Determine if `end()` is necessary in pyttb classes

[dmdunla]

Do we the end() methods in pyttb classes?

The end() method is TTB for MATLAB is used to identify the index of the last element of a tensor using linear indexing or the last index along a dimension when using subscript indexing.

numpy , which is used for most data objects in pyttb classes, supports negative indexing but in a different way than what end() does.

Should we implement end() methods that wrap numpy negative indexing (e.g., X[-1,-1,-1] for subscript indexing or X.flat[-1] for linear indexing? Or remove the end() methods?

[ntjohnson1]

I think we should avoid using end() for all indexing operations since that's very unnatural for python. I believe our slicing support already handles negative values. I was just looking at an example from sptensor X[110:, 110:, 115:].isequal(X[-10:, -10:, -5:]) from this PR #152 (after I apply some fixes to the sptensor indexing but nothing specific to negative indices).

[tgkolda]

It looks like -1 should work fine in lieu of end. I would suggest to have some explanation of this in the documentation, especially for those transitioning from MATLAB.

[dmdunla]

Answered this question. Will use -1 indexing in lieu of end() in class methods.

Follow up:

• Remove end() from class methods #179