149 add nonlinear galaxy bias models

[anicola]

This PR implements a wrapper to CCL nonlinear bias models. For the moment, it's just EPT and LPT but more models can be added once they are in CCL.

[codecov-commenter]

Codecov Report

Merging #164 (39ad7f0) into master (083154c) will increase coverage by 0.50%.
The diff coverage is 86.77%.

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #164      +/-   ##
==========================================
+ Coverage   74.47%   74.98%   +0.50%     
==========================================
  Files          33       33              
  Lines        2018     2123     +105     
==========================================
+ Hits         1503     1592      +89     
- Misses        515      531      +16     

Files	Coverage Δ
soliket/ccl.py	94.11% <100.00%> (+0.27%)	⬆️
soliket/cross_correlation/cross_correlation.py	79.28% <86.32%> (+3.44%)	⬆️

[itrharrison]

@anicola @mgerbino are we happy with the robustness of using the numerical value in converting omnu to mnu?

or we could calculate it from the relevant constants as in camb?

[mgerbino]

@itrharrison the two numbers are basically the same once you rescale the consts in camb for the non-instantaneous nu decoupling, so no specific issue there.
However, what makes me prefer the camb approach is that, should one run with a different Tcmb than the COBE value or with a non-zero nu sterile contribution (i.e., exploring non-standard cosmologies), the current approach in soliket/pyccl does not capture those differences. Is this anything we should be worried about?