Check that time index specified is a valid datetime Type about woodwork

[thehomebrewnerd]

Are we going to support a numeric time index as well? If so we need to add some additional checks just beyond the datetime check.

[gsheni]

Yes, maybe not initially, but eventually we will.

[thehomebrewnerd]

Thinking about this a little more, I think maybe we should follow the approach used by Featuretools, where we check if the specified column contains datetime-like values. For example a column with properly formatted strings can be specified as time index. If the column contains values that can be converted to datetime, we can perform the underlying conversion on the dataframe, if necessary, rather than requiring the user to do this manually beforehand.