Convert functional algo tests to python API

src/orion/algo/hyperband.py

@@ -147,8 +147,6 @@ def __init__(self, space, seed=None, repetitions=numpy.inf):
             for bracket_budgets in self.budgets
         ]
 
-        self.seed_rng(seed)
-
     def sample(self, num, bracket, buffer=10):
         """Sample new points from bracket"""
         sample_buffer = bracket.rungs[0]['n_trials'] * buffer
@@ -186,7 +184,7 @@ def seed_rng(self, seed):
         """
         self.seed = seed
         for i, bracket in enumerate(self.brackets):
-            bracket.seed_rng(seed + i if seed is not None else None)
+            bracket.seed_rng(self.executed_times + seed + i if seed is not None else None)
         self.rng = numpy.random.RandomState(seed)
 
     @property
@@ -276,8 +274,6 @@ def _refresh_bracket(self):
                     Bracket(self, bracket_budgets, self.executed_times + 1)
                     for bracket_budgets in self.budgets
                 ]
-                if self.seed is not None:
-                    self.seed += 1
 
     def _get_bracket(self, point):
         """Get the bracket of a point during observe"""

src/orion/client/__init__.py

@@ -225,18 +225,19 @@ def workon(function, space, name='loop', algorithms=None, max_trials=None):
     # Clear singletons and keep pointers to restore them.
     singletons = update_singletons()
 
-    setup_storage(storage={'type': 'legacy', 'database': {'type': 'EphemeralDB'}})
-
-    experiment = experiment_builder.build(
-        name, version=1, space=space, algorithms=algorithms,
-        strategy='NoParallelStrategy', max_trials=max_trials)
+    try:
+        setup_storage(storage={'type': 'legacy', 'database': {'type': 'EphemeralDB'}})
 
-    producer = Producer(experiment)
+        experiment = experiment_builder.build(
+            name, version=1, space=space, algorithms=algorithms,
+            strategy='NoParallelStrategy', max_trials=max_trials)
 
-    experiment_client = ExperimentClient(experiment, producer)
-    experiment_client.workon(function, max_trials=max_trials)
+        producer = Producer(experiment)
 
-    # Restore singletons
-    update_singletons(singletons)
+        experiment_client = ExperimentClient(experiment, producer)
+        experiment_client.workon(function, max_trials=max_trials)
+    finally:
+        # Restore singletons
+        update_singletons(singletons)
 
     return experiment_client