Simplify getting atomic scattering parameters

[hakonanes]

Is your feature request related to a problem? Please describe. There are now multiple parametrizations of atomic scattering parameters in the code:

• Doyle-Turner: https://github.com/pyxem/diffsims/blob/master/diffsims/structure_factor/atomic_scattering_parameters.py#L39
• International Tables (xtables): https://github.com/pyxem/diffsims/blob/master/diffsims/utils/atomic_scattering_params.py#L22
• Lobato: https://github.com/pyxem/diffsims/blob/master/diffsims/utils/lobato_scattering_params.py#L25
• Peng: https://github.com/pyxem/diffsims/blob/master/diffsims/utils/discretise_utils.py#L907
• Kirkland: https://github.com/pyxem/diffsims/blob/master/diffsims/utils/scattering_params.py#L30

Describe the solution you'd like

• All parametrizations should be placed in one (it's own?) module (directory).
• The module's location should probably be in the diffsims.structure_factor module, which makes most sense?
• The module should have one public function taking in preferably a diffpy.structure.Atom with an element ID in Atom.element (like "H", "Ni", etc.), a string or an integer, and which parametrization of the above to use. The parameters are returned.

[hakonanes]

Got no time for this unfortunately, hopefully someone else can pick this up.

[CSSFrancis]

@hakonanes any thoughts on this? I was going to add in ionic scattering parameters (I guess why not right) and figured I should go all the way if I'm doing that.

[CSSFrancis]

I was thinking that we set up each table in the form of a nested dict with:

{element: 
    {a=[a1,a2, ..., an],
     b=[b1,b2,...., bn],
     z= Atomic Number,
     charge = ionic charge}

[pc494]

This would obviously be an improvement to the API, but unless somebody needs it urgently I think we should deprioritise it (at least until #201 and it's support PR are in). Obviously if @CSSFrancis is using ionic scattering day-to-day we should pick it back up.

[CSSFrancis]

@pc494 I'm not using them and I should probably focus my efforts a bit more so I'll leave this at I'll come back to it unless someone else needs it.

Releasing a 1.0.0 version of pyxem is already looking pretty daunting but we see getting there slowly.