Boolean variables not supported in combination with other categoricals

[Dvermetten]

Describe the bug When the search space contains a boolean variable in addition to another categorical variable which is non-boolean, the search will fail.

To Reproduce The following is a slight modification of an existing test to show the problem `dim_r = 2 # dimension of the real values def obj_fun(x): x_r = np.array([x['continuous_%d'%i] for i in range(dim_r)]) x_i = x['ordinal'] x_d = x['nominal'] _ = 0 if x_d == 'OK' else 1 return np.sum(x_r ** 2) + abs(x_i - 10) / 123. + _ * 2

search_space = ContinuousSpace([-5, 5], var_name='continuous') * dim_r +
OrdinalSpace([5, 15], var_name='ordinal') +
NominalSpace(['OK', 'A', None], var_name='nominal') +
NominalSpace([True, False], var_name='boolvar')

model = RandomForest(levels=search_space.levels)

opt = ParallelBO( search_space=search_space, obj_fun=obj_fun, model=model, max_FEs=6, DoE_size=3, # the initial DoE size eval_type='dict', acquisition_fun='MGFI', acquisition_par={'t' : 2}, n_job=3, # number of processes n_point=3, # number of the candidate solution proposed in each iteration verbose=False # turn this off, if you prefer no output ) xopt, fopt, stop_dict = opt.run()`

Expected behavior This should perform exactly the same as the case without boolean variable

Additional context It seems to be related to the checking of input in the random forest

[Dvermetten]

Somewhere in the random forest these values are converted to strings, which is leading to this issue. I haven't yet found exactly where this issue occurs, but it is not just limited to boolean variable, but any nominal space where the options are not strings originally share this problem.

[MariosKef]

@Dvermetten I actually have a similar issue (maybe the same). I used pip to install the latest version and I am getting errors of the form: ValueError: Found unknown categories ['22', '18', '26', '96'] in column 0 during transform. I am guessing it's what you said above because I have a specific nominal range the elements of which are not strings originally: max_depth = NominalSpace([None] + np.arange(2,102,2).tolist())

Is there an update/fix on this? I also think the pip version is an older version than what is now on master...

Thanks!