[WIP] French-Wilson script rs.if2

[tjlane]

This PR implements a simple script that exposes rs's French-Wilson algorithm to the command line. Addressing #51

[tjlane]

@kmdalton some advice needed here regarding the FW algorithm parameters.

From a user's point of view, I'd like to hide as many of the parameters as possible (bins, bin width, sigma). Those are a bit mysterious for me, so I assume they will also be out of the domain of knowledge of a 99% of users -- and we don't provide any help/guidelines at present.

Are the defaults here robust & sane? Better: can we algorithmically pick optimal (or at least sensible) values?

[kmdalton]

Let's chat offline. The gist is that the default parameters should work for a large range of examples, but it is probably possible and better to infer reasonable values from the inputs.

[tjlane]

Notes from quick chat with @kmdalton on the phone just now:

For the default parameters

• Bandwidth estimator --> use scipy
• Grid size --> function of number of reflections

Then, this script probably should go into rs directly.
Finally, structure factor normalization is not a well-hashed out thing in rs yet. The FW procedure depends on the normalization.