Allow `ep.pp.encode` to produce sparse output

[Zethson]

Description of feature

I looked a bit into sparse encoding. One-hot encoding being the most important:

1. scikit-learn's one-hot encoding supports a sparse_output parameter that should return a CSR matrix.
2. We're getting original_values as numpy arrays when calling the function. May or may not be fine.
3. Currently we default the sparse_output parameter to False without checking the type of matrix.
4. The _update_encoded_data does not take sparse matrices into account

[eroell]

Try to narrow the discussion of this issue down by mentiong ep.pp.encode explicitly.

scipy's sparse arrays accept only numeric data, and no string or object dtypes. So ep.pp.encode would not be about encoding data in a sparse array; but producing an array that is sparse.

This would be a valid argument for encode to take in, and forward to sklearn's encode, and adapt the related functions of encode.