Hi,
Unfortunately I don’t believe it is possible to specify a relative azimuth angle as input to the underlying 6S model, so therefore it isn’t possible using Py6S either. I can see from the 6S code that relative azimuth angles are used internally for some things, but the documentation doesn’t show a way of specifying this as an input parameter.
I don’t know how the people who wrote the other papers did it - I suspect they just ran for lots of solar azimuths and view azimuths, and then calculated the relative azimuth angle for each of them. It might be worth contacting them to see if they can give you more details on how they did it.