fixing the response when there is no data found getting the station data

src/common/pg_impl.py

@@ -137,45 +137,49 @@ def get_station_data(self, **kwargs) -> str:
         # get forecast data
         forecast_data = self.get_forecast_station_data(kwargs['station_name'], kwargs['time_mark'], kwargs['data_source'])
 
-        # derive start date from the time mark
-        start_date = (datetime.fromisoformat(kwargs['time_mark']) - timedelta(4)).isoformat()
+        # check for an error
+        if forecast_data.empty:
+            return ''
+        else:
+            # derive start date from the time mark
+            start_date = (datetime.fromisoformat(kwargs['time_mark']) - timedelta(4)).isoformat()
 
-        # get end_date from last datetime in forecast data
-        end_date = forecast_data['time_stamp'].iloc[-1]
+            # get end_date from last datetime in forecast data
+            end_date = forecast_data['time_stamp'].iloc[-1]
 
-        # get nowcast data_source from forecast data_source
-        nowcast_source = 'NOWCAST_' + "_".join(kwargs['data_source'].split('_')[1:])
+            # get nowcast data_source from forecast data_source
+            nowcast_source = 'NOWCAST_' + "_".join(kwargs['data_source'].split('_')[1:])
 
-        # get obs and nowcast data
-        obs_data = self.get_obs_station_data(kwargs['station_name'], start_date, end_date, nowcast_source)
+            # get obs and nowcast data
+            obs_data = self.get_obs_station_data(kwargs['station_name'], start_date, end_date, nowcast_source)
 
-        # drop empty columns
-        empty_cols = [col for col in obs_data.columns if obs_data[col].isnull().all()]
-        obs_data.drop(empty_cols, axis=1, inplace=True)
+            # drop empty columns
+            empty_cols = [col for col in obs_data.columns if obs_data[col].isnull().all()]
+            obs_data.drop(empty_cols, axis=1, inplace=True)
 
-        # replace any None values with np.nan, in both DataFrames
-        forecast_data.fillna(value=np.nan)
-        obs_data.fillna(value=np.nan)
+            # replace any None values with np.nan, in both DataFrames
+            forecast_data.fillna(value=np.nan)
+            obs_data.fillna(value=np.nan)
 
-        # convert all values after the time mark to nan, in obs data, except in the time_stamp and tidal_predictions columns
-        for col in obs_data.columns:
-            if col not in ('time_stamp', 'tidal_predictions'):
-                obs_data.loc[obs_data.time_stamp >= kwargs['time_mark'], col] = np.nan
-            else:
-                continue
+            # convert all values after the time mark to nan, in obs data, except in the time_stamp and tidal_predictions columns
+            for col in obs_data.columns:
+                if col not in ('time_stamp', 'tidal_predictions'):
+                    obs_data.loc[obs_data.time_stamp >= kwargs['time_mark'], col] = np.nan
+                else:
+                    continue
 
-        # merge the obs DataFrame with the forecast Dataframe
-        station_df = obs_data.merge(forecast_data, on='time_stamp', how='outer')
+            # merge the obs DataFrame with the forecast Dataframe
+            station_df = obs_data.merge(forecast_data, on='time_stamp', how='outer')
 
-        # get the forecast and nowcast column names
-        forecast_column_name = "".join(kwargs['data_source'].split('.')).lower()
-        nowcast_column_name = "".join(nowcast_source.split('.')).lower()
+            # get the forecast and nowcast column names
+            forecast_column_name = "".join(kwargs['data_source'].split('.')).lower()
+            nowcast_column_name = "".join(nowcast_source.split('.')).lower()
 
-        # rename the columns
-        station_df.rename(columns={forecast_column_name: 'forecast_water_level', nowcast_column_name: 'nowcast_water_level'}, inplace=True)
+            # rename the columns
+            station_df.rename(columns={forecast_column_name: 'forecast_water_level', nowcast_column_name: 'nowcast_water_level'}, inplace=True)
 
-        # return the data to the caller
-        return station_df.to_csv(index=False)
+            # return the data to the caller
+            return station_df.to_csv(index=False)
 
     def get_forecast_station_data(self, station_name, time_mark, data_source) -> pd.DataFrame:
         """
@@ -200,6 +204,8 @@ def get_forecast_station_data(self, station_name, time_mark, data_source) -> pd.
         if station_data != -1:
             # convert query output to Pandas dataframe
             ret_val = pd.DataFrame.from_dict(station_data, orient='columns')
+        else:
+            ret_val = pd.DataFrame(None)
 
         # Return Pandas dataframe
         return ret_val