Hyper-parameter Tuning of Data preparation steps

Hi,
In the book, it is mention that we can use some of the data preparation steps as hyper parameters eg `add_bedrooms_per_room` can be hyperparameter

I tried putting it directly in the GridSearchCV, but it gives an error.

```
from sklearn.model_selection import GridSearchCV
param_grid={
'n_estimators':[3,10,30],'max_features':[2,4,6,8],
'bootstrap':[False,True],'n_estimators':[3,10],'max_features':[2,3,4],
'add_bedrooms_per_room':[True,False]
}

forest_reg=RandomForestRegressor()
grid_serach=GridSearchCV(forest_reg,param_grid,cv=5,scoring="neg_mean_squared_error")
grid_serach.fit(housing_prepared,housing_labels)
```

Can you please explain how to do it? and I want to tune this` 'add_bedrooms_per_room':[True,False]` and also use GridSearchCV to find out what can be the best strategy to fill missing values when we pass` 'stategy':["mean","median"]` etc

Please explain it using this example only:
https://github.com/ageron/handson-ml/blob/master/02_end_to_end_machine_learning_project.ipynb