Fixes date and frequency issues in forecasting

docs/source/reference/ai/model-forecasting.rst

@@ -48,7 +48,7 @@ EvaDB's default forecast framework is `statsforecast <https://nixtla.github.io/s
    * - PREDICT (required) 
      - The name of the column we wish to forecast.
    * - TIME
-     - The name of the column that contains the datestamp, wihch should be of a format expected by Pandas, ideally YYYY-MM-DD for a date or YYYY-MM-DD HH:MM:SS for a timestamp. If not provided, an auto increasing ID column will be used.
+     - The name of the column that contains the datestamp, wihch should be of a format expected by Pandas, ideally YYYY-MM-DD for a date or YYYY-MM-DD HH:MM:SS for a timestamp. Please visit the `pandas documentation <https://pandas.pydata.org/docs/reference/api/pandas.to_datetime.html>`_ for details. If not provided, an auto increasing ID column will be used.
    * - ID
      - The name of column that represents an identifier for the series. If not provided, the whole table is considered as one series of data.
    * - MODEL

evadb/executor/create_function_executor.py

@@ -149,15 +149,11 @@ def handle_forecasting_function(self):
             impl_path = Path(f"{self.function_dir}/forecast.py").absolute().as_posix()
         else:
             impl_path = self.node.impl_path.absolute().as_posix()
-        arg_map = {arg.key: arg.value for arg in self.node.metadata}
 
         if "model" not in arg_map.keys():
             arg_map["model"] = "AutoARIMA"
-        if "frequency" not in arg_map.keys():
-            arg_map["frequency"] = "M"
 
         model_name = arg_map["model"]
-        frequency = arg_map["frequency"]
 
         """
         The following rename is needed for statsforecast, which requires the column name to be the following:
@@ -179,6 +175,10 @@ def handle_forecasting_function(self):
         if "ds" not in list(data.columns):
             data["ds"] = [x + 1 for x in range(len(data))]
 
+        if "frequency" not in arg_map.keys():
+            arg_map["frequency"] = pd.infer_freq(data["ds"])
+        frequency = arg_map["frequency"]
+
         try_to_import_forecast()
         from statsforecast import StatsForecast
         from statsforecast.models import AutoARIMA, AutoCES, AutoETS, AutoTheta
@@ -220,7 +220,7 @@ def handle_forecasting_function(self):
         )
 
         weight_file = Path(model_path)
-
+        data["ds"] = pd.to_datetime(data["ds"])
         if not weight_file.exists():
             model.fit(data)
             f = open(model_path, "wb")
@@ -233,6 +233,9 @@ def handle_forecasting_function(self):
             FunctionMetadataCatalogEntry("model_name", model_name),
             FunctionMetadataCatalogEntry("model_path", model_path),
             FunctionMetadataCatalogEntry("output_column_rename", arg_map["predict"]),
+            FunctionMetadataCatalogEntry(
+                "time_column_rename", arg_map["time"] if "time" in arg_map else "ds"
+            ),
         ]
 
         return (

evadb/functions/forecast.py

@@ -28,13 +28,20 @@ def name(self) -> str:
         return "ForecastModel"
 
     @setup(cacheable=False, function_type="Forecasting", batchable=True)
-    def setup(self, model_name: str, model_path: str, output_column_rename: str):
+    def setup(
+        self,
+        model_name: str,
+        model_path: str,
+        output_column_rename: str,
+        time_column_rename: str,
+    ):
         f = open(model_path, "rb")
         loaded_model = pickle.load(f)
         f.close()
         self.model = loaded_model
         self.model_name = model_name
         self.output_column_rename = output_column_rename
+        self.time_column_rename = time_column_rename
 
     def forward(self, data) -> pd.DataFrame:
         horizon = list(data.iloc[:, -1])[0]
@@ -43,6 +50,9 @@ def forward(self, data) -> pd.DataFrame:
         ), "Forecast UDF expects integral horizon in parameter."
         forecast_df = self.model.predict(h=horizon)
         forecast_df = forecast_df.rename(
-            columns={self.model_name: self.output_column_rename}
+            columns={
+                self.model_name: self.output_column_rename,
+                "ds": self.time_column_rename,
+            }
         )
         return forecast_df