Fix catboost to work with has_time

[alex-hse-repository]

🚀 Feature Request

We need to reorder the points in the dataset, which is passed to the catboost model to make it able to use has_time parameter properly.

Proposal

Sort df by the "timestamp" column in the first line of the following methods:

• etna/etna/models/catboost.py

  Line 38 in 0bb7a64

  def fit(self, df: pd.DataFrame, regressors: List[str]) -> "_CatBoostAdapter":

• etna/etna/models/catboost.py

  Line 61 in 0bb7a64

  def predict(self, df: pd.DataFrame) -> np.ndarray:

Test cases

No test cases required

Additional context

No response