Python 2与Python 3下的base64模块

Python2的编解码

python2中程序数据类型默认为ASCII，所以需要先将数据解码（decode）成为Unicode类型，然后再编码（encode）成为想要转换的数据类型（gbk，utf-8，gb18030，gb2312），然后再解码成为对应的数据类型显示在屏幕上；

Python3的编解码

python3中程序默认数据类型为Unicode，所以直接将数据编码（encode）成为想要转换的数据类型（gbk，utf-8，gb18030，gb2312），然后解码成为对应的数据类型显示在屏幕上。

base64

Base64编码是一种“防君子不防小人”的编码方式。广泛应用于MIME协议，作为电子邮件的传输编码，生成的编码可逆，后一两位可能有“=”，生成的编码都是ascii字符。

因此对于python2来说，编解码相对要容易一些。python3因为要从Unicode转换一下，相对麻烦一些。一切见下例：

Python2

def b64encode(s, altchars=None):
    """Encode a string using Base64.

    s is the string to encode.  Optional altchars must be a string of at least
    length 2 (additional characters are ignored) which specifies an
    alternative alphabet for the '+' and '/' characters.  This allows an
    application to e.g. generate url or filesystem safe Base64 strings.

    The encoded string is returned.
    """
    # Strip off the trailing newline
    encoded = binascii.b2a_base64(s)[:-1]
    if altchars is not None:
        return encoded.translate(string.maketrans(b'+/', altchars[:2]))
    return encoded

b64decode源码

def b64decode(s, altchars=None):
    """Decode a Base64 encoded string.

    s is the string to decode.  Optional altchars must be a string of at least
    length 2 (additional characters are ignored) which specifies the
    alternative alphabet used instead of the '+' and '/' characters.

    The decoded string is returned.  A TypeError is raised if s is
    incorrectly padded.  Characters that are neither in the normal base-64
    alphabet nor the alternative alphabet are discarded prior to the padding
    check.
    """
    if altchars is not None:
        s = s.translate(string.maketrans(altchars[:2], '+/'))
    try:
        return binascii.a2b_base64(s)
    except binascii.Error, msg:
        # Transform this exception for consistency
        raise TypeError(msg)

这里面的s是一个字符串类型的对象。

import base64

s = 'Hello, python'
b = base64.b64encode(s)
print 'b为：', b

c = base64.b64decode(b)
print 'c为：', c

# output
b为： SGVsbG8sIHB5dGhvbg==
c为： Hello, python

Python3

def b64encode(s, altchars=None):
    """Encode the bytes-like object s using Base64 and return a bytes object.

    Optional altchars should be a byte string of length 2 which specifies an
    alternative alphabet for the '+' and '/' characters.  This allows an
    application to e.g. generate url or filesystem safe Base64 strings.
    """
    encoded = binascii.b2a_base64(s, newline=False)
    if altchars is not None:
        assert len(altchars) == 2, repr(altchars)
        return encoded.translate(bytes.maketrans(b'+/', altchars))
    return encoded

b64decode源码

def b64decode(s, altchars=None, validate=False):
    """Decode the Base64 encoded bytes-like object or ASCII string s.

    Optional altchars must be a bytes-like object or ASCII string of length 2
    which specifies the alternative alphabet used instead of the '+' and '/'
    characters.

    The result is returned as a bytes object.  A binascii.Error is raised if
    s is incorrectly padded.

    If validate is False (the default), characters that are neither in the
    normal base-64 alphabet nor the alternative alphabet are discarded prior
    to the padding check.  If validate is True, these non-alphabet characters
    in the input result in a binascii.Error.
    """
    s = _bytes_from_decode_data(s)
    if altchars is not None:
        altchars = _bytes_from_decode_data(altchars)
        assert len(altchars) == 2, repr(altchars)
        s = s.translate(bytes.maketrans(altchars, b'+/'))
    if validate and not re.match(b'^[A-Za-z0-9+/]*={0,2}$', s):
        raise binascii.Error('Non-base64 digit found')
    return binascii.a2b_base64(s)

这里面的s是一个bytes对象，则字符串首先要经过编码encode()。经过b64encode/b64decode之后的返回结果也是bytes对象，所以我们要转换为Unicode对象就要再使用decode()方法去解码。

import base64

s = 'Hello, Python!'
b = base64.b64encode(s.encode('utf-8')).decode('utf-8')
print(b)

c = base64.b64decode(b.encode('utf-8')).decode('utf-8')
print(c)

# output
SGVsbG8sIFB5dGhvbiE=
Hello, Python!