Add error handling and warning when interruption is outside of `simul…

…ate` start and end time (#359)

* revise tests so that interventions occur before simulate start_time to demonstrate regression in error handling

* add value error and warning to StaticInterruption

chirho/dynamical/handlers/interruption.py

@@ -33,6 +33,21 @@ def __init__(self, time: R):
         self.time = torch.as_tensor(time)  # TODO enforce this where it is needed
         super().__init__()
 
+    def _pyro_simulate(self, msg) -> None:
+        _, _, start_time, end_time = msg["args"]
+
+        if self.time < start_time:
+            raise ValueError(
+                f"{StaticInterruption.__name__} time {self.time} occurred before the start of the "
+                f"timespan {start_time}. This interruption will have no effect."
+            )
+        elif self.time >= end_time:
+            warnings.warn(
+                f"{StaticInterruption.__name__} time {self.time} occurred after the end of the timespan "
+                f"{end_time}. This interruption will have no effect.",
+                UserWarning,
+            )
+
     def _pyro_get_next_interruptions(self, msg) -> None:
         _, _, _, start_time, end_time = msg["args"]
 
@@ -47,12 +62,6 @@ def _pyro_get_next_interruptions(self, msg) -> None:
                 or self.time < next_static_interruption.time
             ):
                 msg["kwargs"]["next_static_interruption"] = self
-        elif self.time >= end_time:
-            warnings.warn(
-                f"{StaticInterruption.__name__} time {self.time} occurred after the end of the timespan "
-                f"{end_time}. This interruption will have no effect.",
-                UserWarning,
-            )
 
 
 class DynamicInterruption(Generic[T], Interruption):

tests/dynamical/test_static_interventions.py

@@ -28,7 +28,7 @@
 # Points at which to measure the state of the system.
 start_time = torch.tensor(0.0)
 end_time = torch.tensor(10.0)
-logging_times = torch.linspace(start_time + 1, end_time - 2, 5)
+logging_times = torch.linspace(start_time + 0.01, end_time - 2, 5)
 
 # Initial state of the system.
 init_state_values = State(