[ENH] issues and features for the `Empirical` distribution

[fkiraly]

There are a number of issues with the Empirical distribution that should be addressed:

• [ ] the time_indep argument does nothing
• [ ] there should be a similar argument for independent rows
• [x] scalar option currently does not work

[smilingprogrammer]

Hello. Interested in this issue. what fuctionality is expected from the time_indep argument.

[fkiraly]

if True, behaves as current; if False, samples entire time series, so different rows may be dependent