Merge pull request #167 from nasa/bug/calc_error_zero

Fix 0 case

Author: teubert

src/progpy/data_models/lstm_model.py

@@ -585,7 +585,7 @@ def from_data(cls, inputs, outputs, event_states=None, t_met=None, **kwargs):
             output_data.append(t_all)
 
         model = keras.Model(inputs, outputs)
-        model.compile(optimizer="rmsprop", loss="mse", metrics=["mae"]*len(outputs))
+        model.compile(optimizer="rmsprop", loss="mse", metrics=[["mae"]]*len(outputs))
 
         # Train model
         history = model.fit(

src/progpy/utils/calc_error.py

@@ -142,7 +142,7 @@ def MSE(m, times: List[float], inputs: List[dict], outputs: List[dict], **kwargs
             If the model goes unstable before stability_tol is met and short_sim_penalty is  None, then exception is raised
             Else if the model goes unstable before stability_tol is met and short_sim_penalty is not None- the penalty is added to the score
             Else, model goes unstable after stability_tol is met, the error calculated from data up to the instability is returned.
-        short_sim_penalty (float, optional): penalty added for simulation becoming unstable before stability_tol, added for each % below tol. Default is 100
+        short_sim_penalty (float, optional): penalty added for simulation becoming unstable before stability_tol, added for each % below tol. If set to None, operation will return an error if simulation becomes unstable before stability_tol. Default is 100
 
     Returns:
         float: Total error
@@ -190,6 +190,8 @@ def MSE(m, times: List[float], inputs: List[dict], outputs: List[dict], **kwargs
                     warn(f"Model unstable- NAN reached in simulation (t={t}) before cutoff threshold. "
                                      f"Cutoff threshold is {cutoffThreshold}, or roughly {stability_tol * 100}% of the data. Penalty added to score.")
                     # Return value with Penalty added
+                    if counter == 0:
+                        return 100*short_sim_penalty
                     return err_total/counter + (100-(t/cutoffThreshold)*100)*short_sim_penalty
                 else:
                     warn("Model unstable- NaN reached in simulation (t={})".format(t))

tests/test_calc_error.py

@@ -111,11 +111,14 @@ def future_loading(t, x=None):
 
         # With our current set parameters, our model goes unstable immediately
         with self.assertRaises(ValueError) as cm:
-            m.calc_error(simulated_results.times, simulated_results.inputs, simulated_results.outputs, dt=1)
+            m.calc_error(simulated_results.times, simulated_results.inputs, simulated_results.outputs, dt=1, short_sim_penalty=None)
         self.assertEqual(
             "Model unstable- NAN reached in simulation (t=0.0) before cutoff threshold. Cutoff threshold is 1900.0, or roughly 95.0% of the data",
             str(cm.exception)
-        ) 
+        )
+
+        # Shouldn't raise error for default case (i.e., short_sim_penalty is not None)
+        m.calc_error(simulated_results.times, simulated_results.inputs, simulated_results.outputs, dt=1)
 
         # Creating duplicate model to check if consistent results occur
         m1 = BatteryElectroChemEOD()
@@ -131,20 +134,26 @@ def future_loading(t, x=None):
 
         # Checks to see if model goes unstable before default stability tolerance is met.
         with self.assertRaises(ValueError) as cm:
-            m.calc_error(simulated_results.times, simulated_results.inputs, simulated_results.outputs, dt = 1)
+            m.calc_error(simulated_results.times, simulated_results.inputs, simulated_results.outputs, dt=1, short_sim_penalty=None)
         self.assertEqual(
             "Model unstable- NAN reached in simulation (t=1800.0) before cutoff threshold. Cutoff threshold is 1900.0, or roughly 95.0% of the data",
             str(cm.exception)
         )
+
+        # Shouldn't happen for default case (short_sim_penalty is not none)
+        m.calc_error(simulated_results.times, simulated_results.inputs, simulated_results.outputs, dt=1)
 
         # Checks to see if m1 throws the same exception. 
         with self.assertRaises(ValueError):
-            m1.calc_error(m1_sim_results.times, m1_sim_results.inputs, m1_sim_results.outputs, dt = 1)
+            m1.calc_error(m1_sim_results.times, m1_sim_results.inputs, m1_sim_results.outputs, dt=1, short_sim_penalty=None)
         self.assertEqual(
             "Model unstable- NAN reached in simulation (t=1800.0) before cutoff threshold. Cutoff threshold is 1900.0, or roughly 95.0% of the data",
             str(cm.exception)
         )
 
+        # Shouldn't for default case
+        m1.calc_error(m1_sim_results.times, m1_sim_results.inputs, m1_sim_results.outputs, dt=1)
+
         # Checks to see if stability_tolerance throws Warning rather than an Error when the model goes unstable after threshold
         with self.assertWarns(UserWarning) as cm:
             m.calc_error(simulated_results.times, simulated_results.inputs, simulated_results.outputs,