Support Python's Negative Exponent Padding Idiosyncrasy

README.md

@@ -78,6 +78,19 @@ Controls whether indentation ("pretty output") is enabled. Default is `0` (disab
 }
 ```
 
+#### zero_pad_negative_9_to_5_exponent
+
+If true, adds a single `0` padding to the exponent in scientific notation if it
+is between `-9` and `-5` both inclusive, which replicates the Python standard
+library's `json` behavior.  Default is `False`:
+
+```pycon
+>>> ujson.dumps([1e-10, 1e-9, 1e-5, 1e-4])
+'[1e-10,1e-9,1e-5,0.0001]'
+>>> ujson.dumps([1e-10, 1e-9, 1e-5, 1e-4], zero_pad_negative_9_to_5_exponent=True)
+'[1e-10,1e-09,1e-05,0.0001]'
+```
+
 ## Benchmarks
 
 *UltraJSON* calls/sec compared to other popular JSON parsers with performance gain

lib/dconv_wrapper.cc

@@ -12,12 +12,13 @@ namespace double_conversion
                         int decimal_in_shortest_low,
                         int decimal_in_shortest_high,
                         int max_leading_padding_zeroes_in_precision_mode,
-                        int max_trailing_padding_zeroes_in_precision_mode)
+                        int max_trailing_padding_zeroes_in_precision_mode,
+                        int min_exponent_width)
     {
         *d2s = new DoubleToStringConverter(flags, infinity_symbol, nan_symbol,
                                         exponent_character, decimal_in_shortest_low,
                                         decimal_in_shortest_high, max_leading_padding_zeroes_in_precision_mode,
-                                        max_trailing_padding_zeroes_in_precision_mode);
+                                        max_trailing_padding_zeroes_in_precision_mode, min_exponent_width);
     }
 
     int dconv_d2s(void *d2s, double value, char* buf, int buflen, int* strlength)

lib/ultrajson.h

@@ -268,6 +268,12 @@ typedef struct __JSONObjectEncoder
   If true, bytes are rejected. */
   int rejectBytes;
 
+  /*
+  If true, adds a single 0 padding to the exponent in scientific notation if it
+  is between -9 and -5 both inclusive, which replicates the Python standard
+  library's JSON behavior, e.g. 1e-5 will become 1e-05. */
+  int zeroPadNegative9to5Exponent;
+
   /*
   Configuration for item and key separators, e.g. "," and ":" for a compact representation or ", " and ": " to match the Python standard library's defaults. */
   size_t itemSeparatorLength;
@@ -382,7 +388,8 @@ void dconv_d2s_init(void **d2s,
                     int decimal_in_shortest_low,
                     int decimal_in_shortest_high,
                     int max_leading_padding_zeroes_in_precision_mode,
-                    int max_trailing_padding_zeroes_in_precision_mode);
+                    int max_trailing_padding_zeroes_in_precision_mode,
+                    int min_exponent_width);
 int dconv_d2s(void *d2s, double value, char* buf, int buflen, int* strlength);
 void dconv_d2s_free(void **d2s);
 

python/objToJSON.c

@@ -659,7 +659,7 @@ static char *Object_iterGetName(JSOBJ obj, JSONTypeContext *tc, size_t *outLen)
 
 PyObject* objToJSON(PyObject* self, PyObject *args, PyObject *kwargs)
 {
-  static char *kwlist[] = { "obj", "ensure_ascii", "encode_html_chars", "escape_forward_slashes", "sort_keys", "indent", "allow_nan", "reject_bytes", "default", "separators", NULL };
+  static char *kwlist[] = { "obj", "ensure_ascii", "encode_html_chars", "escape_forward_slashes", "sort_keys", "indent", "allow_nan", "reject_bytes", "default", "separators", "zero_pad_negative_9_to_5_exponent", NULL };
 
   char buffer[65536];
   char *ret;
@@ -676,9 +676,11 @@ PyObject* objToJSON(PyObject* self, PyObject *args, PyObject *kwargs)
   PyObject *separatorsItemBytes = NULL;
   PyObject *oseparatorsKey = NULL;
   PyObject *separatorsKeyBytes = NULL;
+  PyObject *ozero_pad_negative_9_to_5_exponent = NULL;
   int allowNan = -1;
   int orejectBytes = -1;
   size_t retLen;
+  int minExponentWidth = 0;
 
   JSONObjectEncoder encoder =
   {
@@ -704,6 +706,7 @@ PyObject* objToJSON(PyObject* self, PyObject *args, PyObject *kwargs)
     0, //indent
     1, //allowNan
     1, //rejectBytes
+    0, //zeroPadNegative9to5Exponent
     0, //itemSeparatorLength
     NULL, //itemSeparatorChars
     0, //keySeparatorLength
@@ -714,7 +717,7 @@ PyObject* objToJSON(PyObject* self, PyObject *args, PyObject *kwargs)
 
   PRINTMARK();
 
-  if (!PyArg_ParseTupleAndKeywords(args, kwargs, "O|OOOOiiiOO", kwlist, &oinput, &oensureAscii, &oencodeHTMLChars, &oescapeForwardSlashes, &osortKeys, &encoder.indent, &allowNan, &orejectBytes, &odefaultFn, &oseparators))
+  if (!PyArg_ParseTupleAndKeywords(args, kwargs, "O|OOOOiiiOOO", kwlist, &oinput, &oensureAscii, &oencodeHTMLChars, &oescapeForwardSlashes, &osortKeys, &encoder.indent, &allowNan, &orejectBytes, &odefaultFn, &oseparators, &ozero_pad_negative_9_to_5_exponent))
   {
     return NULL;
   }
@@ -816,9 +819,14 @@ PyObject* objToJSON(PyObject* self, PyObject *args, PyObject *kwargs)
     }
   }
 
+  if (ozero_pad_negative_9_to_5_exponent != NULL && PyObject_IsTrue(ozero_pad_negative_9_to_5_exponent))
+  {
+    minExponentWidth = 2;
+  }
+
   encoder.d2s = NULL;
   dconv_d2s_init(&encoder.d2s, DCONV_D2S_EMIT_TRAILING_DECIMAL_POINT | DCONV_D2S_EMIT_TRAILING_ZERO_AFTER_POINT | DCONV_D2S_EMIT_POSITIVE_EXPONENT_SIGN,
-                 csInf, csNan, 'e', DCONV_DECIMAL_IN_SHORTEST_LOW, DCONV_DECIMAL_IN_SHORTEST_HIGH, 0, 0);
+                 csInf, csNan, 'e', DCONV_DECIMAL_IN_SHORTEST_LOW, DCONV_DECIMAL_IN_SHORTEST_HIGH, 0, 0, minExponentWidth);
 
   PRINTMARK();
   ret = JSON_EncodeObject (oinput, &encoder, buffer, sizeof (buffer), &retLen);

python/ujson.c

@@ -59,7 +59,8 @@ PyObject* JSONDecodeError;
     "Set encode_html_chars=True to encode < > & as unicode escape sequences. "\
     "Set escape_forward_slashes=False to prevent escaping / characters." \
     "Set allow_nan=False to raise an exception when NaN or Infinity would be serialized." \
-    "Set reject_bytes=True to raise TypeError on bytes."
+    "Set reject_bytes=True to raise TypeError on bytes." \
+    "Set zero_pad_negative_9_to_5_exponent=True to add 0-pad for exponents -9 to -5."
 
 static PyMethodDef ujsonMethods[] = {
   {"encode", (PyCFunction) objToJSON, METH_VARARGS | METH_KEYWORDS, "Converts arbitrary object recursively into JSON. " ENCODER_HELP_TEXT},

tests/fuzz.py

@@ -133,6 +133,13 @@ def __call__(self, parser, namespace, values, option_string=None):
     help="Sets the escape_forward_slashes option to ujson.dumps(). "
     "May be 0 or 1 or 0,1 to test both.",
 )
+parser.add_argument(
+    "--zero_pad_negative_9_to_5_exponent",
+    default=(0, 1),
+    action=ListOption,
+    help="Sets the zero_pad_negative_9_to_5_exponent option to ujson.dumps(). "
+    "May be 0 or 1 or 0,1 to test both.",
+)
 parser.add_argument(
     "--dump-python",
     action="store_true",

tests/test_ujson.py

@@ -84,12 +84,19 @@ def test_double_long_decimal_issue():
     assert sut == decoded
 
 
-# NOTE: can't match exponents -9 to -5; Python 0-pads
+# NOTE: The default behaviour can't match exponents -9 to -5; Python 0-pads
 @pytest.mark.parametrize("val", [1e-10, 1e-4, 1e10, 1e15, 1e16, 1e30])
 def test_encode_float_string_rep(val):
     assert ujson.dumps(val) == json.dumps(val)
 
 
+@pytest.mark.parametrize(
+    "val", [1e-10, 1e-9, 1e-8, 1e-6, 1e-5, 1e-4, 1e10, 1e15, 1e16, 1e30]
+)
+def test_encode_float_string_replicate_python(val):
+    assert ujson.dumps(val, zero_pad_negative_9_to_5_exponent=True) == json.dumps(val)
+
+
 def test_encode_decode_long_decimal():
     sut = {"a": -528656961.4399388}
     encoded = ujson.dumps(sut)