feat: add string_json value type for

sdk/python/feast/on_demand_feature_view.py

@@ -680,6 +680,7 @@ def _construct_random_input(
             ValueType.FLOAT_LIST: [[1.0]],
             ValueType.BOOL_LIST: [[True]],
             ValueType.UNIX_TIMESTAMP_LIST: [[_utc_now()]],
+            ValueType.STRING_JSON: [{"0": "hello world"}],
         }
         if singleton:
             rand_dict_value = {k: rand_dict_value[k][0] for k in rand_dict_value}

sdk/python/feast/types.py

@@ -183,6 +183,7 @@ def __str__(self):
     ValueType.FLOAT_LIST: Array(Float32),
     ValueType.BOOL_LIST: Array(Bool),
     ValueType.UNIX_TIMESTAMP_LIST: Array(UnixTimestamp),
+    ValueType.STRING_JSON: String
 }
 
 FEAST_TYPES_TO_PYARROW_TYPES = {

sdk/python/feast/value_type.py

@@ -48,6 +48,7 @@ class ValueType(enum.Enum):
     BOOL_LIST = 17
     UNIX_TIMESTAMP_LIST = 18
     NULL = 19
+    STRING_JSON = 20
 
 
 ListType = Union[

sdk/python/tests/unit/test_on_demand_feature_view.py

@@ -26,6 +26,7 @@
     PythonTransformation,
 )
 from feast.types import Float32
+from feast.value_type import ValueType
 
 
 def udf1(features_df: pd.DataFrame) -> pd.DataFrame:
@@ -356,3 +357,55 @@ def test_on_demand_feature_view_stored_writes():
     assert transformed_output["output3"] is not None and isinstance(
         transformed_output["output3"], datetime.datetime
     )
+
+
+def test_on_demand_feature_view_string_json():
+    file_source = FileSource(name="my-file-source", path="test.parquet")
+    feature_view = FeatureView(
+        name="driver_hourly_stats_view",
+        entities=[],
+        schema=[Field(name="feature1", dtype=ValueType.STRING_JSON)],
+        source=file_source
+        )
+
+    sources = [feature_view]
+
+    def udf_string_json(inputs: pd.DataFrame) -> pd.DataFrame:
+        df = pd.DataFrame()
+
+        df['output1'] = 0.0
+        for index, row in inputs.iterrows():
+            value_array = eval(row["feature1"])
+            x1 = value_array.get("1", 0.0) 
+            x1 = x1 if x1 is not None else 0.0
+            x2 = value_array.get("2", 0.0)
+            x2 = x2 if x2 is not None else 0.0
+            df.at[index, 'output1'] = x1 + x2
+        return df
+
+
+    on_demand_feature_view = OnDemandFeatureView(
+        name="my-on-demand-feature-view",
+        sources=sources,
+        schema=[
+            Field(name="output1", dtype=Float32),
+        ],
+        feature_transformation=PandasTransformation(
+            udf=udf_string_json, udf_string="udf_string_json source code"
+        ),
+    )
+
+    transformed_output = on_demand_feature_view.transform_dict(
+        {
+            "feature1": '{"1": 1.0, "2": 2.0, "3": 3.0}'
+        }
+    )
+    expected_output = {"feature1": '{"1": 1.0, "2": 2.0, "3": 3.0}', "output1": 3.0}
+    keys_to_validate = [
+        "feature1",
+        "output1",
+    ]
+    for k in keys_to_validate:
+        assert transformed_output[k] == expected_output[k]
+
+