add data validation and error logging

python/calib/pyproject.toml

@@ -1,3 +1,6 @@
 [build-system]
 requires = ["setuptools>=64", "wheel"]
 build-backend = "setuptools.build_meta"
+
+[tool.ruff]
+exclude = ["src/"]

python/calib/src/calib/calibration_set.py

@@ -75,6 +75,13 @@ def __init__(
             if os.path.exists(obsflow_file):
                 logger.info(f"Read observed streamflow from: {obsflow_file}")
                 obs = pd.read_csv(obsflow_file)
+
+                # if obs is empty, raise an error
+                if obs.empty:
+                    msg = f"Streamflow observation file is empty: {obsflow_file}"
+                    logger.error(msg)
+                    raise ValueError(msg)
+
                 cols = obs.columns.str.lower()
                 obs.columns = cols
 
@@ -217,8 +224,10 @@ def output(self) -> "DataFrame":
             except Exception as e:
                 raise e
 
-        if hydrograph is None:
-            logger.info("Output could not be read after multiple attempts.")
+        if hydrograph is None or hydrograph.empty:
+            msg = "Simulated hydrograph is unavailable or empty."
+            logger.error(msg)
+            raise ValueError(msg)
 
         return hydrograph
 

python/calib/src/calib/configuration.py

@@ -246,6 +246,11 @@ def get_sim_df(self, _output_file, _wb_lst) -> "DataFrame":
         except Exception as e:
             raise (e)
 
+        if hydrograph is None or hydrograph.empty:
+            msg = "Simulated hydrograph is unavailable or empty."
+            logger.error(msg)
+            raise ValueError(msg)
+
         return hydrograph
 
     def postprocess_single_calibration_output(self, agent):
@@ -295,6 +300,13 @@ def postprocess_single_calibration_output(self, agent):
 
             # Load observed data
             obs_df = pd.read_csv(self.obsflow, parse_dates=["value_date"])
+
+            # if obs is empty, raise an error
+            if obs_df.empty:
+                msg = f"Streamflow observation file is empty: {self.obsflow}"
+                logger.error(msg)
+                raise ValueError(msg)
+
             obs_df = obs_df.rename(
                 columns={"value_date": "Time", obs_df.columns[1]: obs_flow_col}
             ).set_index("Time")
@@ -316,8 +328,6 @@ def postprocess_single_calibration_output(self, agent):
             metrics_df = pd.DataFrame([metrics])
             metrics_df.insert(0, "iteration", 0, True)
 
-            # IS THIS CORRECT???????
-            # metrics_df["objFunVal"] = metrics_df[self.eval_params.objective.upper()]
             metrics_best_path = workdir / f"{basin_id}_metrics_iteration.csv"
             metrics_df.to_csv(metrics_best_path, index=False)
 

python/calib/src/calib/search.py

@@ -130,14 +130,35 @@ def _calc_metrics(
         simulated_hydrograph, observed_hydrograph, left_index=True, right_index=True
     )
     if df.empty:
-        _logger().warning("Cannot compute objective function, do time indicies align?")
+        msg = "No overlapping time period between simulated and observed streamflow. Metrics cannot be calculated. Exit."
+        _logger().error(msg)
+        raise ValueError(msg)
+
+    # If eval_range is provided, filter the dataframe to only include data within that range
     if eval_range:
         df = df.loc[eval_range[0] : eval_range[1]]
 
     df.reset_index(inplace=True)
 
+    # treat the data by removing negative values, NaN values, and replacing zero values with a small positive value
     df = treat_values(df, remove_neg=True, remove_na=True, replace_zero=True)
 
+    # if df is empty, log an error and raise an exception
+    if df.empty:
+        if eval_range:
+            eval_range_str = (
+                f" within the evaluation datetime range "
+                f"{eval_range[0].strftime('%Y-%m-%d %H:%M')} "
+                f"to {eval_range[1].strftime('%Y-%m-%d %H:%M')}"
+            )
+        else:
+            eval_range_str = ""
+
+        msg = f"There are no valid observed or simulated streamflow data{eval_range_str}. Metrics cannot be calculated. Exit."
+
+        _logger().error(msg)
+        raise ValueError(msg)
+
     # reset the time index (needed for calculation of event-based metrics)
     df.set_index(df.columns[0], inplace=True)
 

python/calib/src/calib/validation_run.py

@@ -92,15 +92,14 @@ def run_valid_ctrl_best(agent):
 
         outputs = [primary_set.output]
         runs = [agent.run_name]
-        if agent.run_name != "valid_control":
-            if agent.nwmflow is not None:
-                outputs.append(agent.nwmflow)
-                runs.append("nwm_retro")
+        if agent.run_name != "valid_control" and agent.nwmflow is not None:
+            outputs.append(agent.nwmflow)
+            runs.append("nwm_retro")
 
         for out1, run1 in zip(outputs, runs):
             metrics = pd.DataFrame()
-            # _logger().info(f"Computing metrics for out1 : {out1}, run1: {run1}")
             for key, value in time_period.items():
+                _logger().info(f"Computing metrics for {run1} and time period: {key}")
                 result = _calc_metrics(
                     out1,
                     primary_set.observed,