Add some errorhandling for custom estimator

freqtrade/optimize/hyperopt.py

@@ -45,7 +45,7 @@ progressbar.streams.wrap_stdout()
 logger = logging.getLogger(__name__)
 
 
-INITIAL_POINTS = 30
+INITIAL_POINTS = 5
 
 # Keep no more than SKOPT_MODEL_QUEUE_SIZE models
 # in the skopt model queue, to optimize memory consumption
@@ -366,13 +366,19 @@ class Hyperopt:
 
     def get_optimizer(self, dimensions: List[Dimension], cpu_count) -> Optimizer:
         estimator = self.custom_hyperopt.generate_estimator()
+
+        acq_optimizer = "sampling"
+        if isinstance(estimator, str):
+            if estimator not in ("GP", "RF", "ET", "GBRT"):
+                raise OperationalException(f"Estimator {estimator} not supported.")
+            else:
+                acq_optimizer = "auto"
+
         logger.info(f"Using estimator {estimator}.")
-        # TODO: Impact of changing acq_optimizer to "sampling" is unclear
-        # (other than that it fails with other optimizers when using custom sklearn regressors)
         return Optimizer(
             dimensions,
             base_estimator=estimator,
-            acq_optimizer="sampling",
+            acq_optimizer=acq_optimizer,
             n_initial_points=INITIAL_POINTS,
             acq_optimizer_kwargs={'n_jobs': cpu_count},
             random_state=self.random_state,

tests/optimize/test_hyperopt.py

@@ -884,6 +884,10 @@ def test_in_strategy_auto_hyperopt(mocker, hyperopt_conf, tmpdir, fee) -> None:
     assert hyperopt.backtesting.strategy.buy_rsi.value != 35
     assert hyperopt.backtesting.strategy.sell_rsi.value != 74
 
+    hyperopt.custom_hyperopt.generate_estimator = lambda *args, **kwargs: 'ET1'
+    with pytest.raises(OperationalException, match="Estimator ET1 not supported."):
+        hyperopt.get_optimizer([], 2)
+
 
 def test_SKDecimal():
     space = SKDecimal(1, 2, decimals=2)