Add support for preprocessed JSON strings (with optimizations) in encoder

In some situations, you may have a large python dictionary you need to JSONify but one of the values inside the dict is already a JSON string. This is common when pulling an object from a database, for example, where one of the fields is a JSON blob/string. Previously you would have to deserialize and then reserialize that string just to serialize the high level object, but obviously this is unnecessarily slow. This changes adds a method/type that can be used to wrap a str and tell the serializer to just pass it through instead.

4 files changed, 63 insertions, 1 deletions

diff --git a/simplejson/__init__.py b/simplejson/__init__.py
index b7fe828..eb8f295 100644
--- a/simplejson/__init__.py
+++ b/simplejson/__init__.py
@@ -110,7 +110,7 @@ from decimal import Decimal
 
 from .scanner import JSONDecodeError
 from .decoder import JSONDecoder
-from .encoder import JSONEncoder, JSONEncoderForHTML
+from .encoder import JSONEncoder, JSONEncoderForHTML, RawJSON
 def _import_OrderedDict():
     import collections
     try:
diff --git a/simplejson/_speedups.c b/simplejson/_speedups.c
index 9976464..2d81063 100644
--- a/simplejson/_speedups.c
+++ b/simplejson/_speedups.c
@@ -245,6 +245,8 @@ static void
 encoder_dealloc(PyObject *self);
 static int
 encoder_clear(PyObject *self);
+static int
+is_raw_json(PyObject *obj);
 static PyObject *
 encoder_stringify_key(PyEncoderObject *s, PyObject *key);
 static int
@@ -277,6 +279,20 @@ moduleinit(void);
 
 #define MIN_EXPANSION 6
 
+static PyObject* RawJSONType;
+static int
+is_raw_json(PyObject *obj)
+{
+    if (RawJSONType == NULL) {
+        PyObject *encoder_module = PyImport_ImportModule("simplejson.encoder");
+        RawJSONType = PyObject_GetAttrString(encoder_module, "RawJSON");
+        Py_DECREF(encoder_module);
+        if (RawJSONType == NULL)
+            return 0;
+    }
+    return PyObject_IsInstance(obj, RawJSONType) ? 1 : 0;
+}
+
 static int
 JSON_Accu_Init(JSON_Accu *acc)
 {
@@ -2868,6 +2884,12 @@ encoder_listencode_obj(PyEncoderObject *s, JSON_Accu *rval, PyObject *obj, Py_ss
             if (encoded != NULL)
                 rv = _steal_accumulate(rval, encoded);
         }
+        else if (is_raw_json(obj))
+        {
+            PyObject *encoded = PyObject_GetAttrString(obj, "encoded_json");
+            if (encoded != NULL)
+                rv = _steal_accumulate(rval, encoded);
+        }
         else if (PyInt_Check(obj) || PyLong_Check(obj)) {
             PyObject *encoded;
             if (PyInt_CheckExact(obj) || PyLong_CheckExact(obj)) {
diff --git a/simplejson/encoder.py b/simplejson/encoder.py
index 5b9bda7..9a4c0c5 100644
--- a/simplejson/encoder.py
+++ b/simplejson/encoder.py
@@ -39,6 +39,10 @@ for i in [0x2028, 0x2029]:
 
 FLOAT_REPR = repr
 
+class RawJSON(object):
+    def __init__(self, encoded_json):
+        self.encoded_json = encoded_json
+
 def encode_basestring(s, _PY3=PY3, _q=u('"')):
     """Return a JSON representation of a Python string
 
@@ -464,6 +468,8 @@ def _make_iterencode(markers, _default, _encoder, _indent, _floatstr,
             if (isinstance(value, string_types) or
                 (_PY3 and isinstance(value, binary_type))):
                 yield buf + _encoder(value)
+            elif isinstance(value, RawJSON):
+                yield buf + value.encoded_json
             elif value is None:
                 yield buf + 'null'
             elif value is True:
@@ -582,6 +588,8 @@ def _make_iterencode(markers, _default, _encoder, _indent, _floatstr,
             if (isinstance(value, string_types) or
                 (_PY3 and isinstance(value, binary_type))):
                 yield _encoder(value)
+            elif isinstance(value, RawJSON):
+                yield value.encoded_json
             elif value is None:
                 yield 'null'
             elif value is True:
@@ -624,6 +632,8 @@ def _make_iterencode(markers, _default, _encoder, _indent, _floatstr,
         if (isinstance(o, string_types) or
             (_PY3 and isinstance(o, binary_type))):
             yield _encoder(o)
+        elif isinstance(o, RawJSON):
+            buf + o.encoded_json
         elif o is None:
             yield 'null'
         elif o is True:
diff --git a/simplejson/tests/test_raw_json.py b/simplejson/tests/test_raw_json.py
new file mode 100644
index 0000000..6fa6349
--- /dev/null
+++ b/simplejson/tests/test_raw_json.py
@@ -0,0 +1,30 @@
+import unittest
+import simplejson as json
+
+dct1 = {
+    'key1': 'value1'
+}
+
+dct2 = {
+    'key2': 'value2',
+    'd1': dct1
+}
+
+dct3 = {
+    'key2': 'value2',
+    'd1': json.dumps(dct1)
+}
+
+dct4 = {
+    'key2': 'value2',
+    'd1': json.RawJSON(json.dumps(dct1))
+}
+
+
+class TestRawJson(unittest.TestCase):
+
+    def test_normal_str(self):
+        self.assertNotEqual(json.dumps(dct2), json.dumps(dct3))
+
+    def test_raw_json_str(self):
+        self.assertEqual(json.dumps(dct2), json.dumps(dct4))