Calibration plotting

As per #360 (closed), I've added a method to the CBCResult object to plot the calibration, where I pretty much just copied over and slightly amended the code from bayespputils that LALInference uses. I have kept the call to the bayespputils functions for the actual plotting and calculations. Are people ok with this or would it be preferable to copy over those functions directly? I still need to add some sort of switch to check if the results include the calibration posteriors, because obviously it will fail if those are missing. I'm also kind of confused about the CBCResult object. I tested the standalone function on my run with GW170814 before incorporating it into the CBCResult object, and it produces nice plots (see attached). However, I don't really understand how to instantiate the CBCResult object class. If I use bilby.core.result.read_in_result to load the results file, can I cast this to a CBCResult object somehow? Otherwise I think it would make more sense to just keep this as a separate plotting function somewhere.