failproof

Author: michielkallenberg

cybench/datasets/configured.py

@@ -134,14 +134,17 @@ def load_dfs(
     dfs_x = {"soil": df_x_soil}
 
     # location
-    df_x_location = pd.read_csv(
-        os.path.join(path_data_cn, "_".join(["location", crop, country_code]) + ".csv"),
-        header=0,
-    )
-    df_x_location = df_x_location[[KEY_LOC] + LOCATION_PROPERTIES]
-    df_x_location.set_index([KEY_LOC], inplace=True)
+    if LOCATION_PROPERTIES:
+        df_x_location = pd.read_csv(
+            os.path.join(
+                path_data_cn, "_".join(["location", crop, country_code]) + ".csv"
+            ),
+            header=0,
+        )
+        df_x_location = df_x_location[[KEY_LOC] + LOCATION_PROPERTIES]
+        df_x_location.set_index([KEY_LOC], inplace=True)
 
-    dfs_x["location"] = df_x_location
+        dfs_x["location"] = df_x_location
 
     # crop calendar
     df_crop_cal = pd.read_csv(

cybench/models/sklearn_models.py

@@ -197,12 +197,14 @@ def _design_features(self, crop: str, data_items: Iterable):
         soil_df = data_to_pandas(data_items, data_cols=[KEY_LOC] + SOIL_PROPERTIES)
         soil_df = soil_df.drop_duplicates()
 
-        location_df = data_to_pandas(
-            data_items, data_cols=[KEY_LOC] + LOCATION_PROPERTIES
-        )
-        location_df = location_df.drop_duplicates()
+        dfs_x = {"soil": soil_df}
 
-        dfs_x = {"soil": soil_df, "location": location_df}
+        if LOCATION_PROPERTIES:
+            location_df = data_to_pandas(
+                data_items, data_cols=[KEY_LOC] + LOCATION_PROPERTIES
+            )
+            location_df = location_df.drop_duplicates()
+            dfs_x["location"] = location_df
 
         for x, ts_cols in TIME_SERIES_INPUTS.items():
             df_ts = data_to_pandas(