Cleanup of plotting choices in correlation function SysTests

[msimet]

We've got feedback from our first non-developer user, yay!

Here are the requests:

• Add definitions of xi_pm, xi_pm,re, etc to the docstring
• Add units to the x-axis
• Add a dashed line at xi=0
• Use the form \xi_re instead of \xi, since it's the real part of \xi_ee
• Perhaps add e1e1, e2e2, e1e2, e2e1?

I will look into all of these. (I'm not sure if the e1e1, e1e2, etc is possible with TreeCorr, but the others are all definitely doable.) Any other plotting choices that have been bugging anyone while I'm working on this?

[rmandelb]

These sound good. Re: e1e1 etc. - open feature request on TreeCorr issue page?

[msimet]

I'll check if it's possible first--I'm not sure all of what's available as TreeCorr internals (or what we could get through linear combinations of what's available...). But otherwise, can open an issue, yes.