DaskEngine: Repeated Use with Context Manager

Per #2667 , it was noted that a DaskEngine can't be used in a context manager as the manager will shutdown the associated cluster. We should probably modify the `__enter__()` function to restart the cluster if it's shutdown.

```
def test_automl_dask_called_twice(
X_y_binary_cls,
):
"""Making sure that the max_iterations parameter limits the number of pipelines run."""

X, y = X_y_binary_cls
max_iterations = 4
engine = DaskEngine()
with engine:
par_automl = AutoMLSearch(
X_train=X,
y_train=y,
problem_type="binary",
engine=engine,
max_iterations=max_iterations,
)
par_automl.search()

with engine:
par_automl = AutoMLSearch(
X_train=X,
y_train=y,
problem_type="binary",
engine=engine,
max_iterations=max_iterations,
)
par_automl.search()
```

The above code snippet should be modified into a test that passes.