finalize reward function handling

dominoes/datasets/base.py

@@ -147,14 +147,21 @@ def _check_task(self, task):
 
     def _check_parameters(self, reference=None, init=False, **task_parameters):
         """
-        check if the parameters provided at initialization are valid and complete
+        check if parameters provided in the task_parameters are valid (and complete)
 
         checks two things:
         1. If any parameters are provided that are not recognized for the task, an error will be generated
+
+        ... if init=True, will additionally check:
         2. If any parameters are required for the task but not provided, an error will be generated
 
         args:
+            reference: dict, the reference parameters to check against (if not provided, uses self._required_parameters())
+            init: bool, whether this is being called by the constructor's __init__ method
+                  in practive, this determines whether any required parameters without defaults are set properly
             task_parameters: dict, the parameters provided at initialization
+
+        raise ValueError if any parameters are not recognized or required parameters are not provided
         """
         if reference is None:
             reference = self._required_parameters()
@@ -317,11 +324,17 @@ def reward_function(self, choices, batch, **kwargs):
         args:
             choice: torch.Tensor, index to the choices made by the network
             batch: tuple, the batch of data generated for this training step
+            kwargs: optional kwargs for each task-specific reward function
 
         returns:
             torch.Tensor, the rewards for the network
         """
-        pass
+        if self.task == "sequencer":
+            return self._measurereward_sequencer(choices, batch, **kwargs)
+        elif self.task == "sorting":
+            return self._measurereward_sorter(choices, batch, **kwargs):
+        else:
+            raise ValueError(f"task {self.task} not recognized")
 
     @torch.no_grad()
     def _gettarget_sequencer(self, dominoes, selection, available, **prms):
@@ -391,7 +404,7 @@ def _gettarget_sorting(self, dominoes, selection, **prms):
         return dict(target=target, value=value)
 
     @torch.no_grad()
-    def _measurereward_sequencer(self, choices, batch, return_direction=False):
+    def _measurereward_sequencer(self, choices, batch, return_direction=False, verbose=None):
         """
         reward function for sequencer
 
@@ -432,7 +445,6 @@ def _measurereward_sequencer(self, choices, batch, return_direction=False):
         null_index = copy(num_in_hand)
 
         # check verbose
-        verbose = batch.get("verbose", None)
         if verbose is not None:
             debug = True
             assert 0 <= verbose < batch_size, "verbose should be an index corresponding to one of the batch elements"
@@ -554,7 +566,7 @@ def _measurereward_sequencer(self, choices, batch, return_direction=False):
             return rewards
 
     @torch.no_grad()
-    def _measurereward_sorter(self, choices, batch):
+    def _measurereward_sorter(self, choices, batch, **kwargs):
         """
         measure the reward for the sorting task