Use underlying index for mutual info with pandas

- As a user, I wish Woodwork's mutual information used a DataFrame's underlying index when including the index in calculations, that way, if a user hasn't set a Woodwork index but has an underlying index that matters to them, setting `include_index=True` will allow the index to be included in mutual information calculations.

This is only relevant for Pandas DataFrames, as we do not set the underlying index for Dask and Koalas DataFrames.