_like
self._feeding = False
#: array of bytes fed.
self._buffer = bytearray()
#: Which position we currently reads
self._buff_i = 0
# When Unpacker is used as an iterable, between the calls to next(),
# the buffer is not "consumed" completely, for efficiency sake.
# Instead, it is done sloppily. To make sure we raise BufferFull at
# the correct moments, we have to keep track of how sloppy we were.
# Furthermore, when the buffer is incomplete (that is: in the case
# we raise an OutOfData) we need to rollback the buffer to the correct
# state, which _buf_checkpoint records.
self._buf_checkpoint = 0
if not max_buffer_size:
max_buffer_size = 2**31 - 1
if max_str_len == -1:
max_str_len = max_buffer_size
if max_bin_len == -1:
max_bin_len = max_buffer_size
if max_array_len == -1:
max_array_len = max_buffer_size
if max_map_len == -1:
max_map_len = max_buffer_size // 2
if max_ext_len == -1:
max_ext_len = max_buffer_size
self._max_buffer_size = max_buffer_size
if read_size > self._max_buffer_size:
raise ValueError("read_size must be smaller than max_buffer_size")
self._read_size = read_size or min(self._max_buffer_size, 16 * 1024)
self._raw = bool(raw)
self._strict_map_key = bool(strict_map_key)
self._unicode_errors = unicode_errors
self._use_list = use_list
if not (0 0:
del self._buffer[: self._buf_checkpoint]
self._buff_i -= self._buf_checkpoint
self._buf_checkpoint = 0
# Use extend here: INPLACE_ADD += doesn't reliably typecast memoryview in jython
self._buffer.extend(view if view.contiguous else view.tobytes())
view.release()
def _consume(self):
"""Gets rid of the used parts of the buffer."""
self._stream_offset += self._buff_i - self._buf_checkpoint
self._buf_checkpoint = self._buff_i
def _got_extradata(self):
return self._buff_i < len(self._buffer)
def _get_extradata(self):
return self._buffer[self._buff_i :]
def read_bytes(self, n):
ret = self._read(n, raise_outofdata=False)
self._consume()
return ret
def _read(self, n, raise_outofdata=True):
# (int) -> bytearray
self._reserve(n, raise_outofdata=raise_outofdata)
i = self._buff_i
ret = self._buffer[i : i + n]
self._buff_i = i + len(ret)
return ret
def _reserve(self, n, raise_outofdata=True):
remain_bytes = len(self._buffer) - self._buff_i - n
# Fast path: buffer has n bytes already
if remain_bytes >= 0:
return
if self._feeding:
self._buff_i = self._buf_checkpoint
raise OutOfData
# Strip buffer before checkpoint before reading file.
if self._buf_checkpoint > 0:
del self._buffer[: self._buf_checkpoint]
self._buff_i -= self._buf_checkpoint
self._buf_checkpoint = 0
# Read from file
remain_bytes = -remain_bytes
if remain_bytes + len(self._buffer) > self._max_buffer_size:
raise BufferFull
while remain_bytes > 0:
to_read_bytes = max(self._read_size, remain_bytes)
read_data = self.file_like.read(to_read_bytes)
if not read_data:
break
assert isinstance(read_data, bytes)
self._buffer += read_data
remain_bytes -= len(read_data)
if len(self._buffer) < n + self._buff_i and raise_outofdata:
self._buff_i = 0 # rollback
raise OutOfData
def _read_header(self):
typ = TYPE_IMMEDIATE
n = 0
obj = None
self._reserve(1)
b = self._buffer[self._buff_i]
self._buff_i += 1
if b & 0b10000000 == 0:
obj = b
elif b & 0b11100000 == 0b11100000:
obj = -1 - (b ^ 0xFF)
elif b & 0b11100000 == 0b10100000:
n = b & 0b00011111
typ = TYPE_RAW
if n > self._max_str_len:
raise ValueError(f"{n} exceeds max_str_len({self._max_str_len})")
obj = self._read(n)
elif b & 0b11110000 == 0b10010000:
n = b & 0b00001111
typ = TYPE_ARRAY
if n > self._max_array_len:
raise ValueError(f"{n} exceeds max_array_len({self._max_array_len})")
elif b & 0b11110000 == 0b10000000:
n = b & 0b00001111
typ = TYPE_MAP
if n > self._max_map_len:
raise ValueError(f"{n} exceeds max_map_len({self._max_map_len})")
elif b == 0xC0:
obj = None
elif b == 0xC2:
obj = False
elif b == 0xC3:
obj = True
elif 0xC4 self._max_bin_len:
raise ValueError(f"{n} exceeds max_bin_len({self._max_bin_len})")
obj = self._read(n)
elif 0xC7