pickle和json模块

json模块

json模块是实现序列化和反序列化的，主要用户不同程序之间的数据交换，首先来看一下:

dumps()序列化

import json
'''json模块是实现序列化和反序列话功能的'''
users = ["alex","tom","wupeiqi","sb","耿长学"]
mes = json.dumps(users)    #实例化，并打印
print(mes)
运行结果如下：
["alex", "tom", "wupeiqi", "sb", "\u803f\u957f\u5b66"]

从上面可以看出，dumps其实是生成一个序列化的实例，这个后面会和dump进行区分，而且汉字非英文转化成的是字节码。

loads()反序列化

import json
'''json模块是实现序列化和反序列话功能的'''
users = ["alex","tom","wupeiqi","sb","耿长学"]
mes = json.dumps(users)    #实例化，并打印
print("序列化:",mes)
 
data = json.loads(mes)
print("反序列化：",data)
运行结果如下：
序列化: ["alex", "tom", "wupeiqi", "sb", "\u803f\u957f\u5b66"]
反序列化： ['alex', 'tom', 'wupeiqi', 'sb', '耿长学']

上面dumps()和loads()序列化和反序列化是在程序之间交互，即没有通过文件，只是以json.dumps()序列化一个实例，然后接收之后json.loads()进行反序列化，依次实现程序之间的交互。那么如何在文件中交互呢？

文件中的交互：

import json
'''json模块是实现序列化和反序列话功能的'''
users = ["alex","tom","wupeiqi","sb","耿长学"]
mes = json.dumps(users)    #实例化，并打印
'''把mes序列化实例写入文件'''
with open("users",'w+') as fw:
    fw.write(mes)

上面代码中，我们把序列化的实例写入到文件中，使用json.dumps()生成实例，fw.write()写入文件，需要通过文件的write()写入。

文件反序列化：

import json
with open('users','r+') as fr:
    mess = json.loads(fr.read())
    print(mess)
运行结果如下：
['alex', 'tom', 'wupeiqi', 'sb', '耿长学']

上面文件反序列化过程中，首先要把文件中的内容读取出来，因为是使用write()写进去的，因此要read()出来，然后再进行反序列化，生成一个实例。

下面我们来看看load()和dump()实现序列化和反序列化：

    dump()序列化

import json
users = {"alex":"sb","wupeiqi":,"耿长学":}
with open("users","w+") as fw:
    json.dump(users,fw)

从上面可以看出，json.dump()是直接把实例化的信息序列化到指定文件中，不需要write()来写入，直接可以自己来写入，而dumps()只是序列化成一个实例，还需要自己write()到文件中。

    load()反序列化

import json
with open('users','r+') as fr:
    users = json.load(fr)
    print(users)
 
运行结果如下：
{'alex': 'sb', '耿长学': , 'wupeiqi': }

load()反序列化不需要读取之后才反序列化，直接可以从文件反序列化，因为是dump()进去的。

    总结：

从上面dumps()、loads()和dump()、load()序列化和反序列化可以看出，dumps()是序列化生成一个实例，loads()是读取写进去的实例，dumps()和loads()主要用于不同程序或者接口之间的数据交换传输，而dump()和load()更适合于文件级别之间的读取和写入，两者之间还是有侧重点的。

这个不是说两者功能一样，用法不同，其实设计的时候侧重方向就是不一样的。

    json源代码

r"""JSON (JavaScript Object Notation) <http://json.org> is a subset of
JavaScript syntax (ECMA- 3rd edition) used as a lightweight data
interchange format.
 
:mod:`json` exposes an API familiar to users of the standard library
:mod:`marshal` and :mod:`pickle` modules.  It is derived from a
version of the externally maintained simplejson library.
 
Encoding basic Python object hierarchies::
 
    >>> import json
    >>> json.dumps(['foo', {'bar': ('baz', None, 1.0, )}])
    '["foo", {"bar": ["baz", null, 1.0, 2]}]'
    >>> print(json.dumps("\"foo\bar"))
    "\"foo\bar"
    >>> print(json.dumps('\u1234'))
    "\u1234"
    >>> print(json.dumps('\\'))
    "\\"
    >>> print(json.dumps({"c": , "b": , "a": }, sort_keys=True))
    {"a": , "b": , "c": }
    >>> from io import StringIO
    >>> io = StringIO()
    >>> json.dump(['streaming API'], io)
    >>> io.getvalue()
    '["streaming API"]'
 
Compact encoding::
 
    >>> import json
    >>> from collections import OrderedDict
    >>> mydict = OrderedDict([('', ), ('', )])
    >>> json.dumps([,,,mydict], separators=(',', ':'))
    '[1,2,3,{"4":5,"6":7}]'
 
Pretty printing::
 
    >>> import json
    >>> print(json.dumps({'': , '': }, sort_keys=True, indent=))
    {
        "": ,
        "":
    }
 
Decoding JSON::
 
    >>> import json
    >>> obj = ['foo', {'bar': ['baz', None, 1.0, ]}]
    >>> json.loads('["foo", {"bar":["baz", null, 1.0, 2]}]') == obj
    True
    >>> json.loads('"\\"foo\\bar"') == '"foo\x08ar'
    True
    >>> from io import StringIO
    >>> io = StringIO('["streaming API"]')
    >>> json.load(io)[] == 'streaming API'
    True
 
Specializing JSON object decoding::
 
    >>> import json
    >>> def as_complex(dct):
    ...     if '__complex__' in dct:
    ...         return complex(dct['real'], dct['imag'])
    ...     return dct
    ...
    >>> json.loads('{"__complex__": true, "real": 1, "imag": 2}',
    ...     object_hook=as_complex)
    (+2j)
    >>> from decimal import Decimal
    >>> json.loads('1.1', parse_float=Decimal) == Decimal('1.1')
    True
 
Specializing JSON object encoding::
 
    >>> import json
    >>> def encode_complex(obj):
    ...     if isinstance(obj, complex):
    ...         return [obj.real, obj.imag]
    ...     raise TypeError(repr(o) + " is not JSON serializable")
    ...
    >>> json.dumps( + 1j, default=encode_complex)
    '[2.0, 1.0]'
    >>> json.JSONEncoder(default=encode_complex).encode( + 1j)
    '[2.0, 1.0]'
    >>> ''.join(json.JSONEncoder(default=encode_complex).iterencode( + 1j))
    '[2.0, 1.0]'
 
Using json.tool from the shell to validate and pretty-print::
 
    $ echo '{"json":"obj"}' | python -m json.tool
    {
        "json": "obj"
    }
    $ echo '{ 1.2:3.4}' | python -m json.tool
    Expecting property name enclosed in double quotes: line  column  (char )
"""
__version__ = '2.0.9'
__all__ = [
    'dump', 'dumps', 'load', 'loads',
    'JSONDecoder', 'JSONDecodeError', 'JSONEncoder',
]
 
__author__ = 'Bob Ippolito <bob@redivi.com>'
 
from .decoder import JSONDecoder, JSONDecodeError
from .encoder import JSONEncoder
 
_default_encoder = JSONEncoder(
    skipkeys=False,
    ensure_ascii=True,
    check_circular=True,
    allow_nan=True,
    indent=None,
    separators=None,
    default=None,
)
 
def dump(obj, fp, skipkeys=False, ensure_ascii=True, check_circular=True,
        allow_nan=True, cls=None, indent=None, separators=None,
        default=None, sort_keys=False, **kw):
    """Serialize ``obj`` as a JSON formatted stream to ``fp`` (a
    ``.write()``-supporting file-like object).
 
    If ``skipkeys`` is true then ``dict`` keys that are not basic types
    (``str``, ``int``, ``float``, ``bool``, ``None``) will be skipped
    instead of raising a ``TypeError``.
 
    If ``ensure_ascii`` is false, then the strings written to ``fp`` can
    contain non-ASCII characters if they appear in strings contained in
    ``obj``. Otherwise, all such characters are escaped in JSON strings.
 
    If ``check_circular`` is false, then the circular reference check
    for container types will be skipped and a circular reference will
    result in an ``OverflowError`` (or worse).
 
    If ``allow_nan`` is false, then it will be a ``ValueError`` to
    serialize out of range ``float`` values (``nan``, ``inf``, ``-inf``)
    in strict compliance of the JSON specification, instead of using the
    JavaScript equivalents (``NaN``, ``Infinity``, ``-Infinity``).
 
    If ``indent`` is a non-negative integer, then JSON array elements and
    object members will be pretty-printed with that indent level. An indent
    level of  will only insert newlines. ``None`` is the most compact
    representation.
 
    If specified, ``separators`` should be an ``(item_separator, key_separator)``
    tuple.  The default is ``(', ', ': ')`` if *indent* is ``None`` and
    ``(',', ': ')`` otherwise.  To get the most compact JSON representation,
    you should specify ``(',', ':')`` to eliminate whitespace.
 
    ``default(obj)`` is a function that should return a serializable version
    of obj or raise TypeError. The default simply raises TypeError.
 
    If *sort_keys* is ``True`` (default: ``False``), then the output of
    dictionaries will be sorted by key.
 
    To use a custom ``JSONEncoder`` subclass (e.g. one that overrides the
    ``.default()`` method to serialize additional types), specify it with
    the ``cls`` kwarg; otherwise ``JSONEncoder`` is used.
 
    """
    # cached encoder
    if (not skipkeys and ensure_ascii and
        check_circular and allow_nan and
        cls is None and indent is None and separators is None and
        default is None and not sort_keys and not kw):
        iterable = _default_encoder.iterencode(obj)
    else:
        if cls is None:
            cls = JSONEncoder
        iterable = cls(skipkeys=skipkeys, ensure_ascii=ensure_ascii,
            check_circular=check_circular, allow_nan=allow_nan, indent=indent,
            separators=separators,
            default=default, sort_keys=sort_keys, **kw).iterencode(obj)
    # could accelerate with writelines in some versions of Python, at
    # a debuggability cost
    for chunk in iterable:
        fp.write(chunk)
 
def dumps(obj, skipkeys=False, ensure_ascii=True, check_circular=True,
        allow_nan=True, cls=None, indent=None, separators=None,
        default=None, sort_keys=False, **kw):
    """Serialize ``obj`` to a JSON formatted ``str``.
 
    If ``skipkeys`` is true then ``dict`` keys that are not basic types
    (``str``, ``int``, ``float``, ``bool``, ``None``) will be skipped
    instead of raising a ``TypeError``.
 
    If ``ensure_ascii`` is false, then the return value can contain non-ASCII
    characters if they appear in strings contained in ``obj``. Otherwise, all
    such characters are escaped in JSON strings.
 
    If ``check_circular`` is false, then the circular reference check
    for container types will be skipped and a circular reference will
    result in an ``OverflowError`` (or worse).
 
    If ``allow_nan`` is false, then it will be a ``ValueError`` to
    serialize out of range ``float`` values (``nan``, ``inf``, ``-inf``) in
    strict compliance of the JSON specification, instead of using the
    JavaScript equivalents (``NaN``, ``Infinity``, ``-Infinity``).
 
    If ``indent`` is a non-negative integer, then JSON array elements and
    object members will be pretty-printed with that indent level. An indent
    level of  will only insert newlines. ``None`` is the most compact
    representation.
 
    If specified, ``separators`` should be an ``(item_separator, key_separator)``
    tuple.  The default is ``(', ', ': ')`` if *indent* is ``None`` and
    ``(',', ': ')`` otherwise.  To get the most compact JSON representation,
    you should specify ``(',', ':')`` to eliminate whitespace.
 
    ``default(obj)`` is a function that should return a serializable version
    of obj or raise TypeError. The default simply raises TypeError.
 
    If *sort_keys* is ``True`` (default: ``False``), then the output of
    dictionaries will be sorted by key.
 
    To use a custom ``JSONEncoder`` subclass (e.g. one that overrides the
    ``.default()`` method to serialize additional types), specify it with
    the ``cls`` kwarg; otherwise ``JSONEncoder`` is used.
 
    """
    # cached encoder
    if (not skipkeys and ensure_ascii and
        check_circular and allow_nan and
        cls is None and indent is None and separators is None and
        default is None and not sort_keys and not kw):
        return _default_encoder.encode(obj)
    if cls is None:
        cls = JSONEncoder
    return cls(
        skipkeys=skipkeys, ensure_ascii=ensure_ascii,
        check_circular=check_circular, allow_nan=allow_nan, indent=indent,
        separators=separators, default=default, sort_keys=sort_keys,
        **kw).encode(obj)
 
_default_decoder = JSONDecoder(object_hook=None, object_pairs_hook=None)
 
def load(fp, cls=None, object_hook=None, parse_float=None,
        parse_int=None, parse_constant=None, object_pairs_hook=None, **kw):
    """Deserialize ``fp`` (a ``.read()``-supporting file-like object containing
    a JSON document) to a Python object.
 
    ``object_hook`` is an optional function that will be called with the
    result of any object literal decode (a ``dict``). The return value of
    ``object_hook`` will be used instead of the ``dict``. This feature
    can be used to implement custom decoders (e.g. JSON-RPC class hinting).
 
    ``object_pairs_hook`` is an optional function that will be called with the
    result of any object literal decoded with an ordered list of pairs.  The
    return value of ``object_pairs_hook`` will be used instead of the ``dict``.
    This feature can be used to implement custom decoders that rely on the
    order that the key and value pairs are decoded (for example,
    collections.OrderedDict will remember the order of insertion). If
    ``object_hook`` is also defined, the ``object_pairs_hook`` takes priority.
 
    To use a custom ``JSONDecoder`` subclass, specify it with the ``cls``
    kwarg; otherwise ``JSONDecoder`` is used.
 
    """
    return loads(fp.read(),
        cls=cls, object_hook=object_hook,
        parse_float=parse_float, parse_int=parse_int,
        parse_constant=parse_constant, object_pairs_hook=object_pairs_hook, **kw)
 
def loads(s, encoding=None, cls=None, object_hook=None, parse_float=None,
        parse_int=None, parse_constant=None, object_pairs_hook=None, **kw):
    """Deserialize ``s`` (a ``str`` instance containing a JSON
    document) to a Python object.
 
    ``object_hook`` is an optional function that will be called with the
    result of any object literal decode (a ``dict``). The return value of
    ``object_hook`` will be used instead of the ``dict``. This feature
    can be used to implement custom decoders (e.g. JSON-RPC class hinting).
 
    ``object_pairs_hook`` is an optional function that will be called with the
    result of any object literal decoded with an ordered list of pairs.  The
    return value of ``object_pairs_hook`` will be used instead of the ``dict``.
    This feature can be used to implement custom decoders that rely on the
    order that the key and value pairs are decoded (for example,
    collections.OrderedDict will remember the order of insertion). If
    ``object_hook`` is also defined, the ``object_pairs_hook`` takes priority.
 
    ``parse_float``, if specified, will be called with the string
    of every JSON float to be decoded. By default this is equivalent to
    float(num_str). This can be used to use another datatype or parser
    for JSON floats (e.g. decimal.Decimal).
 
    ``parse_int``, if specified, will be called with the string
    of every JSON int to be decoded. By default this is equivalent to
    int(num_str). This can be used to use another datatype or parser
    for JSON integers (e.g. float).
 
    ``parse_constant``, if specified, will be called with one of the
    following strings: -Infinity, Infinity, NaN, null, true, false.
    This can be used to raise an exception if invalid JSON numbers
    are encountered.
 
    To use a custom ``JSONDecoder`` subclass, specify it with the ``cls``
    kwarg; otherwise ``JSONDecoder`` is used.
 
    The ``encoding`` argument is ignored and deprecated.
 
    """
    if not isinstance(s, str):
        raise TypeError('the JSON object must be str, not {!r}'.format(
                            s.__class__.__name__))
    if s.startswith(u'\ufeff'):
        raise JSONDecodeError("Unexpected UTF-8 BOM (decode using utf-8-sig)",
                              s, )
    if (cls is None and object_hook is None and
            parse_int is None and parse_float is None and
            parse_constant is None and object_pairs_hook is None and not kw):
        return _default_decoder.decode(s)
    if cls is None:
        cls = JSONDecoder
    if object_hook is not None:
        kw['object_hook'] = object_hook
    if object_pairs_hook is not None:
        kw['object_pairs_hook'] = object_pairs_hook
    if parse_float is not None:
        kw['parse_float'] = parse_float
    if parse_int is not None:
        kw['parse_int'] = parse_int
    if parse_constant is not None:
        kw['parse_constant'] = parse_constant
    return cls(**kw).decode(s)

    pickle源代码

"""Create portable serialized representations of Python objects.
 
See module copyreg for a mechanism for registering custom picklers.
See module pickletools source for extensive comments.
 
Classes:
 
    Pickler
    Unpickler
 
Functions:
 
    dump(object, file)
    dumps(object) -> string
    load(file) -> object
    loads(string) -> object
 
Misc variables:
 
    __version__
    format_version
    compatible_formats
 
"""
 
from types import FunctionType
from copyreg import dispatch_table
from copyreg import _extension_registry, _inverted_registry, _extension_cache
from itertools import islice
import sys
from sys import maxsize
from struct import pack, unpack
import re
import io
import codecs
import _compat_pickle
 
__all__ = ["PickleError", "PicklingError", "UnpicklingError", "Pickler",
           "Unpickler", "dump", "dumps", "load", "loads"]
 
# Shortcut for use in isinstance testing
bytes_types = (bytes, bytearray)
 
# These are purely informational; no code uses these.
format_version = "4.0"                  # File format version we write
compatible_formats = ["1.0",            # Original protocol
                      "1.1",            # Protocol  with INST added
                      "1.2",            # Original protocol
                      "1.3",            # Protocol  with BINFLOAT added
                      "2.0",            # Protocol
                      "3.0",            # Protocol
                      "4.0",            # Protocol
                      ]                 # Old format versions we can read
 
# This is the highest protocol number we know how to read.
HIGHEST_PROTOCOL = 
 
# The protocol we write by default.  May be less than HIGHEST_PROTOCOL.
# We intentionally write a protocol that Python .x cannot read;
# there are too many issues with that.
DEFAULT_PROTOCOL = 
 
class PickleError(Exception):
    """A common base class for the other pickling exceptions."""
    pass
 
class PicklingError(PickleError):
    """This exception is raised when an unpicklable object is passed to the
    dump() method.
 
    """
    pass
 
class UnpicklingError(PickleError):
    """This exception is raised when there is a problem unpickling an object,
    such as a security violation.
 
    Note that other exceptions may also be raised during unpickling, including
    (but not necessarily limited to) AttributeError, EOFError, ImportError,
    and IndexError.
 
    """
    pass
 
# An instance of _Stop is raised by Unpickler.load_stop() in response to
# the STOP opcode, passing the object that is the result of unpickling.
class _Stop(Exception):
    def __init__(self, value):
        self.value = value
 
# Jython has PyStringMap; it's a dict subclass with string keys
try:
    from org.python.core import PyStringMap
except ImportError:
    PyStringMap = None
 
# Pickle opcodes.  See pickletools.py for extensive docs.  The listing
# here is in kind-of alphabetical order of -character pickle code.
# pickletools groups them by purpose.
 
MARK           = b'('   # push special markobject on stack
STOP           = b'.'   # every pickle ends with STOP
POP            = b''   # discard topmost stack item
POP_MARK       = b''   # discard stack top through topmost markobject
DUP            = b''   # duplicate top stack item
FLOAT          = b'F'   # push float object; decimal string argument
INT            = b'I'   # push integer or bool; decimal string argument
BININT         = b'J'   # push four-byte signed int
BININT1        = b'K'   # push -byte unsigned int
LONG           = b'L'   # push long; decimal string argument
BININT2        = b'M'   # push -byte unsigned int
NONE           = b'N'   # push None
PERSID         = b'P'   # push persistent object; id is taken from string arg
BINPERSID      = b'Q'   #  "       "         "  ;  "  "   "     "  stack
REDUCE         = b'R'   # apply callable to argtuple, both on stack
STRING         = b'S'   # push string; NL-terminated string argument
BINSTRING      = b'T'   # push string; counted binary string argument
SHORT_BINSTRING= b'U'   #  "     "   ;    "      "       "      " <  bytes
UNICODE        = b'V'   # push Unicode string; raw-unicode-escaped'd argument
BINUNICODE     = b'X'   #   "     "       "  ; counted UTF-8 string argument
APPEND         = b'a'   # append stack top to list below it
BUILD          = b'b'   # call __setstate__ or __dict__.update()
GLOBAL         = b'c'   # push self.find_class(modname, name);  string args
DICT           = b'd'   # build a dict from stack items
EMPTY_DICT     = b'}'   # push empty dict
APPENDS        = b'e'   # extend list on stack by topmost stack slice
GET            = b'g'   # push item from memo on stack; index is string arg
BINGET         = b'h'   #   "    "    "    "   "   "  ;   "    " -byte arg
INST           = b'i'   # build & push class instance
LONG_BINGET    = b'j'   # push item from memo on stack; index is -byte arg
LIST           = b'l'   # build list from topmost stack items
EMPTY_LIST     = b']'   # push empty list
OBJ            = b'o'   # build & push class instance
PUT            = b'p'   # store stack top in memo; index is string arg
BINPUT         = b'q'   #   "     "    "   "   " ;   "    " 1-byte arg
LONG_BINPUT    = b'r'   #   "     "    "   "   " ;   "    " 4-byte arg
SETITEM        = b's'   # add key+value pair to dict
TUPLE          = b't'   # build tuple from topmost stack items
EMPTY_TUPLE    = b')'   # push empty tuple
SETITEMS       = b'u'   # modify dict by adding topmost key+value pairs
BINFLOAT       = b'G'   # push float; arg is -byte float encoding
 
TRUE           = b'I01\n'  # not an opcode; see INT docs in pickletools.py
FALSE          = b'I00\n'  # not an opcode; see INT docs in pickletools.py
 
# Protocol 
 
PROTO          = b'\x80'  # identify pickle protocol
NEWOBJ         = b'\x81'  # build object by applying cls.__new__ to argtuple
EXT1           = b'\x82'  # push object from extension registry; -byte index
EXT2           = b'\x83'  # ditto, but -byte index
EXT4           = b'\x84'  # ditto, but -byte index
TUPLE1         = b'\x85'  # build -tuple from stack top
TUPLE2         = b'\x86'  # build -tuple from two topmost stack items
TUPLE3         = b'\x87'  # build -tuple from three topmost stack items
NEWTRUE        = b'\x88'  # push True
NEWFALSE       = b'\x89'  # push False
LONG1          = b'\x8a'  # push long from <  bytes
LONG4          = b'\x8b'  # push really big long
 
_tuplesize2code = [EMPTY_TUPLE, TUPLE1, TUPLE2, TUPLE3]
 
# Protocol  (Python .x)
 
BINBYTES       = b'B'   # push bytes; counted binary string argument
SHORT_BINBYTES = b'C'   #  "     "   ;    "      "       "      " <  bytes
 
# Protocol
SHORT_BINUNICODE = b'\x8c'  # push short string; UTF- length <  bytes
BINUNICODE8      = b'\x8d'  # push very long string
BINBYTES8        = b'\x8e'  # push very long bytes string
EMPTY_SET        = b'\x8f'  # push empty set on the stack
ADDITEMS         = b'\x90'  # modify set by adding topmost stack items
FROZENSET        = b'\x91'  # build frozenset from topmost stack items
NEWOBJ_EX        = b'\x92'  # like NEWOBJ but work with keyword only arguments
STACK_GLOBAL     = b'\x93'  # same as GLOBAL but using names on the stacks
MEMOIZE          = b'\x94'  # store top of the stack in memo
FRAME            = b'\x95'  # indicate the beginning of a new frame
 
__all__.extend([x for x in dir() if re.match("[A-Z][A-Z0-9_]+$", x)])
 
class _Framer:
 
    _FRAME_SIZE_TARGET =  * 
 
    def __init__(self, file_write):
        self.file_write = file_write
        self.current_frame = None
 
    def start_framing(self):
        self.current_frame = io.BytesIO()
 
    def end_framing(self):
        if self.current_frame and self.current_frame.tell() > :
            self.commit_frame(force=True)
            self.current_frame = None
 
    def commit_frame(self, force=False):
        if self.current_frame:
            f = self.current_frame
            if f.tell() >= self._FRAME_SIZE_TARGET or force:
                with f.getbuffer() as data:
                    n = len(data)
                    write = self.file_write
                    write(FRAME)
                    write(pack("<Q", n))
                    write(data)
                f.seek()
                f.truncate()
 
    def write(self, data):
        if self.current_frame:
            return self.current_frame.write(data)
        else:
            return self.file_write(data)
 
class _Unframer:
 
    def __init__(self, file_read, file_readline, file_tell=None):
        self.file_read = file_read
        self.file_readline = file_readline
        self.current_frame = None
 
    def read(self, n):
        if self.current_frame:
            data = self.current_frame.read(n)
            if not data and n != :
                self.current_frame = None
                return self.file_read(n)
            if len(data) < n:
                raise UnpicklingError(
                    "pickle exhausted before end of frame")
            return data
        else:
            return self.file_read(n)
 
    def readline(self):
        if self.current_frame:
            data = self.current_frame.readline()
            if not data:
                self.current_frame = None
                return self.file_readline()
            if data[-] != b'\n'[]:
                raise UnpicklingError(
                    "pickle exhausted before end of frame")
            return data
        else:
            return self.file_readline()
 
    def load_frame(self, frame_size):
        if self.current_frame and self.current_frame.read() != b'':
            raise UnpicklingError(
                "beginning of a new frame before end of current frame")
        self.current_frame = io.BytesIO(self.file_read(frame_size))
 
# Tools used for pickling.
 
def _getattribute(obj, name):
    for subpath in name.split('.'):
        if subpath == '<locals>':
            raise AttributeError("Can't get local attribute {!r} on {!r}"
                                 .format(name, obj))
        try:
            parent = obj
            obj = getattr(obj, subpath)
        except AttributeError:
            raise AttributeError("Can't get attribute {!r} on {!r}"
                                 .format(name, obj))
    return obj, parent
 
def whichmodule(obj, name):
    """Find the module an object belong to."""
    module_name = getattr(obj, '__module__', None)
    if module_name is not None:
        return module_name
    # Protect the iteration by using a list copy of sys.modules against dynamic
    # modules that trigger imports of other modules upon calls to getattr.
    for module_name, module in list(sys.modules.items()):
        if module_name == '__main__' or module is None:
            continue
        try:
            if _getattribute(module, name)[] is obj:
                return module_name
        except AttributeError:
            pass
    return '__main__'
 
def encode_long(x):
    r"""Encode a long to a two's complement little-endian binary string.
    Note that  is a special case, returning an empty string, to save a
    byte in the LONG1 pickling context.
 
    >>> encode_long()
    b''
    >>> encode_long()
    b'\xff\x00'
    >>> encode_long()
    b'\xff\x7f'
    >>> encode_long(-)
    b'\x00\xff'
    >>> encode_long(-)
    b'\x00\x80'
    >>> encode_long(-)
    b'\x80'
    >>> encode_long()
    b'\x7f'
    >>>
    """
    if x == :
        return b''
    nbytes = (x.bit_length() >> ) +
    result = x.to_bytes(nbytes, byteorder='little', signed=True)
    if x <  and nbytes > :
        if result[-] == 0xff and (result[-] & 0x80) != :
            result = result[:-]
    return result
 
def decode_long(data):
    r"""Decode a long from a two's complement little-endian binary string.
 
    >>> decode_long(b'')
 
    >>> decode_long(b"\xff\x00")
 
    >>> decode_long(b"\xff\x7f")
 
    >>> decode_long(b"\x00\xff")
    -
    >>> decode_long(b"\x00\x80")
    -
    >>> decode_long(b"\x80")
    -
    >>> decode_long(b"\x7f")
 
    """
    return int.from_bytes(data, byteorder='little', signed=True)
 
# Pickling machinery
 
class _Pickler:
 
    def __init__(self, file, protocol=None, *, fix_imports=True):
        """This takes a binary file for writing a pickle data stream.
 
        The optional *protocol* argument tells the pickler to use the
        given protocol; supported protocols are , , ,  and .  The
        default protocol is ; a backward-incompatible protocol designed
        for Python .
 
        Specifying a negative protocol version selects the highest
        protocol version supported.  The higher the protocol used, the
        more recent the version of Python needed to read the pickle
        produced.
 
        The *file* argument must have a write() method that accepts a
        single bytes argument. It can thus be a file object opened for
        binary writing, an io.BytesIO instance, or any other custom
        object that meets this interface.
 
        If *fix_imports* is True and *protocol* is less than , pickle
        will try to map the new Python  names to the old module names
        used in Python , so that the pickle data stream is readable
        with Python .
        """
        if protocol is None:
            protocol = DEFAULT_PROTOCOL
        if protocol < :
            protocol = HIGHEST_PROTOCOL
        elif not  <= protocol <= HIGHEST_PROTOCOL:
            raise ValueError("pickle protocol must be <= %d" % HIGHEST_PROTOCOL)
        try:
            self._file_write = file.write
        except AttributeError:
            raise TypeError("file must have a 'write' attribute")
        self.framer = _Framer(self._file_write)
        self.write = self.framer.write
        self.memo = {}
        self.proto = int(protocol)
        self.bin = protocol >=
        self.fast =
        self.fix_imports = fix_imports and protocol < 
 
    def clear_memo(self):
        """Clears the pickler's "memo".
 
        The memo is the data structure that remembers which objects the
        pickler has already seen, so that shared or recursive objects
        are pickled by reference and not by value.  This method is
        useful when re-using picklers.
        """
        self.memo.clear()
 
    def dump(self, obj):
        """Write a pickled representation of obj to the open file."""
        # Check whether Pickler was initialized correctly. This is
        # only needed to mimic the behavior of _pickle.Pickler.dump().
        if not hasattr(self, "_file_write"):
            raise PicklingError("Pickler.__init__() was not called by "
                                "%s.__init__()" % (self.__class__.__name__,))
        if self.proto >= :
            self.write(PROTO + pack("<B", self.proto))
        if self.proto >= :
            self.framer.start_framing()
        self.save(obj)
        self.write(STOP)
        self.framer.end_framing()
 
    def memoize(self, obj):
        """Store an object in the memo."""
 
        # The Pickler memo is a dictionary mapping object ids to -tuples
        # that contain the Unpickler memo key and the object being memoized.
        # The memo key is written to the pickle and will become
        # the key in the Unpickler's memo.  The object is stored in the
        # Pickler memo so that transient objects are kept alive during
        # pickling.
 
        # The use of the Unpickler memo length as the memo key is just a
        # convention.  The only requirement is that the memo values be unique.
        # But there appears no advantage to any other scheme, and this
        # scheme allows the Unpickler memo to be implemented as a plain (but
        # growable) array, indexed by memo key.
        if self.fast:
            return
        assert id(obj) not in self.memo
        idx = len(self.memo)
        self.write(self.put(idx))
        self.memo[id(obj)] = idx, obj
 
    # Return a PUT (BINPUT, LONG_BINPUT) opcode string, with argument i.
    def put(self, idx):
        if self.proto >= :
            return MEMOIZE
        elif self.bin:
            if idx < :
                return BINPUT + pack("<B", idx)
            else:
                return LONG_BINPUT + pack("<I", idx)
        else:
            return PUT + repr(idx).encode("ascii") + b'\n'
 
    # Return a GET (BINGET, LONG_BINGET) opcode string, with argument i.
    def get(self, i):
        if self.bin:
            if i < :
                return BINGET + pack("<B", i)
            else:
                return LONG_BINGET + pack("<I", i)
 
        return GET + repr(i).encode("ascii") + b'\n'
 
    def save(self, obj, save_persistent_id=True):
        self.framer.commit_frame()
 
        # Check for persistent id (defined by a subclass)
        pid = self.persistent_id(obj)
        if pid is not None and save_persistent_id:
            self.save_pers(pid)
            return
 
        # Check the memo
        x = self.memo.get(id(obj))
        if x is not None:
            self.write(self.get(x[]))
            return
 
        # Check the type dispatch table
        t = type(obj)
        f = self.dispatch.get(t)
        if f is not None:
            f(self, obj) # Call unbound method with explicit self
            return
 
        # Check private dispatch table if any, or else copyreg.dispatch_table
        reduce = getattr(self, 'dispatch_table', dispatch_table).get(t)
        if reduce is not None:
            rv = reduce(obj)
        else:
            # Check for a class with a custom metaclass; treat as regular class
            try:
                issc = issubclass(t, type)
            except TypeError: # t is not a class (old Boost; see SF #)
                issc = False
            if issc:
                self.save_global(obj)
                return
 
            # Check for a __reduce_ex__ method, fall back to __reduce__
            reduce = getattr(obj, "__reduce_ex__", None)
            if reduce is not None:
                rv = reduce(self.proto)
            else:
                reduce = getattr(obj, "__reduce__", None)
                if reduce is not None:
                    rv = reduce()
                else:
                    raise PicklingError("Can't pickle %r object: %r" %
                                        (t.__name__, obj))
 
        # Check for string returned by reduce(), meaning "save as global"
        if isinstance(rv, str):
            self.save_global(obj, rv)
            return
 
        # Assert that reduce() returned a tuple
        if not isinstance(rv, tuple):
            raise PicklingError("%s must return string or tuple" % reduce)
 
        # Assert that it returned an appropriately sized tuple
        l = len(rv)
        if not ( <= l <= ):
            raise PicklingError("Tuple returned by %s must have "
                                "two to five elements" % reduce)
 
        # Save the reduce() output and finally memoize the object
        self.save_reduce(obj=obj, *rv)
 
    def persistent_id(self, obj):
        # This exists so a subclass can override it
        return None
 
    def save_pers(self, pid):
        # Save a persistent id reference
        if self.bin:
            self.save(pid, save_persistent_id=False)
            self.write(BINPERSID)
        else:
            self.write(PERSID + str(pid).encode("ascii") + b'\n')
 
    def save_reduce(self, func, args, state=None, listitems=None,
                    dictitems=None, obj=None):
        # This API is called by some subclasses
 
        if not isinstance(args, tuple):
            raise PicklingError("args from save_reduce() must be a tuple")
        if not callable(func):
            raise PicklingError("func from save_reduce() must be callable")
 
        save = self.save
        write = self.write
 
        func_name = getattr(func, "__name__", "")
        if self.proto >=  and func_name == "__newobj_ex__":
            cls, args, kwargs = args
            if not hasattr(cls, "__new__"):
                raise PicklingError("args[0] from {} args has no __new__"
                                    .format(func_name))
            if obj is not None and cls is not obj.__class__:
                raise PicklingError("args[0] from {} args has the wrong class"
                                    .format(func_name))
            save(cls)
            save(args)
            save(kwargs)
            write(NEWOBJ_EX)
        elif self.proto >=  and func_name == "__newobj__":
            # A __reduce__ implementation can direct protocol  or newer to
            # use the more efficient NEWOBJ opcode, while still
            # allowing protocol  and  to work normally.  For this to
            # work, the function returned by __reduce__ should be
            # called __newobj__, and its first argument should be a
            # class.  The implementation for __newobj__
            # should be as follows, although pickle has no way to
            # verify this:
            #
            # def __newobj__(cls, *args):
            #     return cls.__new__(cls, *args)
            #
            # Protocols  and  will pickle a reference to __newobj__,
            # while protocol  (and above) will pickle a reference to
            # cls, the remaining args tuple, and the NEWOBJ code,
            # which calls cls.__new__(cls, *args) at unpickling time
            # (see load_newobj below).  If __reduce__ returns a
            # three-tuple, the state from the third tuple item will be
            # pickled regardless of the protocol, calling __setstate__
            # at unpickling time (see load_build below).
            #
            # Note that no standard __newobj__ implementation exists;
            # you have to provide your own.  This is to enforce
            # compatibility with Python 2.2 (pickles written using
            # protocol  or  in Python 2.3 should be unpicklable by
            # Python 2.2).
            cls = args[]
            if not hasattr(cls, "__new__"):
                raise PicklingError(
                    "args[0] from __newobj__ args has no __new__")
            if obj is not None and cls is not obj.__class__:
                raise PicklingError(
                    "args[0] from __newobj__ args has the wrong class")
            args = args[:]
            save(cls)
            save(args)
            write(NEWOBJ)
        else:
            save(func)
            save(args)
            write(REDUCE)
 
        if obj is not None:
            # If the object is already in the memo, this means it is
            # recursive. In this case, throw away everything we put on the
            # stack, and fetch the object back from the memo.
            if id(obj) in self.memo:
                write(POP + self.get(self.memo[id(obj)][]))
            else:
                self.memoize(obj)
 
        # More new special cases (that work with older protocols as
        # well): when __reduce__ returns a tuple with  or  items,
        # the 4th and 5th item should be iterators that provide list
        # items and dict items (as (key, value) tuples), or None.
 
        if listitems is not None:
            self._batch_appends(listitems)
 
        if dictitems is not None:
            self._batch_setitems(dictitems)
 
        if state is not None:
            save(state)
            write(BUILD)
 
    # Methods below this point are dispatched through the dispatch table
 
    dispatch = {}
 
    def save_none(self, obj):
        self.write(NONE)
    dispatch[type(None)] = save_none
 
    def save_bool(self, obj):
        if self.proto >= :
            self.write(NEWTRUE if obj else NEWFALSE)
        else:
            self.write(TRUE if obj else FALSE)
    dispatch[bool] = save_bool
 
    def save_long(self, obj):
        if self.bin:
            # If the int is small enough to fit in a signed -byte 's-comp
            # format, we can store it more efficiently than the general
            # case.
            # First one- and two-byte unsigned ints:
            if obj >= :
                if obj <= 0xff:
                    self.write(BININT1 + pack("<B", obj))
                    return
                if obj <= 0xffff:
                    self.write(BININT2 + pack("<H", obj))
                    return
            # Next check for -byte signed ints:
            if -0x80000000 <= obj <= 0x7fffffff:
                self.write(BININT + pack("<i", obj))
                return
        if self.proto >= :
            encoded = encode_long(obj)
            n = len(encoded)
            if n < :
                self.write(LONG1 + pack("<B", n) + encoded)
            else:
                self.write(LONG4 + pack("<i", n) + encoded)
            return
        self.write(LONG + repr(obj).encode("ascii") + b'L\n')
    dispatch[int] = save_long
 
    def save_float(self, obj):
        if self.bin:
            self.write(BINFLOAT + pack('>d', obj))
        else:
            self.write(FLOAT + repr(obj).encode("ascii") + b'\n')
    dispatch[float] = save_float
 
    def save_bytes(self, obj):
        if self.proto < :
            if not obj: # bytes object is empty
                self.save_reduce(bytes, (), obj=obj)
            else:
                self.save_reduce(codecs.encode,
                                 (str(obj, 'latin1'), 'latin1'), obj=obj)
            return
        n = len(obj)
        if n <= 0xff:
            self.write(SHORT_BINBYTES + pack("<B", n) + obj)
        elif n > 0xffffffff and self.proto >= :
            self.write(BINBYTES8 + pack("<Q", n) + obj)
        else:
            self.write(BINBYTES + pack("<I", n) + obj)
        self.memoize(obj)
    dispatch[bytes] = save_bytes
 
    def save_str(self, obj):
        if self.bin:
            encoded = obj.encode('utf-8', 'surrogatepass')
            n = len(encoded)
            if n <= 0xff and self.proto >= :
                self.write(SHORT_BINUNICODE + pack("<B", n) + encoded)
            elif n > 0xffffffff and self.proto >= :
                self.write(BINUNICODE8 + pack("<Q", n) + encoded)
            else:
                self.write(BINUNICODE + pack("<I", n) + encoded)
        else:
            obj = obj.replace("\\", "\\u005c")
            obj = obj.replace("\n", "\\u000a")
            self.write(UNICODE + obj.encode('raw-unicode-escape') +
                       b'\n')
        self.memoize(obj)
    dispatch[str] = save_str
 
    def save_tuple(self, obj):
        if not obj: # tuple is empty
            if self.bin:
                self.write(EMPTY_TUPLE)
            else:
                self.write(MARK + TUPLE)
            return
 
        n = len(obj)
        save = self.save
        memo = self.memo
        if n <=  and self.proto >= :
            for element in obj:
                save(element)
            # Subtle.  Same as in the big comment below.
            if id(obj) in memo:
                get = self.get(memo[id(obj)][])
                self.write(POP * n + get)
            else:
                self.write(_tuplesize2code[n])
                self.memoize(obj)
            return
 
        # proto  or proto  and tuple isn't empty, or proto > 1 and tuple
        # has more than  elements.
        write = self.write
        write(MARK)
        for element in obj:
            save(element)
 
        if id(obj) in memo:
            # Subtle.  d was not in memo when we entered save_tuple(), so
            # the process of saving the tuple's elements must have saved
            # the tuple itself:  the tuple is recursive.  The proper action
            # now is to throw away everything we put on the stack, and
            # simply GET the tuple (it's already constructed).  This check
            # could have been done in the "for element" loop instead, but
            # recursive tuples are a rare thing.
            get = self.get(memo[id(obj)][])
            if self.bin:
                write(POP_MARK + get)
            else:   # proto  -- POP_MARK not available
                write(POP * (n+) + get)
            return
 
        # No recursion.
        write(TUPLE)
        self.memoize(obj)
 
    dispatch[tuple] = save_tuple
 
    def save_list(self, obj):
        if self.bin:
            self.write(EMPTY_LIST)
        else:   # proto  -- can't use EMPTY_LIST
            self.write(MARK + LIST)
 
        self.memoize(obj)
        self._batch_appends(obj)
 
    dispatch[list] = save_list
 
    _BATCHSIZE = 
 
    def _batch_appends(self, items):
        # Helper to batch up APPENDS sequences
        save = self.save
        write = self.write
 
        if not self.bin:
            for x in items:
                save(x)
                write(APPEND)
            return
 
        it = iter(items)
        while True:
            tmp = list(islice(it, self._BATCHSIZE))
            n = len(tmp)
            if n > :
                write(MARK)
                for x in tmp:
                    save(x)
                write(APPENDS)
            elif n:
                save(tmp[])
                write(APPEND)
            # else tmp is empty, and we're done
            if n < self._BATCHSIZE:
                return
 
    def save_dict(self, obj):
        if self.bin:
            self.write(EMPTY_DICT)
        else:   # proto  -- can't use EMPTY_DICT
            self.write(MARK + DICT)
 
        self.memoize(obj)
        self._batch_setitems(obj.items())
 
    dispatch[dict] = save_dict
    if PyStringMap is not None:
        dispatch[PyStringMap] = save_dict
 
    def _batch_setitems(self, items):
        # Helper to batch up SETITEMS sequences; proto >=  only
        save = self.save
        write = self.write
 
        if not self.bin:
            for k, v in items:
                save(k)
                save(v)
                write(SETITEM)
            return
 
        it = iter(items)
        while True:
            tmp = list(islice(it, self._BATCHSIZE))
            n = len(tmp)
            if n > :
                write(MARK)
                for k, v in tmp:
                    save(k)
                    save(v)
                write(SETITEMS)
            elif n:
                k, v = tmp[]
                save(k)
                save(v)
                write(SETITEM)
            # else tmp is empty, and we're done
            if n < self._BATCHSIZE:
                return
 
    def save_set(self, obj):
        save = self.save
        write = self.write
 
        if self.proto < :
            self.save_reduce(set, (list(obj),), obj=obj)
            return
 
        write(EMPTY_SET)
        self.memoize(obj)
 
        it = iter(obj)
        while True:
            batch = list(islice(it, self._BATCHSIZE))
            n = len(batch)
            if n > :
                write(MARK)
                for item in batch:
                    save(item)
                write(ADDITEMS)
            if n < self._BATCHSIZE:
                return
    dispatch[set] = save_set
 
    def save_frozenset(self, obj):
        save = self.save
        write = self.write
 
        if self.proto < :
            self.save_reduce(frozenset, (list(obj),), obj=obj)
            return
 
        write(MARK)
        for item in obj:
            save(item)
 
        if id(obj) in self.memo:
            # If the object is already in the memo, this means it is
            # recursive. In this case, throw away everything we put on the
            # stack, and fetch the object back from the memo.
            write(POP_MARK + self.get(self.memo[id(obj)][]))
            return
 
        write(FROZENSET)
        self.memoize(obj)
    dispatch[frozenset] = save_frozenset
 
    def save_global(self, obj, name=None):
        write = self.write
        memo = self.memo
 
        if name is None:
            name = getattr(obj, '__qualname__', None)
        if name is None:
            name = obj.__name__
 
        module_name = whichmodule(obj, name)
        try:
            __import__(module_name, level=)
            module = sys.modules[module_name]
            obj2, parent = _getattribute(module, name)
        except (ImportError, KeyError, AttributeError):
            raise PicklingError(
                "Can't pickle %r: it's not found as %s.%s" %
                (obj, module_name, name))
        else:
            if obj2 is not obj:
                raise PicklingError(
                    "Can't pickle %r: it's not the same object as %s.%s" %
                    (obj, module_name, name))
 
        if self.proto >= :
            code = _extension_registry.get((module_name, name))
            if code:
                assert code >
                if code <= 0xff:
                    write(EXT1 + pack("<B", code))
                elif code <= 0xffff:
                    write(EXT2 + pack("<H", code))
                else:
                    write(EXT4 + pack("<i", code))
                return
        lastname = name.rpartition('.')[]
        if parent is module:
            name = lastname
        # Non-ASCII identifiers are supported only with protocols >= .
        if self.proto >= :
            self.save(module_name)
            self.save(name)
            write(STACK_GLOBAL)
        elif parent is not module:
            self.save_reduce(getattr, (parent, lastname))
        elif self.proto >= :
            write(GLOBAL + bytes(module_name, "utf-8") + b'\n' +
                  bytes(name, "utf-8") + b'\n')
        else:
            if self.fix_imports:
                r_name_mapping = _compat_pickle.REVERSE_NAME_MAPPING
                r_import_mapping = _compat_pickle.REVERSE_IMPORT_MAPPING
                if (module_name, name) in r_name_mapping:
                    module_name, name = r_name_mapping[(module_name, name)]
                elif module_name in r_import_mapping:
                    module_name = r_import_mapping[module_name]
            try:
                write(GLOBAL + bytes(module_name, "ascii") + b'\n' +
                      bytes(name, "ascii") + b'\n')
            except UnicodeEncodeError:
                raise PicklingError(
                    "can't pickle global identifier '%s.%s' using "
                    "pickle protocol %i" % (module, name, self.proto))
 
        self.memoize(obj)
 
    def save_type(self, obj):
        if obj is type(None):
            return self.save_reduce(type, (None,), obj=obj)
        elif obj is type(NotImplemented):
            return self.save_reduce(type, (NotImplemented,), obj=obj)
        elif obj is type(...):
            return self.save_reduce(type, (...,), obj=obj)
        return self.save_global(obj)
 
    dispatch[FunctionType] = save_global
    dispatch[type] = save_type
 
# Unpickling machinery
 
class _Unpickler:
 
    def __init__(self, file, *, fix_imports=True,
                 encoding="ASCII", errors="strict"):
        """This takes a binary file for reading a pickle data stream.
 
        The protocol version of the pickle is detected automatically, so
        no proto argument is needed.
 
        The argument *file* must have two methods, a read() method that
        takes an integer argument, and a readline() method that requires
        no arguments.  Both methods should return bytes.  Thus *file*
        can be a binary file object opened for reading, an io.BytesIO
        object, or any other custom object that meets this interface.
 
        The file-like object must have two methods, a read() method
        that takes an integer argument, and a readline() method that
        requires no arguments.  Both methods should return bytes.
        Thus file-like object can be a binary file object opened for
        reading, a BytesIO object, or any other custom object that
        meets this interface.
 
        Optional keyword arguments are *fix_imports*, *encoding* and
        *errors*, which are used to control compatibility support for
        pickle stream generated by Python .  If *fix_imports* is True,
        pickle will try to map the old Python  names to the new names
        used in Python .  The *encoding* and *errors* tell pickle how
        to decode -bit string instances pickled by Python ; these
        default to 'ASCII' and 'strict', respectively. *encoding* can be
        'bytes' to read theses -bit string instances as bytes objects.
        """
        self._file_readline = file.readline
        self._file_read = file.read
        self.memo = {}
        self.encoding = encoding
        self.errors = errors
        self.proto =
        self.fix_imports = fix_imports
 
    def load(self):
        """Read a pickled object representation from the open file.
 
        Return the reconstituted object hierarchy specified in the file.
        """
        # Check whether Unpickler was initialized correctly. This is
        # only needed to mimic the behavior of _pickle.Unpickler.dump().
        if not hasattr(self, "_file_read"):
            raise UnpicklingError("Unpickler.__init__() was not called by "
                                  "%s.__init__()" % (self.__class__.__name__,))
        self._unframer = _Unframer(self._file_read, self._file_readline)
        self.read = self._unframer.read
        self.readline = self._unframer.readline
        self.mark = object() # any new unique object
        self.stack = []
        self.append = self.stack.append
        self.proto =
        read = self.read
        dispatch = self.dispatch
        try:
            while True:
                key = read()
                if not key:
                    raise EOFError
                assert isinstance(key, bytes_types)
                dispatch[key[]](self)
        except _Stop as stopinst:
            return stopinst.value
 
    # Return largest index k such that self.stack[k] is self.mark.
    # If the stack doesn't contain a mark, eventually raises IndexError.
    # This could be sped by maintaining another stack, of indices at which
    # the mark appears.  For that matter, the latter stack would suffice,
    # and we wouldn't need to push mark objects on self.stack at all.
    # Doing so is probably a good thing, though, since if the pickle is
    # corrupt (or hostile) we may get a clue from finding self.mark embedded
    # in unpickled objects.
    def marker(self):
        stack = self.stack
        mark = self.mark
        k = len(stack)-
        while stack[k] is not mark: k = k-
        return k
 
    def persistent_load(self, pid):
        raise UnpicklingError("unsupported persistent id encountered")
 
    dispatch = {}
 
    def load_proto(self):
        proto = self.read()[]
        if not  <= proto <= HIGHEST_PROTOCOL:
            raise ValueError("unsupported pickle protocol: %d" % proto)
        self.proto = proto
    dispatch[PROTO[]] = load_proto
 
    def load_frame(self):
        frame_size, = unpack('<Q', self.read())
        if frame_size > sys.maxsize:
            raise ValueError("frame size > sys.maxsize: %d" % frame_size)
        self._unframer.load_frame(frame_size)
    dispatch[FRAME[]] = load_frame
 
    def load_persid(self):
        pid = self.readline()[:-].decode("ascii")
        self.append(self.persistent_load(pid))
    dispatch[PERSID[]] = load_persid
 
    def load_binpersid(self):
        pid = self.stack.pop()
        self.append(self.persistent_load(pid))
    dispatch[BINPERSID[]] = load_binpersid
 
    def load_none(self):
        self.append(None)
    dispatch[NONE[]] = load_none
 
    def load_false(self):
        self.append(False)
    dispatch[NEWFALSE[]] = load_false
 
    def load_true(self):
        self.append(True)
    dispatch[NEWTRUE[]] = load_true
 
    def load_int(self):
        data = self.readline()
        if data == FALSE[:]:
            val = False
        elif data == TRUE[:]:
            val = True
        else:
            val = int(data, )
        self.append(val)
    dispatch[INT[]] = load_int
 
    def load_binint(self):
        self.append(unpack('<i', self.read())[])
    dispatch[BININT[]] = load_binint
 
    def load_binint1(self):
        self.append(self.read()[])
    dispatch[BININT1[]] = load_binint1
 
    def load_binint2(self):
        self.append(unpack('<H', self.read())[])
    dispatch[BININT2[]] = load_binint2
 
    def load_long(self):
        val = self.readline()[:-]
        if val and val[-] == b'L'[]:
            val = val[:-]
        self.append(int(val, ))
    dispatch[LONG[]] = load_long
 
    def load_long1(self):
        n = self.read()[]
        data = self.read(n)
        self.append(decode_long(data))
    dispatch[LONG1[]] = load_long1
 
    def load_long4(self):
        n, = unpack('<i', self.read())
        if n < :
            # Corrupt or hostile pickle -- we never write one like this
            raise UnpicklingError("LONG pickle has negative byte count")
        data = self.read(n)
        self.append(decode_long(data))
    dispatch[LONG4[]] = load_long4
 
    def load_float(self):
        self.append(float(self.readline()[:-]))
    dispatch[FLOAT[]] = load_float
 
    def load_binfloat(self):
        self.append(unpack('>d', self.read())[])
    dispatch[BINFLOAT[]] = load_binfloat
 
    def _decode_string(self, value):
        # Used to allow strings from Python  to be decoded either as
        # bytes or Unicode strings.  This should be used only with the
        # STRING, BINSTRING and SHORT_BINSTRING opcodes.
        if self.encoding == "bytes":
            return value
        else:
            return value.decode(self.encoding, self.errors)
 
    def load_string(self):
        data = self.readline()[:-]
        # Strip outermost quotes
        if len(data) >=  and data[] == data[-] and data[] in b'"\'':
            data = data[:-]
        else:
            raise UnpicklingError("the STRING opcode argument must be quoted")
        self.append(self._decode_string(codecs.escape_decode(data)[]))
    dispatch[STRING[]] = load_string
 
    def load_binstring(self):
        # Deprecated BINSTRING uses signed -bit length
        len, = unpack('<i', self.read())
        if len < :
            raise UnpicklingError("BINSTRING pickle has negative byte count")
        data = self.read(len)
        self.append(self._decode_string(data))
    dispatch[BINSTRING[]] = load_binstring
 
    def load_binbytes(self):
        len, = unpack('<I', self.read())
        if len > maxsize:
            raise UnpicklingError("BINBYTES exceeds system's maximum size "
                                  "of %d bytes" % maxsize)
        self.append(self.read(len))
    dispatch[BINBYTES[]] = load_binbytes
 
    def load_unicode(self):
        self.append(str(self.readline()[:-], 'raw-unicode-escape'))
    dispatch[UNICODE[]] = load_unicode
 
    def load_binunicode(self):
        len, = unpack('<I', self.read())
        if len > maxsize:
            raise UnpicklingError("BINUNICODE exceeds system's maximum size "
                                  "of %d bytes" % maxsize)
        self.append(str(self.read(len), 'utf-8', 'surrogatepass'))
    dispatch[BINUNICODE[]] = load_binunicode
 
    def load_binunicode8(self):
        len, = unpack('<Q', self.read())
        if len > maxsize:
            raise UnpicklingError("BINUNICODE8 exceeds system's maximum size "
                                  "of %d bytes" % maxsize)
        self.append(str(self.read(len), 'utf-8', 'surrogatepass'))
    dispatch[BINUNICODE8[]] = load_binunicode8
 
    def load_binbytes8(self):
        len, = unpack('<Q', self.read())
        if len > maxsize:
            raise UnpicklingError("BINBYTES8 exceeds system's maximum size "
                                  "of %d bytes" % maxsize)
        self.append(self.read(len))
    dispatch[BINBYTES8[]] = load_binbytes8
 
    def load_short_binstring(self):
        len = self.read()[]
        data = self.read(len)
        self.append(self._decode_string(data))
    dispatch[SHORT_BINSTRING[]] = load_short_binstring
 
    def load_short_binbytes(self):
        len = self.read()[]
        self.append(self.read(len))
    dispatch[SHORT_BINBYTES[]] = load_short_binbytes
 
    def load_short_binunicode(self):
        len = self.read()[]
        self.append(str(self.read(len), 'utf-8', 'surrogatepass'))
    dispatch[SHORT_BINUNICODE[]] = load_short_binunicode
 
    def load_tuple(self):
        k = self.marker()
        self.stack[k:] = [tuple(self.stack[k+:])]
    dispatch[TUPLE[]] = load_tuple
 
    def load_empty_tuple(self):
        self.append(())
    dispatch[EMPTY_TUPLE[]] = load_empty_tuple
 
    def load_tuple1(self):
        self.stack[-] = (self.stack[-],)
    dispatch[TUPLE1[]] = load_tuple1
 
    def load_tuple2(self):
        self.stack[-:] = [(self.stack[-], self.stack[-])]
    dispatch[TUPLE2[]] = load_tuple2
 
    def load_tuple3(self):
        self.stack[-:] = [(self.stack[-], self.stack[-], self.stack[-])]
    dispatch[TUPLE3[]] = load_tuple3
 
    def load_empty_list(self):
        self.append([])
    dispatch[EMPTY_LIST[]] = load_empty_list
 
    def load_empty_dictionary(self):
        self.append({})
    dispatch[EMPTY_DICT[]] = load_empty_dictionary
 
    def load_empty_set(self):
        self.append(set())
    dispatch[EMPTY_SET[]] = load_empty_set
 
    def load_frozenset(self):
        k = self.marker()
        self.stack[k:] = [frozenset(self.stack[k+:])]
    dispatch[FROZENSET[]] = load_frozenset
 
    def load_list(self):
        k = self.marker()
        self.stack[k:] = [self.stack[k+:]]
    dispatch[LIST[]] = load_list
 
    def load_dict(self):
        k = self.marker()
        items = self.stack[k+:]
        d = {items[i]: items[i+]
             for i in range(, len(items), )}
        self.stack[k:] = [d]
    dispatch[DICT[]] = load_dict
 
    # INST and OBJ differ only in how they get a class object.  It's not
    # only sensible to do the rest in a common routine, the two routines
    # previously diverged and grew different bugs.
    # klass is the class to instantiate, and k points to the topmost mark
    # object, following which are the arguments for klass.__init__.
    def _instantiate(self, klass, k):
        args = tuple(self.stack[k+:])
        del self.stack[k:]
        if (args or not isinstance(klass, type) or
            hasattr(klass, "__getinitargs__")):
            try:
                value = klass(*args)
            except TypeError as err:
                raise TypeError("in constructor for %s: %s" %
                                (klass.__name__, str(err)), sys.exc_info()[])
        else:
            value = klass.__new__(klass)
        self.append(value)
 
    def load_inst(self):
        module = self.readline()[:-].decode("ascii")
        name = self.readline()[:-].decode("ascii")
        klass = self.find_class(module, name)
        self._instantiate(klass, self.marker())
    dispatch[INST[]] = load_inst
 
    def load_obj(self):
        # Stack is ... markobject classobject arg1 arg2 ...
        k = self.marker()
        klass = self.stack.pop(k+)
        self._instantiate(klass, k)
    dispatch[OBJ[]] = load_obj
 
    def load_newobj(self):
        args = self.stack.pop()
        cls = self.stack.pop()
        obj = cls.__new__(cls, *args)
        self.append(obj)
    dispatch[NEWOBJ[]] = load_newobj
 
    def load_newobj_ex(self):
        kwargs = self.stack.pop()
        args = self.stack.pop()
        cls = self.stack.pop()
        obj = cls.__new__(cls, *args, **kwargs)
        self.append(obj)
    dispatch[NEWOBJ_EX[]] = load_newobj_ex
 
    def load_global(self):
        module = self.readline()[:-].decode("utf-8")
        name = self.readline()[:-].decode("utf-8")
        klass = self.find_class(module, name)
        self.append(klass)
    dispatch[GLOBAL[]] = load_global
 
    def load_stack_global(self):
        name = self.stack.pop()
        module = self.stack.pop()
        if type(name) is not str or type(module) is not str:
            raise UnpicklingError("STACK_GLOBAL requires str")
        self.append(self.find_class(module, name))
    dispatch[STACK_GLOBAL[]] = load_stack_global
 
    def load_ext1(self):
        code = self.read()[]
        self.get_extension(code)
    dispatch[EXT1[]] = load_ext1
 
    def load_ext2(self):
        code, = unpack('<H', self.read())
        self.get_extension(code)
    dispatch[EXT2[]] = load_ext2
 
    def load_ext4(self):
        code, = unpack('<i', self.read())
        self.get_extension(code)
    dispatch[EXT4[]] = load_ext4
 
    def get_extension(self, code):
        nil = []
        obj = _extension_cache.get(code, nil)
        if obj is not nil:
            self.append(obj)
            return
        key = _inverted_registry.get(code)
        if not key:
            if code <= : # note that  is forbidden
                # Corrupt or hostile pickle.
                raise UnpicklingError("EXT specifies code <= 0")
            raise ValueError("unregistered extension code %d" % code)
        obj = self.find_class(*key)
        _extension_cache[code] = obj
        self.append(obj)
 
    def find_class(self, module, name):
        # Subclasses may override this.
        if self.proto <  and self.fix_imports:
            if (module, name) in _compat_pickle.NAME_MAPPING:
                module, name = _compat_pickle.NAME_MAPPING[(module, name)]
            elif module in _compat_pickle.IMPORT_MAPPING:
                module = _compat_pickle.IMPORT_MAPPING[module]
        __import__(module, level=)
        if self.proto >= :
            return _getattribute(sys.modules[module], name)[]
        else:
            return getattr(sys.modules[module], name)
 
    def load_reduce(self):
        stack = self.stack
        args = stack.pop()
        func = stack[-]
        stack[-] = func(*args)
    dispatch[REDUCE[]] = load_reduce
 
    def load_pop(self):
        del self.stack[-]
    dispatch[POP[]] = load_pop
 
    def load_pop_mark(self):
        k = self.marker()
        del self.stack[k:]
    dispatch[POP_MARK[]] = load_pop_mark
 
    def load_dup(self):
        self.append(self.stack[-])
    dispatch[DUP[]] = load_dup
 
    def load_get(self):
        i = int(self.readline()[:-])
        self.append(self.memo[i])
    dispatch[GET[]] = load_get
 
    def load_binget(self):
        i = self.read()[]
        self.append(self.memo[i])
    dispatch[BINGET[]] = load_binget
 
    def load_long_binget(self):
        i, = unpack('<I', self.read())
        self.append(self.memo[i])
    dispatch[LONG_BINGET[]] = load_long_binget
 
    def load_put(self):
        i = int(self.readline()[:-])
        if i < :
            raise ValueError("negative PUT argument")
        self.memo[i] = self.stack[-]
    dispatch[PUT[]] = load_put
 
    def load_binput(self):
        i = self.read()[]
        if i < :
            raise ValueError("negative BINPUT argument")
        self.memo[i] = self.stack[-]
    dispatch[BINPUT[]] = load_binput
 
    def load_long_binput(self):
        i, = unpack('<I', self.read())
        if i > maxsize:
            raise ValueError("negative LONG_BINPUT argument")
        self.memo[i] = self.stack[-]
    dispatch[LONG_BINPUT[]] = load_long_binput
 
    def load_memoize(self):
        memo = self.memo
        memo[len(memo)] = self.stack[-]
    dispatch[MEMOIZE[]] = load_memoize
 
    def load_append(self):
        stack = self.stack
        value = stack.pop()
        list = stack[-]
        list.append(value)
    dispatch[APPEND[]] = load_append
 
    def load_appends(self):
        stack = self.stack
        mark = self.marker()
        list_obj = stack[mark - ]
        items = stack[mark + :]
        if isinstance(list_obj, list):
            list_obj.extend(items)
        else:
            append = list_obj.append
            for item in items:
                append(item)
        del stack[mark:]
    dispatch[APPENDS[]] = load_appends
 
    def load_setitem(self):
        stack = self.stack
        value = stack.pop()
        key = stack.pop()
        dict = stack[-]
        dict[key] = value
    dispatch[SETITEM[]] = load_setitem
 
    def load_setitems(self):
        stack = self.stack
        mark = self.marker()
        dict = stack[mark - ]
        for i in range(mark + , len(stack), ):
            dict[stack[i]] = stack[i + ]
 
        del stack[mark:]
    dispatch[SETITEMS[]] = load_setitems
 
    def load_additems(self):
        stack = self.stack
        mark = self.marker()
        set_obj = stack[mark - ]
        items = stack[mark + :]
        if isinstance(set_obj, set):
            set_obj.update(items)
        else:
            add = set_obj.add
            for item in items:
                add(item)
        del stack[mark:]
    dispatch[ADDITEMS[]] = load_additems
 
    def load_build(self):
        stack = self.stack
        state = stack.pop()
        inst = stack[-]
        setstate = getattr(inst, "__setstate__", None)
        if setstate is not None:
            setstate(state)
            return
        slotstate = None
        if isinstance(state, tuple) and len(state) == :
            state, slotstate = state
        if state:
            inst_dict = inst.__dict__
            intern = sys.intern
            for k, v in state.items():
                if type(k) is str:
                    inst_dict[intern(k)] = v
                else:
                    inst_dict[k] = v
        if slotstate:
            for k, v in slotstate.items():
                setattr(inst, k, v)
    dispatch[BUILD[]] = load_build
 
    def load_mark(self):
        self.append(self.mark)
    dispatch[MARK[]] = load_mark
 
    def load_stop(self):
        value = self.stack.pop()
        raise _Stop(value)
    dispatch[STOP[]] = load_stop
 
# Shorthands
 
def _dump(obj, file, protocol=None, *, fix_imports=True):
    _Pickler(file, protocol, fix_imports=fix_imports).dump(obj)
 
def _dumps(obj, protocol=None, *, fix_imports=True):
    f = io.BytesIO()
    _Pickler(f, protocol, fix_imports=fix_imports).dump(obj)
    res = f.getvalue()
    assert isinstance(res, bytes_types)
    return res
 
def _load(file, *, fix_imports=True, encoding="ASCII", errors="strict"):
    return _Unpickler(file, fix_imports=fix_imports,
                     encoding=encoding, errors=errors).load()
 
def _loads(s, *, fix_imports=True, encoding="ASCII", errors="strict"):
    if isinstance(s, str):
        raise TypeError("Can't load pickle from unicode string")
    file = io.BytesIO(s)
    return _Unpickler(file, fix_imports=fix_imports,
                      encoding=encoding, errors=errors).load()
 
# Use the faster _pickle if possible
try:
    from _pickle import (
        PickleError,
        PicklingError,
        UnpicklingError,
        Pickler,
        Unpickler,
        dump,
        dumps,
        load,
        loads
    )
except ImportError:
    Pickler, Unpickler = _Pickler, _Unpickler
    dump, dumps, load, loads = _dump, _dumps, _load, _loads
 
# Doctest
def _test():
    import doctest
    return doctest.testmod()
 
if __name__ == "__main__":
    import argparse
    parser = argparse.ArgumentParser(
        description='display contents of the pickle files')
    parser.add_argument(
        'pickle_file', type=argparse.FileType('br'),
        nargs='*', help='the pickle file')
    parser.add_argument(
        '-t', '--test', action='store_true',
        help='run self-test suite')
    parser.add_argument(
        '-v', action='store_true',
        help='run verbosely; only affects self-test run')
    args = parser.parse_args()
    if args.test:
        _test()
    else:
        if not args.pickle_file:
            parser.print_help()
        else:
            import pprint
            for f in args.pickle_file:
                obj = load(f)
                pprint.pprint(obj)