Reloading fit results is slow

[Jhsmit]

Performance is poor when using pyhdx.fileIO.load_fitresult to load a previous fit result.

This is because of the methods generate_output and estimate_errors are invoked, although this results are known so this is not strictly necessary.