Rename example "dataset" classes, for clarity and collision avoidance

[apjanke]

Tablicious has a few different classes with base namedataset or datasets, enough to be confusing.

They're all example-dataset stuff, and are not equivalents of Matlab's dataset or the dataset in OF Statistics. Plus, similar class names that differ only in their namespace qualification or the "s" at the end can be confusing. And a name collision on just the base name of a namespaced class with a standard class could be a problem, e.g. when cd'ing around, in doctest, or in Octave stuff that doesn't properly support namespaces.

Rename them to something better. Do it for a minor version release, bc even though these are only in the examples code, that would be a bit of a breaking change for users.

Here, "ex dataset" is short for "example dataset".

Current classes:

• tblish.datasets - public interface to our ex datasets
• tblish.dataset - convenience wrapper around tblish.datasets, presenting them as static methods for tab-completion etc
• tblish.internal.dataset - internal implementation stuff for tblish.datasets (not tblish.dataset directly)
• +tblish.internal.datasets - namespace holding each ex dataset as an individual class
• tblish.internal.datasets.<name> - the actual ex dataseets, one per class, in @<name> directories to support resource files

What to do

I'm thinking it should be like "exdatasets" instead, and the +internal ones with "impl" or something in their base name. Maybe capitalize the class names, too? Though tblish.dataset is kinda mimicking a package in terms of the interface it presents. I dunno.

References

• Test-abort error for Tablicious, maybe around classes in namespaces, or categorical.missing gnu-octave/octave-doctest#283