Mitchell O'Hara-Wild

Thanks for investigating this further. I believe this is due to scoping issues with parallel modelling in fabletools: https://github.com/tidyverts/fabletools/issues/146 The model formula will need to be carefully parsed to identify...

Currently the python version does not support providing a set of innovations - so this method can either re-implement the simulation method, or disallow specifying the innovations.

This is lower priority now that some form of vectorisation (#52) is implemented. The value proposition here is less enticing than the complexity it introduces when compared to #52.

The documentation should definitely be improved to explain how the distributions and arguments are vectorised. Essentially the operation's argument (e.g. `density(at=)`) is vectorised across the distributions and the result is...

From discussions with @robjhyndman, it is reasonable to return a transformed distribution when comparing two distributions (or reasonably pointed out, a distribution with a degenerate distribution).

Great, this is exactly what I had in mind for this functionality. Thanks for your comment!

The last issue with supporting `inv_box_cox` would be the use of `[`. I think it is important if `dist[numeric]` is used, then it should index the distribution vector. However if...

Hmm, I hadn't considered the use of indexing by a singular discrete distribution - your proposal sounds reasonable (although fairly specific for the inputs). I currently have an experimental [`dist_mixture()`](http://pkg.mitchelloharawild.com/distributional/reference/dist_mixture.html)...

Thanks for working on this, it'll take a little while to go through it all and review appropriately. One thing I'm not convinced by yet is nesting the transformed distribution...

Can definitely modify the print method, although I think needing to do that is a symptom of bad design. Storing the functions in a list could work, although the user...