Update min_asks to directly inherit from TransitionCriterion (#328)

Summary:
Pull Request resolved: #328

This diff we update the 3rd of 4 aepsych completion criterion to directly rely on transitioncriterion

In following diffs we will:
- Completely remove the completion criterion file
- revisit storage
- remove all todos in gennode, genstrat, and transitioncriterion classes related to maintaining this deprecated code
- update AEPsych GSs as needed
- determine if run indefinetly can be replaced by simply having gen_unlimited_trials = true
- determine if additional updates to min_asks and run_indefinetly are necessary

Reviewed By: crasanders

Differential Revision: D52852941

fbshipit-source-id: b515689c199f5d5f8ba1d547b12d171c4bbae6ba

aepsych/generators/completion_criterion/min_asks.py

@@ -5,21 +5,30 @@
 # This source code is licensed under the license found in the
 # LICENSE file in the root directory of this source tree.
 
-from typing import Any, Dict
-
-from ax.core.experiment import Experiment
-from ax.modelbridge.completion_criterion import CompletionCriterion
+from typing import Any, Dict, Optional, Set
 
 from aepsych.config import Config, ConfigurableMixin
 
+from ax.core.experiment import Experiment
+from ax.modelbridge.transition_criterion import TransitionCriterion
 
-class MinAsks(CompletionCriterion, ConfigurableMixin):
+
+class MinAsks(TransitionCriterion, ConfigurableMixin):
     def __init__(self, threshold: int) -> None:
         self.threshold = threshold
 
     def is_met(self, experiment: Experiment) -> bool:
         return experiment.num_asks >= self.threshold
 
+    def block_continued_generation_error(
+        self,
+        node_name: Optional[str],
+        model_name: Optional[str],
+        experiment: Optional[Experiment],
+        trials_from_node: Optional[Set[int]] = None,
+    ) -> None:
+        pass
+
     @classmethod
     def get_config_options(cls, config: Config, name: str) -> Dict[str, Any]:
         min_asks = config.getint(name, "min_asks", fallback=1)