Add test

CHANGELOG.md

@@ -5,8 +5,8 @@
 - `recommend` now provides an evaluated candidate when possible. For non-deterministic parametrization like `Choice`, this means we won't
   resample, and we will actually recommend the best past evaluated candidate [#668](https://github.com/facebookresearch/nevergrad/pull/668).
   Still, some optimizers (like `TBPSA`) may recommend a non-evaluated point.
-- `Choice` now takes a new `repetitions` parameters for sampling several times,
-  it is equivalent to :code:`Tuple(*[Choice(options) for _ in range(repetitions)])` but can be be around 30x faster for large numbers of repetitions [#670](https://github.com/facebookresearch/nevergrad/pull/670).
+- `Choice` and `TransitionChoice` can now take a `repetitions` parameters for sampling several times,
+  it is equivalent to :code:`Tuple(*[Choice(options) for _ in range(repetitions)])` but can be be up to 30x faster for large numbers of repetitions [#670](https://github.com/facebookresearch/nevergrad/pull/670).
 - Defaults for bounds in `Array` is now `bouncing`, which is a variant of `clipping` avoiding over-sompling on the bounds [#684](https://github.com/facebookresearch/nevergrad/pull/684)
   and [#691](https://github.com/facebookresearch/nevergrad/pull/691).
 

nevergrad/parametrization/test_parameter.py

@@ -315,6 +315,17 @@ def test_choice_repetitions() -> None:
     choice.mutate()
 
 
+def test_transition_choice_repetitions() -> None:
+    choice = par.TransitionChoice([0, 1, 2, 3], repetitions=2)
+    choice.random_state.seed(12)
+    assert len(choice) == 4
+    assert choice.value == (2, 2)
+    choice.value = (3, 1)
+    np.testing.assert_almost_equal(choice.position.value, [3.5, 1.5], decimal=3)
+    choice.mutate()
+    assert choice.value == (3, 0)
+
+
 def test_descriptors() -> None:
     d1 = utils.Descriptors()
     d2 = utils.Descriptors(continuous=False)