Fix PTSwap Output

[AaronDJohnson]

• Changes the PTSwap method to also scatter a numpy array of acceptances allowing the sampler to output PT swap acceptance rates as it did previously.

[kdolum]

The code looks right to me. Is the functionality documented somewhere that with each chain it outputs the acceptance rate of swaps between this chain and the next hotter one?
The header comment of PTswap discusses a return value swapReturn. Does that not exist? Maybe we should remove it from the documentation if so.
Would it be better to remove self.nswap_accepted = 0 from the initialization code?

[vhaasteren]

Aside from what @kdolum says, you also seem to have committed new versions of two example notebooks. But the notebooks do not seem to have changed, except for the figures. Looking at the figures nothing is updated.

So please remove the ipynb edits from the PR, they do not seem different. Or create a new PR that doesn't contain them

[AaronDJohnson]

@vhaasteren I will remove the ipynbs and make a separate PR. These are extremely minor errors that my linter picked up.

[AaronDJohnson]

@kdolum I've updated the docstring to hopefully make it clear that the function does not return swapAccept anymore, but it does still use the nswap_accepted parameter. It's necessary to keep this in the initialization so that we can increment it when swaps happen. The actual acceptance rates are computed elsewhere in the code.

[AaronDJohnson]

Notebook changes have been removed and moved to #49

[vhaasteren]

Looks good to me now. Also happy, @kdolum?

[kdolum]

Looks good to me now too.