Allow to slice not just rows but columns too for Pools

According to catboost [documentation ](https://catboost.ai/docs/en/concepts/python-reference_pool_slice) you can slice dataset passing index values as argument to pool.slice().
```python
from catboost import Pool

data = [[1, 3],
[0, 4],
[1, 7],
[6, 4],
[5, 3]]

dataset = Pool(data)
print(dataset.num_row())
dataset_part = dataset.slice([0, 1, 2])
print(dataset_part.num_row())
```
However during feature selection it may be very handy to slice columns instead of creating new pool with necessary columns which would occupy CPU. Perhaps passing list of feature names or feature indexes to slice function?

```python
dataset_part = dataset.slice(rows=[0, 1, 2],columns=['feature_2','feature_3'])
```

Best regards!