Handling data without N-dimensional form

[ssomnath]

The most popular reasons for data not to have an N-dimensional form are:

1. Incomplete acquisition - incomplete (position) dimensions
2. Inability to fold dimensions - eg. Relaxation BEPS - dependent (spectroscopic) dimension
3. Sparse sampling - incomplete (position) dimensions

One solution is to add more (optional) attributes to the (position or spectroscopic) ancillary datasets:

1. incomplete_dimensions = ['X', 'Y']
2. dependent_dimensions = ['Field']

The reshape, visualization, etc. functions could ignore dimensions listed in dependent_dimensions and not even try to reshape if incomplete_dimensions is set.

Others are welcome to propose alternate / improvements to this solution.

This solution needs to be tested out in pyUSID.