finish base gpr

src/conf/cs/models/gpr.yaml

@@ -13,7 +13,3 @@ gpr:
       - 3
     kernel:
       - DotProduct + WhiteKernel
-      - RBF
-      - Matern
-      - RationalQuadratic
-      - Exponential  

src/permutation/exporter.py

@@ -49,6 +49,29 @@ def pandas_to_csv(
         dir_path = f"{self.export_path}/{self.experiment}/{model}/{stage.name}"
         self._save_df(dir_path, filename, dataframe)
 
+    def make_json_serializable(self, data):
+        """
+        Recursively checks if all elements in a dictionary are JSON-serializable.
+        If not, converts non-serializable elements to strings.
+        
+        Parameters:
+        - data (dict): Dictionary of parameters to check.
+
+        Returns:
+        - dict: JSON-serializable dictionary.
+        """
+        if isinstance(data, dict):
+            return {key: self.make_json_serializable(value) for key, value in data.items()}
+        elif isinstance(data, list):
+            return [self.make_json_serializable(item) for item in data]
+        try:
+            # Attempt to serialize to JSON
+            json.dumps(data)
+            return data
+        except (TypeError, OverflowError):
+            # If not serializable, convert to string
+            return str(data)
+
     def _save_df(self, dir_path: str, name: str, df: pd.DataFrame | pd.Series) -> None:
         """Validates save path and then saves a Dataframe as a CSV file"""
         validate_dir(dir_path)
@@ -67,7 +90,8 @@ def save_model_json(self, runner: Runner) -> None:
         )
         with open(dir_path, "w", encoding="UTF-8") as outfile:
             if runner.model.hparams:
-                json.dump(runner.model.hparams.as_dict(), outfile)
+                serialized_hp = self.make_json_serializable(runner.model.hparams.as_dict())
+                json.dump(serialized_hp, outfile)
 
     def save_predictions(self, model: str, runner: Runner) -> None:
         """Saves the model predictions as a CSV file"""

src/permutation/models/gpr.py

@@ -71,9 +71,7 @@ def set_model(
             kernel_name = hparams_dict.get("kernel")
             kernel = map_kernel(kernel_name)
             hparams_dict['kernel'] = kernel
-            print(hparams_dict)
-            print(hparams)
-            hparams = HParams(**hparams_dict)
+            hparams = HParams(param_dict=hparams_dict)
         else:
             hparams_dict = {'kernel': kernel} if kernel else {}
 

src/permutation/models/sklearnmodel.py

@@ -51,10 +51,6 @@ def _set_model(
         """
         model = cls()
         model.hparams = hparams
-        print(model.hparams)
-        print(model)
-        print(model_dependency)
-        #asd()
         pipeline_list = []
 
         if preprocessing_dependencies:
@@ -70,7 +66,6 @@ def _set_model(
             model.model = model_dependency()
 
         pipeline_list.append((model.algorithm_name, model.model))  # pylint: disable=no-member
-
         model.pipeline = Pipeline(pipeline_list)
         return model