Various Audition Improvements (#844)

* remove alpha for selection rule plots

* add support for distinguishing between baseline model groups for plotting but not model selection

* ensure stable colors for audition plotting

* debug

* debug and update tutorial notebook

* remove old example audition notebook

* fix up audition tutorial imports

* fix for no baselines

src/tests/audition_tests/test_preaudition.py

@@ -78,10 +78,13 @@ class ImmediateEvalFactory(EvaluationFactory):
         pre_aud = PreAudition(db_engine)
 
         # Expect the number of model groups with label_1
-        assert len(pre_aud.get_model_groups_from_label("label_1")) == sum(
+        assert len(pre_aud.get_model_groups_from_label("label_1")['model_groups']) == sum(
             [x["label_definition"] == "label_1" for x in model_configs]
         )
 
+        # Expect no baseline model groups
+        assert len(pre_aud.get_model_groups_from_label("label_1")['baseline_model_groups']) == 0
+
         # Expect the number of model groups with certain experiment_hash
         experiment_hash = list(
             pd.read_sql(
@@ -92,7 +95,7 @@ class ImmediateEvalFactory(EvaluationFactory):
                 con=db_engine,
             )["experiment_hash"]
         )[0]
-        assert len(pre_aud.get_model_groups_from_experiment(experiment_hash)) == 1
+        assert len(pre_aud.get_model_groups_from_experiment(experiment_hash)['model_groups']) == 1
 
         # Expect the number of model groups for customs SQL
         query = """