File: canonicaljson.py

package info (click to toggle)
python-canonicaljson 1.1.4-1~bpo9+1
  • links: PTS
  • area: main
  • in suites: stretch-backports
  • size: 1,736 kB
  • sloc: python: 20,709; sh: 19; makefile: 3
file content (158 lines) | stat: -rw-r--r-- 5,400 bytes parent folder | download | duplicates (3)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
# -*- coding: utf-8 -*-

# Copyright 2014 OpenMarket Ltd
# Copyright 2018 New Vector Ltd
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

import re
from six import unichr, PY2, PY3

# using simplejson rather than regular json gives approximately a 100%
# performance improvement (as measured on python 2.7.12/simplejson 3.13.2)
import simplejson as json

from frozendict import frozendict

__version__ = '1.1.4'


def _default(obj):
    if type(obj) is frozendict:
        # fishing the protected dict out of the object is a bit nasty,
        # but we don't really want the overhead of copying the dict.
        return obj._dict
    raise TypeError('Object of type %s is not JSON serializable' %
                    obj.__class__.__name__)


# ideally we'd set ensure_ascii=False, but the ensure_ascii codepath is so
# much quicker (assuming c speedups are enabled) that it's actually much
# quicker to let it do that and then substitute back (it's about 2.5x faster).
#
# (in any case, simplejson's ensure_ascii doesn't get U+2028 and U+2029 right,
# as per https://github.com/simplejson/simplejson/issues/206).
#
_canonical_encoder = json.JSONEncoder(
    ensure_ascii=True,
    separators=(',', ':'),
    sort_keys=True,
    default=_default,
)

_pretty_encoder = json.JSONEncoder(
    ensure_ascii=True,
    indent=4,
    sort_keys=True,
    default=_default,
)

# This regexp matches either `\uNNNN` or `\\`. We match '\\' (and leave it
# unchanged) to make sure that the regex doesn't accidentally capture the uNNNN
# in `\\uNNNN`, which is an escaped backslash followed by 'uNNNN'.
_U_ESCAPE = re.compile(r"\\u([0-9a-f]{4})|\\\\")


def _unascii(s):
    """Unpack `\\uNNNN` escapes in 's' and encode the result as UTF-8

    This method takes the output of the JSONEncoder and expands any \\uNNNN
    escapes it finds (except for \\u0000 to \\u001F, which are converted to
    \\xNN escapes).

    For performance, it assumes that the input is valid JSON, and performs few
    sanity checks.
    """

    # make the fast path fast: if there are no matches in the string, the
    # whole thing is ascii. On python 2, that means we're done. On python 3,
    # we have to turn it into a bytes, which is quickest with encode('utf-8')
    m = _U_ESCAPE.search(s)
    if not m:
        return s if PY2 else s.encode('utf-8')

    # appending to a string (or a bytes) is slooow, so we accumulate sections
    # of string result in 'chunks', and join them all together later.
    # (It doesn't seem to make much difference whether we accumulate
    # utf8-encoded bytes, or strings which we utf-8 encode after rejoining)
    #
    chunks = []

    # 'pos' tracks the index in 's' that we have processed into 'chunks' so
    # far.
    pos = 0

    while m:
        start = m.start()
        end = m.end()

        g = m.group(1)

        if g is None:
            # escaped backslash: pass it through along with anything before the
            # match
            chunks.append(s[pos:end])
        else:
            # \uNNNN, but we have to watch out for surrogate pairs.
            #
            # On python 2, str.encode("utf-8") will decode utf-16 surrogates
            # before re-encoding, so it's fine for us to pass the surrogates
            # through. (Indeed we must, to deal with UCS-2 python builds, per
            # https://github.com/matrix-org/python-canonicaljson/issues/12).
            #
            # On python 3, str.encode("utf-8") complains about surrogates, so
            # we have to unpack them.
            c = int(g, 16)

            if c < 0x20:
                # leave as a \uNNNN escape
                chunks.append(s[pos:end])
            else:
                if PY3:   # pragma nocover
                    if c & 0xfc00 == 0xd800 and s[end:end + 2] == '\\u':
                        esc2 = s[end + 2:end + 6]
                        c2 = int(esc2, 16)
                        if c2 & 0xfc00 == 0xdc00:
                            c = 0x10000 + (((c - 0xd800) << 10) |
                                           (c2 - 0xdc00))
                            end += 6

                chunks.append(s[pos:start])
                chunks.append(unichr(c))

        pos = end
        m = _U_ESCAPE.search(s, pos)

    # pass through anything after the last match
    chunks.append(s[pos:])

    return (''.join(chunks)).encode("utf-8")


def encode_canonical_json(json_object):
    """Encodes the shortest UTF-8 JSON encoding with dictionary keys
    lexicographically sorted by unicode code point.

    Args:
        json_object (dict): The JSON object to encode.

    Returns:
        bytes encoding the JSON object"""
    s = _canonical_encoder.encode(json_object)
    return _unascii(s)


def encode_pretty_printed_json(json_object):
    """Encodes the JSON object dict as human readable ascii bytes."""

    return _pretty_encoder.encode(json_object).encode("ascii")