Switched from using pop to get in _initiateSolverAndObjFunc.  It was causing the fit to fail on large datasets.

Swapped calls to multiproccessing to use joblib instead.
Updated method functions to work with joblib.

Fixed h5_spec_vals attribute being the Spectroscopic Indices dataset

Also clean up unused imports

Fixed bug when using the new as_scalar kwarg

Added as_scalar parameter to get_n_dim_form and slice methods.

Now goes 1. PycroDataset, 2. reshape_to_Ndims, 3. manual slicing

Cades dev

Removed call to self.get_n_dim_form()
Fixed bug in loop over labels and sizes
Fixed bug when only a single position selected.

Bug if h5_pos is array of size 1.

Removed some unneeded lines of code and slight docstring change

Added exceptions to getH5DsetRefs and findDataset to allow returning of normal H5 Dataset if not a 'main' dataset

Function calls that used to return h5py.Dataset objects will now return PycroDataset objects.

Refactor how dataset attributes are set.
Added more documentation for the class.

slice method now operates directly on the h5 dataset.
Added some documentation

Fixed errors with labeling due to not having datasets for Position and/or Spectroscopic dimensions

Use the matplotlib notebook backend
Added new visualization

Updated all functions to work with the matplotlib notebook backend.
Replaced x,y position sliders in all functions with crosshair based on click location
Added new plotting function plot_loop_sho_raw_comparison.

Exports the data in the selected figure to a text file.