Random 3D Tomo Loader

Description

A hdf5 dataset of a specified size is created at runtime using numpy random sampling (numpy.random), saved with relevant meta_data to a NeXus file, and used as input. It recreates the behaviour of the nxtomo loader but with random data. The input file path passed to Savu will be ignored (use a dummy). Note: Further extensions planned to allow the generated data to be re-loaded with the nxtomo_loader.