Fixed catboost and added new tqdm

[Tialo]

Hi, I see you've been working on the project, you are doing great job!

Catboost does not have keys of parameters, which weren't passed in initialization, so your _get_tree_num will raise KeyError. My PR fixes it.

from arfs.feature_selection import Leshy
from catboost import CatBoostClassifier
from sklearn.datasets import load_iris


x, y = load_iris(return_X_y=True, as_frame=True)

estimator = CatBoostClassifier()
leshy = Leshy(estimator, n_estimators="auto")
leshy.fit(x, y)

gives KeyError

Also when I was using your module in jupyter notebook, it has been using not fancy looking tqdm. It is easy to fix by using
from tqdm.auto import tqdm which will import nice tqdm if it can, otherwise will use usual version.