Support skip_runners_and_metrics option (facebook#3105)

Summary:
Pull Request resolved: facebook#3105

Support options for not loading runners and metric types for MultiTypeExperiment

Reviewed By: sdaulton

Differential Revision: D66280618

fbshipit-source-id: 1a5c0acc6b9737b1ed0eaf056a079fb49554fe9d

Author: andycylmeta

ax/core/multi_type_experiment.py

@@ -47,7 +47,7 @@ def __init__(
         name: str,
         search_space: SearchSpace,
         default_trial_type: str,
-        default_runner: Runner,
+        default_runner: Runner | None,
         optimization_config: OptimizationConfig | None = None,
         tracking_metrics: list[Metric] | None = None,
         status_quo: Arm | None = None,
@@ -79,7 +79,7 @@ def __init__(
         self._default_trial_type = default_trial_type
 
         # Map from trial type to default runner of that type
-        self._trial_type_to_runner: dict[str, Runner] = {
+        self._trial_type_to_runner: dict[str, Runner | None] = {
             default_trial_type: default_runner
         }
 

ax/storage/sqa_store/decoder.py

@@ -197,7 +197,8 @@ def _init_experiment_from_sqa(
         )
 
     def _init_mt_experiment_from_sqa(
-        self, experiment_sqa: SQAExperiment
+        self,
+        experiment_sqa: SQAExperiment,
     ) -> MultiTypeExperiment:
         """First step of conversion within experiment_from_sqa."""
         opt_config, tracking_metrics = self.opt_config_and_tracking_metrics_from_sqa(
@@ -217,24 +218,41 @@ def _init_mt_experiment_from_sqa(
             if experiment_sqa.status_quo_parameters is not None
             else None
         )
+
+        default_trial_type = none_throws(experiment_sqa.default_trial_type)
         trial_type_to_runner = {
             none_throws(sqa_runner.trial_type): self.runner_from_sqa(sqa_runner)
             for sqa_runner in experiment_sqa.runners
         }
-        default_trial_type = none_throws(experiment_sqa.default_trial_type)
+        if len(trial_type_to_runner) == 0:
+            trial_type_to_runner = {default_trial_type: None}
+            trial_types_with_metrics = {
+                metric.trial_type
+                for metric in experiment_sqa.metrics
+                if metric.trial_type
+            }
+            # trial_type_to_runner is instantiated to map all trial types to None,
+            # so the trial types are associated with the expeirment. This is
+            # important for adding metrics.
+            trial_type_to_runner.update(
+                {t_type: None for t_type in trial_types_with_metrics}
+            )
         properties = dict(experiment_sqa.properties or {})
         default_data_type = experiment_sqa.default_data_type
         experiment = MultiTypeExperiment(
             name=experiment_sqa.name,
             description=experiment_sqa.description,
             search_space=search_space,
             default_trial_type=default_trial_type,
-            default_runner=trial_type_to_runner[default_trial_type],
+            default_runner=trial_type_to_runner.get(default_trial_type),
             optimization_config=opt_config,
             status_quo=status_quo,
             properties=properties,
             default_data_type=default_data_type,
         )
+        # pyre-ignore Imcompatible attribute type [8]: attribute _trial_type_to_runner
+        # has type Dict[str, Optional[Runner]] but is used as type
+        # Uniont[Dict[str, Optional[Runner]], Dict[str, None]]
         experiment._trial_type_to_runner = trial_type_to_runner
         sqa_metric_dict = {metric.name: metric for metric in experiment_sqa.metrics}
         for tracking_metric in tracking_metrics:

ax/storage/sqa_store/encoder.py

@@ -195,7 +195,7 @@ def experiment_to_sqa(self, experiment: Experiment) -> SQAExperiment:
         if isinstance(experiment, MultiTypeExperiment):
             properties[Keys.SUBCLASS] = "MultiTypeExperiment"
             for trial_type, runner in experiment._trial_type_to_runner.items():
-                runner_sqa = self.runner_to_sqa(runner, trial_type)
+                runner_sqa = self.runner_to_sqa(none_throws(runner), trial_type)
                 runners.append(runner_sqa)
 
             for metric in tracking_metrics:

ax/storage/sqa_store/tests/test_sqa_store.py

@@ -609,6 +609,23 @@ def test_MTExperimentSaveAndLoad(self) -> None:
         self.assertEqual(loaded_experiment._metric_to_canonical_name["m2"], "m1")
         self.assertEqual(len(loaded_experiment.trials), 2)
 
+    def test_MTExperimentSaveAndLoadSkipRunnersAndMetrics(self) -> None:
+        experiment = get_multi_type_experiment(add_trials=True)
+        save_experiment(experiment)
+        loaded_experiment = load_experiment(
+            experiment.name, skip_runners_and_metrics=True
+        )
+        self.assertEqual(loaded_experiment.default_trial_type, "type1")
+        # pyre-fixme[16]: `Experiment` has no attribute `_trial_type_to_runner`.
+        self.assertIsNone(loaded_experiment._trial_type_to_runner["type1"])
+        self.assertIsNone(loaded_experiment._trial_type_to_runner["type2"])
+        # pyre-fixme[16]: `Experiment` has no attribute `metric_to_trial_type`.
+        self.assertEqual(loaded_experiment.metric_to_trial_type["m1"], "type1")
+        self.assertEqual(loaded_experiment.metric_to_trial_type["m2"], "type2")
+        # pyre-fixme[16]: `Experiment` has no attribute `_metric_to_canonical_name`.
+        self.assertEqual(loaded_experiment._metric_to_canonical_name["m2"], "m1")
+        self.assertEqual(len(loaded_experiment.trials), 2)
+
     def test_ExperimentNewTrial(self) -> None:
         # Create a new trial without data
         save_experiment(self.experiment)