1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/addon-sdk/source/python-lib/simplejson/encoder.py Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,385 @@
1.4 +"""
1.5 +Implementation of JSONEncoder
1.6 +"""
1.7 +import re
1.8 +
1.9 +try:
1.10 + from simplejson._speedups import encode_basestring_ascii as c_encode_basestring_ascii
1.11 +except ImportError:
1.12 + pass
1.13 +
1.14 +ESCAPE = re.compile(r'[\x00-\x1f\\"\b\f\n\r\t]')
1.15 +ESCAPE_ASCII = re.compile(r'([\\"]|[^\ -~])')
1.16 +HAS_UTF8 = re.compile(r'[\x80-\xff]')
1.17 +ESCAPE_DCT = {
1.18 + '\\': '\\\\',
1.19 + '"': '\\"',
1.20 + '\b': '\\b',
1.21 + '\f': '\\f',
1.22 + '\n': '\\n',
1.23 + '\r': '\\r',
1.24 + '\t': '\\t',
1.25 +}
1.26 +for i in range(0x20):
1.27 + ESCAPE_DCT.setdefault(chr(i), '\\u%04x' % (i,))
1.28 +
1.29 +# Assume this produces an infinity on all machines (probably not guaranteed)
1.30 +INFINITY = float('1e66666')
1.31 +FLOAT_REPR = repr
1.32 +
1.33 +def floatstr(o, allow_nan=True):
1.34 + # Check for specials. Note that this type of test is processor- and/or
1.35 + # platform-specific, so do tests which don't depend on the internals.
1.36 +
1.37 + if o != o:
1.38 + text = 'NaN'
1.39 + elif o == INFINITY:
1.40 + text = 'Infinity'
1.41 + elif o == -INFINITY:
1.42 + text = '-Infinity'
1.43 + else:
1.44 + return FLOAT_REPR(o)
1.45 +
1.46 + if not allow_nan:
1.47 + raise ValueError("Out of range float values are not JSON compliant: %r"
1.48 + % (o,))
1.49 +
1.50 + return text
1.51 +
1.52 +
1.53 +def encode_basestring(s):
1.54 + """
1.55 + Return a JSON representation of a Python string
1.56 + """
1.57 + def replace(match):
1.58 + return ESCAPE_DCT[match.group(0)]
1.59 + return '"' + ESCAPE.sub(replace, s) + '"'
1.60 +
1.61 +
1.62 +def py_encode_basestring_ascii(s):
1.63 + if isinstance(s, str) and HAS_UTF8.search(s) is not None:
1.64 + s = s.decode('utf-8')
1.65 + def replace(match):
1.66 + s = match.group(0)
1.67 + try:
1.68 + return ESCAPE_DCT[s]
1.69 + except KeyError:
1.70 + n = ord(s)
1.71 + if n < 0x10000:
1.72 + return '\\u%04x' % (n,)
1.73 + else:
1.74 + # surrogate pair
1.75 + n -= 0x10000
1.76 + s1 = 0xd800 | ((n >> 10) & 0x3ff)
1.77 + s2 = 0xdc00 | (n & 0x3ff)
1.78 + return '\\u%04x\\u%04x' % (s1, s2)
1.79 + return '"' + str(ESCAPE_ASCII.sub(replace, s)) + '"'
1.80 +
1.81 +
1.82 +try:
1.83 + encode_basestring_ascii = c_encode_basestring_ascii
1.84 +except NameError:
1.85 + encode_basestring_ascii = py_encode_basestring_ascii
1.86 +
1.87 +
1.88 +class JSONEncoder(object):
1.89 + """
1.90 + Extensible JSON <http://json.org> encoder for Python data structures.
1.91 +
1.92 + Supports the following objects and types by default:
1.93 +
1.94 + +-------------------+---------------+
1.95 + | Python | JSON |
1.96 + +===================+===============+
1.97 + | dict | object |
1.98 + +-------------------+---------------+
1.99 + | list, tuple | array |
1.100 + +-------------------+---------------+
1.101 + | str, unicode | string |
1.102 + +-------------------+---------------+
1.103 + | int, long, float | number |
1.104 + +-------------------+---------------+
1.105 + | True | true |
1.106 + +-------------------+---------------+
1.107 + | False | false |
1.108 + +-------------------+---------------+
1.109 + | None | null |
1.110 + +-------------------+---------------+
1.111 +
1.112 + To extend this to recognize other objects, subclass and implement a
1.113 + ``.default()`` method with another method that returns a serializable
1.114 + object for ``o`` if possible, otherwise it should call the superclass
1.115 + implementation (to raise ``TypeError``).
1.116 + """
1.117 + __all__ = ['__init__', 'default', 'encode', 'iterencode']
1.118 + item_separator = ', '
1.119 + key_separator = ': '
1.120 + def __init__(self, skipkeys=False, ensure_ascii=True,
1.121 + check_circular=True, allow_nan=True, sort_keys=False,
1.122 + indent=None, separators=None, encoding='utf-8', default=None):
1.123 + """
1.124 + Constructor for JSONEncoder, with sensible defaults.
1.125 +
1.126 + If skipkeys is False, then it is a TypeError to attempt
1.127 + encoding of keys that are not str, int, long, float or None. If
1.128 + skipkeys is True, such items are simply skipped.
1.129 +
1.130 + If ensure_ascii is True, the output is guaranteed to be str
1.131 + objects with all incoming unicode characters escaped. If
1.132 + ensure_ascii is false, the output will be unicode object.
1.133 +
1.134 + If check_circular is True, then lists, dicts, and custom encoded
1.135 + objects will be checked for circular references during encoding to
1.136 + prevent an infinite recursion (which would cause an OverflowError).
1.137 + Otherwise, no such check takes place.
1.138 +
1.139 + If allow_nan is True, then NaN, Infinity, and -Infinity will be
1.140 + encoded as such. This behavior is not JSON specification compliant,
1.141 + but is consistent with most JavaScript based encoders and decoders.
1.142 + Otherwise, it will be a ValueError to encode such floats.
1.143 +
1.144 + If sort_keys is True, then the output of dictionaries will be
1.145 + sorted by key; this is useful for regression tests to ensure
1.146 + that JSON serializations can be compared on a day-to-day basis.
1.147 +
1.148 + If indent is a non-negative integer, then JSON array
1.149 + elements and object members will be pretty-printed with that
1.150 + indent level. An indent level of 0 will only insert newlines.
1.151 + None is the most compact representation.
1.152 +
1.153 + If specified, separators should be a (item_separator, key_separator)
1.154 + tuple. The default is (', ', ': '). To get the most compact JSON
1.155 + representation you should specify (',', ':') to eliminate whitespace.
1.156 +
1.157 + If specified, default is a function that gets called for objects
1.158 + that can't otherwise be serialized. It should return a JSON encodable
1.159 + version of the object or raise a ``TypeError``.
1.160 +
1.161 + If encoding is not None, then all input strings will be
1.162 + transformed into unicode using that encoding prior to JSON-encoding.
1.163 + The default is UTF-8.
1.164 + """
1.165 +
1.166 + self.skipkeys = skipkeys
1.167 + self.ensure_ascii = ensure_ascii
1.168 + self.check_circular = check_circular
1.169 + self.allow_nan = allow_nan
1.170 + self.sort_keys = sort_keys
1.171 + self.indent = indent
1.172 + self.current_indent_level = 0
1.173 + if separators is not None:
1.174 + self.item_separator, self.key_separator = separators
1.175 + if default is not None:
1.176 + self.default = default
1.177 + self.encoding = encoding
1.178 +
1.179 + def _newline_indent(self):
1.180 + return '\n' + (' ' * (self.indent * self.current_indent_level))
1.181 +
1.182 + def _iterencode_list(self, lst, markers=None):
1.183 + if not lst:
1.184 + yield '[]'
1.185 + return
1.186 + if markers is not None:
1.187 + markerid = id(lst)
1.188 + if markerid in markers:
1.189 + raise ValueError("Circular reference detected")
1.190 + markers[markerid] = lst
1.191 + yield '['
1.192 + if self.indent is not None:
1.193 + self.current_indent_level += 1
1.194 + newline_indent = self._newline_indent()
1.195 + separator = self.item_separator + newline_indent
1.196 + yield newline_indent
1.197 + else:
1.198 + newline_indent = None
1.199 + separator = self.item_separator
1.200 + first = True
1.201 + for value in lst:
1.202 + if first:
1.203 + first = False
1.204 + else:
1.205 + yield separator
1.206 + for chunk in self._iterencode(value, markers):
1.207 + yield chunk
1.208 + if newline_indent is not None:
1.209 + self.current_indent_level -= 1
1.210 + yield self._newline_indent()
1.211 + yield ']'
1.212 + if markers is not None:
1.213 + del markers[markerid]
1.214 +
1.215 + def _iterencode_dict(self, dct, markers=None):
1.216 + if not dct:
1.217 + yield '{}'
1.218 + return
1.219 + if markers is not None:
1.220 + markerid = id(dct)
1.221 + if markerid in markers:
1.222 + raise ValueError("Circular reference detected")
1.223 + markers[markerid] = dct
1.224 + yield '{'
1.225 + key_separator = self.key_separator
1.226 + if self.indent is not None:
1.227 + self.current_indent_level += 1
1.228 + newline_indent = self._newline_indent()
1.229 + item_separator = self.item_separator + newline_indent
1.230 + yield newline_indent
1.231 + else:
1.232 + newline_indent = None
1.233 + item_separator = self.item_separator
1.234 + first = True
1.235 + if self.ensure_ascii:
1.236 + encoder = encode_basestring_ascii
1.237 + else:
1.238 + encoder = encode_basestring
1.239 + allow_nan = self.allow_nan
1.240 + if self.sort_keys:
1.241 + keys = dct.keys()
1.242 + keys.sort()
1.243 + items = [(k, dct[k]) for k in keys]
1.244 + else:
1.245 + items = dct.iteritems()
1.246 + _encoding = self.encoding
1.247 + _do_decode = (_encoding is not None
1.248 + and not (_encoding == 'utf-8'))
1.249 + for key, value in items:
1.250 + if isinstance(key, str):
1.251 + if _do_decode:
1.252 + key = key.decode(_encoding)
1.253 + elif isinstance(key, basestring):
1.254 + pass
1.255 + # JavaScript is weakly typed for these, so it makes sense to
1.256 + # also allow them. Many encoders seem to do something like this.
1.257 + elif isinstance(key, float):
1.258 + key = floatstr(key, allow_nan)
1.259 + elif isinstance(key, (int, long)):
1.260 + key = str(key)
1.261 + elif key is True:
1.262 + key = 'true'
1.263 + elif key is False:
1.264 + key = 'false'
1.265 + elif key is None:
1.266 + key = 'null'
1.267 + elif self.skipkeys:
1.268 + continue
1.269 + else:
1.270 + raise TypeError("key %r is not a string" % (key,))
1.271 + if first:
1.272 + first = False
1.273 + else:
1.274 + yield item_separator
1.275 + yield encoder(key)
1.276 + yield key_separator
1.277 + for chunk in self._iterencode(value, markers):
1.278 + yield chunk
1.279 + if newline_indent is not None:
1.280 + self.current_indent_level -= 1
1.281 + yield self._newline_indent()
1.282 + yield '}'
1.283 + if markers is not None:
1.284 + del markers[markerid]
1.285 +
1.286 + def _iterencode(self, o, markers=None):
1.287 + if isinstance(o, basestring):
1.288 + if self.ensure_ascii:
1.289 + encoder = encode_basestring_ascii
1.290 + else:
1.291 + encoder = encode_basestring
1.292 + _encoding = self.encoding
1.293 + if (_encoding is not None and isinstance(o, str)
1.294 + and not (_encoding == 'utf-8')):
1.295 + o = o.decode(_encoding)
1.296 + yield encoder(o)
1.297 + elif o is None:
1.298 + yield 'null'
1.299 + elif o is True:
1.300 + yield 'true'
1.301 + elif o is False:
1.302 + yield 'false'
1.303 + elif isinstance(o, (int, long)):
1.304 + yield str(o)
1.305 + elif isinstance(o, float):
1.306 + yield floatstr(o, self.allow_nan)
1.307 + elif isinstance(o, (list, tuple)):
1.308 + for chunk in self._iterencode_list(o, markers):
1.309 + yield chunk
1.310 + elif isinstance(o, dict):
1.311 + for chunk in self._iterencode_dict(o, markers):
1.312 + yield chunk
1.313 + else:
1.314 + if markers is not None:
1.315 + markerid = id(o)
1.316 + if markerid in markers:
1.317 + raise ValueError("Circular reference detected")
1.318 + markers[markerid] = o
1.319 + for chunk in self._iterencode_default(o, markers):
1.320 + yield chunk
1.321 + if markers is not None:
1.322 + del markers[markerid]
1.323 +
1.324 + def _iterencode_default(self, o, markers=None):
1.325 + newobj = self.default(o)
1.326 + return self._iterencode(newobj, markers)
1.327 +
1.328 + def default(self, o):
1.329 + """
1.330 + Implement this method in a subclass such that it returns
1.331 + a serializable object for ``o``, or calls the base implementation
1.332 + (to raise a ``TypeError``).
1.333 +
1.334 + For example, to support arbitrary iterators, you could
1.335 + implement default like this::
1.336 +
1.337 + def default(self, o):
1.338 + try:
1.339 + iterable = iter(o)
1.340 + except TypeError:
1.341 + pass
1.342 + else:
1.343 + return list(iterable)
1.344 + return JSONEncoder.default(self, o)
1.345 + """
1.346 + raise TypeError("%r is not JSON serializable" % (o,))
1.347 +
1.348 + def encode(self, o):
1.349 + """
1.350 + Return a JSON string representation of a Python data structure.
1.351 +
1.352 + >>> JSONEncoder().encode({"foo": ["bar", "baz"]})
1.353 + '{"foo": ["bar", "baz"]}'
1.354 + """
1.355 + # This is for extremely simple cases and benchmarks.
1.356 + if isinstance(o, basestring):
1.357 + if isinstance(o, str):
1.358 + _encoding = self.encoding
1.359 + if (_encoding is not None
1.360 + and not (_encoding == 'utf-8')):
1.361 + o = o.decode(_encoding)
1.362 + if self.ensure_ascii:
1.363 + return encode_basestring_ascii(o)
1.364 + else:
1.365 + return encode_basestring(o)
1.366 + # This doesn't pass the iterator directly to ''.join() because the
1.367 + # exceptions aren't as detailed. The list call should be roughly
1.368 + # equivalent to the PySequence_Fast that ''.join() would do.
1.369 + chunks = list(self.iterencode(o))
1.370 + return ''.join(chunks)
1.371 +
1.372 + def iterencode(self, o):
1.373 + """
1.374 + Encode the given object and yield each string
1.375 + representation as available.
1.376 +
1.377 + For example::
1.378 +
1.379 + for chunk in JSONEncoder().iterencode(bigobject):
1.380 + mysocket.write(chunk)
1.381 + """
1.382 + if self.check_circular:
1.383 + markers = {}
1.384 + else:
1.385 + markers = None
1.386 + return self._iterencode(o, markers)
1.387 +
1.388 +__all__ = ['JSONEncoder']
