ujson-ia 2.1.1

Ultra fast JSON encoder and decoder for Python (Internet Archive fork)

About this fork

We use this version at the Internet Archive. We have merged @vdmit11’s changes from https://github.com/dignio/ultrajson and the latest from upstream master, and may continue to make other tweaks.

To install:

$ pip install ujson-ia

Back to your regularly scheduled readme

UltraJSON is an ultra fast JSON encoder and decoder written in pure C with bindings for Python 2.5+ and 3.

For a more painless day to day C/C++ JSON decoder experience please checkout ujson4c, based on UltraJSON.

Please checkout the rest of the projects in the Ultra series:

http://github.com/esnme/ultramemcache

http://github.com/esnme/ultramysql

Usage

May be used as a drop in replacement for most other JSON parsers for Python:

>>> import ujson
>>> ujson.dumps([{"key": "value"}, 81, True])
'[{"key":"value"},81,true]'
>>> ujson.loads("""[{"key": "value"}, 81, true]""")
[{u'key': u'value'}, 81, True]

Encoder options

encode_html_chars

Used to enable special encoding of “unsafe” HTML characters into safer Unicode sequences. Default is False:

>>> ujson.dumps("<script>John&Doe", encode_html_chars=True)
'"\\u003cscript\\u003eJohn\\u0026Doe"'

ensure_ascii

Limits output to ASCII and escapes all extended characters above 127. Default is true. If your end format supports UTF-8 setting this option to false is highly recommended to save space:

>>> ujson.dumps(u"\xe5\xe4\xf6")
'"\\u00e5\\u00e4\\u00f6"'
>>> ujson.dumps(u"\xe5\xe4\xf6", ensure_ascii=False)
'"\xc3\xa5\xc3\xa4\xc3\xb6"'

escape_forward_slashes

Controls whether forward slashes (/) are escaped. Default is True:

>>> ujson.dumps("http://esn.me")
'"http:\/\/esn.me"'
>>> ujson.dumps("http://esn.me", escape_forward_slashes=False)
'"http://esn.me"'

indent

Controls whether indention (“pretty output”) is enabled. Default is 0 (disabled):

>>> ujson.dumps({"foo": "bar"})
'{"foo":"bar"}'
>>> ujson.dumps({"foo": "bar"}, indent=4)
{
    "foo":"bar"
}

pre_encode_hook

Allows to provide a custom function which is called for every encoded Python object.

The hook function semantics is similar to the standard JSONEncoder.default() method, but the pre_encode_hook() is called before any other serialization attempts, while the default() is called when all other options didn’t work.

That allows to override already exsiting behavior and define custom serialization formats for things like dates. For example:

# Default behavior: datetime is converted to timestamp
>>> ujson.dumps({"a": "foo", "b": datetime.now()})
'{"a":"foo","b":1454523657}'

# Hook is involved: the datetime object is replaced with the .isoformat() string
>>> def hook(obj):
        return obj.isoformat() if hasattr(obj, 'isoformat') else obj

>>> ujson.dumps({"a": "foo", "b": datetime.now()}, pre_encode_hook=hook)
'{"a":"foo","b":"2016-02-03T18:21:55.351081"}'

The hook may be used to replace any object with any other arbitrary object before encoding it. However, it doesn’t cancel all further encoding transformations. For example, if you return a datetime object from the hook instead of a string, it will be transformed to a timestamp.

pre_encode_primitive

The boolean flag that indicates that pre_encode_hook() should also be called for Python objects that serialized to primitive JSON types (Number, String, Boolean, null).

Usually you don’t need to define any special serialization format for these types, so the flag is false by default.

Enabling this flag may produce huge amount of pre_encode_hook() calls (the hook will be called for every single JSON value) and thus affect the performance.

Decoders options

precise_float

Set to enable usage of higher precision (strtod) function when decoding string to double values. Default is to use fast but less precise builtin functionality:

>>> ujson.loads("4.56")
4.5600000000000005
>>> ujson.loads("4.56", precise_float=True)
4.5599999999999996

object_hook

A custom Python function which is called after a JSON object is decoded.

The hook semantics is similar to the standard JSONDecoder.object_hook() behavior. You may use it to transform a dictionary (the decoded JSON object) into a more specific object.

For example:

>>> def hook(obj):
        if '__complex__' in obj:
            return complex(obj['real'], obj['imag'])
        return obj

>>> ujson.loads('{"__complex__": true, "real": 1, "imag": 2}', object_hook=hook)
(1+2j)

string_hook

Similar to object_hook, but called for every decoded string.

Useful for deserializing objects like dates from their textual representations, e.g.:

>>> def hook(s):
        if s.startswith('__DATE'):
            return datetime.strptime(s, '__DATE: %Y-%m-%d')
        return s

>>> ujson.loads('{"a": "foo", "b": "__DATE: 2016-01-01"}', string_hook=hook)
{'a': 'foo', 'b': datetime.datetime(2016, 1, 1, 0, 0)}

Benchmarks

UltraJSON calls/sec compared to three other popular JSON parsers with performance gain specified below each.

Test machine:

Linux 3.13.0-66-generic x86_64 #108-Ubuntu SMP Wed Oct 7 15:20:27 UTC 2015

Versions:

• CPython 2.7.6 (default, Jun 22 2015, 17:58:13) [GCC 4.8.2]

• blist : 1.3.6

• simplejson: 3.8.1

• ujson : 1.34 (0c52200eb4e2d97e548a765d5f089858c41967b0)

• yajl : 0.3.5

	ujson	yajl	simplejson	json
Array with 256 doubles
encode	3508.19	5742.00	3232.38	3309.09
decode	25103.37	11257.83	11696.26	11871.04
Array with 256 UTF-8 strings
encode	3189.71	2717.14	2006.38	2961.72
decode	1354.94	630.54	356.35	344.05
Array with 256 strings
encode	18127.47	12537.39	12541.23	20001.00
decode	23264.70	12788.85	25427.88	9352.36
Medium complex object
encode	10519.38	5021.29	3686.86	4643.47
decode	9676.53	5326.79	8515.77	3017.30
Array with 256 True values
encode	105998.03	102067.28	44758.51	60424.80
decode	163869.96	78341.57	110859.36	115013.90
Array with 256 dict{string, int} pairs
encode	13471.32	12109.09	3876.40	8833.92
decode	16890.63	8946.07	12218.55	3350.72
Dict with 256 arrays with 256 dict{string, int} pairs
encode	50.25	46.45	13.82	29.28
decode	33.27	22.10	27.91	10.43
Dict with 256 arrays with 256 dict{string, int} pairs, outputting sorted keys
encode	27.19		7.75	2.39
Complex object
encode	577.98		387.81	470.02
decode	496.73	234.44	151.00	145.16

Versions:

• CPython 3.4.3 (default, Oct 14 2015, 20:28:29) [GCC 4.8.4]

• blist : 1.3.6

• simplejson: 3.8.1

• ujson : 1.34 (0c52200eb4e2d97e548a765d5f089858c41967b0)

• yajl : 0.3.5

	ujson	yajl	simplejson	json
Array with 256 doubles
encode	3477.15	5732.24	3016.76	3071.99
decode	23625.20	9731.45	9501.57	9901.92
Array with 256 UTF-8 strings
encode	1995.89	2151.61	1771.98	1817.20
decode	1425.04	625.38	327.14	305.95
Array with 256 strings
encode	25461.75	12188.64	13054.76	14429.81
decode	21981.31	17014.22	23869.48	22483.58
Medium complex object
encode	10821.46	4837.04	3114.04	4254.46
decode	7887.77	5126.67	4934.60	6204.97
Array with 256 True values
encode	100452.86	94639.42	46657.63	60358.63
decode	148312.69	75485.90	88434.91	116395.51
Array with 256 dict{string, int} pairs
encode	11698.13	8886.96	3043.69	6302.35
decode	10686.40	7061.77	5646.80	7702.29
Dict with 256 arrays with 256 dict{string, int} pairs
encode	44.26	34.43	10.40	21.97
decode	28.46	23.95	18.70	22.83
Dict with 256 arrays with 256 dict{string, int} pairs, outputting sorted keys
encode	33.60		6.94	22.34
Complex object
encode	432.30		351.47	379.34
decode	434.40	221.97	149.57	147.79