Various TODO's in the code

[jgallowa07]

I have a small litany of things I noticed during the last few PR's that I didn't want to get side-tracked with at the time. Here is the list of small things we could do to cleanup the source code and interface a little.

• Rename "huber_scale_huber" parameter -> "huber_scale"
• Get this to follow the -1 heuristic for using all CPU's
• we should probably remove the get_ prefix from get_mutations_df and get_variants_df . get should only be for getter/setter operations
• we should add method to output wide version of mutations df. This is really useful for many of the plots we make for the manuscript
• make sure for all plotting functions (that it's appropriate) we have the R^2 option to show coefficient of determination.
• remove the ipywidgets dependency and the fancy progress bars. They're annoying and unnecessary.
• rename the ModelCollection.fit_models attribute -> ModelCollection.models_df
• We could allow for singletons in the params dict passed to fit_models. It would make for a nicer interface
• Let's do the kwargs thing here
• Make the epistatic and activation function arguments to model a string instead of the function definition.
• Allow output of Model.get_mutations_df() to output either wide, or tall formats.
• Print more useful error message when fits fail in fit_models call
• Use pytest fixtures to speed things up and make global test inputs more consistent
• Make method for printing convergence trajectory for models.
• Let's make sure condition names match a certain pattern, like that there's no "." in the condition name. That would clean up a lot of the altair plotting code.