Implement "quick" pre-check of CUBE file contents for faster rejection at cmdline

[bskinn]

In cases where the data contained in a CUBE file is invalid in some way, a "quick" pre-check to ensure the correct data dimensionality would allow a faster fail-exit at the commandline.

The best way is probably, per here, something like sum(1 for _ in re.finditer('[^\\s\\n][\\s\\n]+[^\\s\\n]')). There would still be the unavoidable overhead of loading the file, but maybe this would be faster....

OTOH, the UX load wait could be improved by a file-load status bar, implemented as part of #50.