Quasar luminosity function doing strange things in the edges

[andreufont]

@julienguy - take a look at examples/plot_qsoLF.py

The new file nzr_qso.dat does strange things at the faint end, and at high-z, probably because of the way we are interpolating / extrapolating.

It would be good to regularize it, otherwise it might impact our BAO forecast.

[andreufont]

I added a comparison of both files in plot_qsoLF.ipynb

The differences are large, and the BAO errors as a function of z are quite different, but somehow the integrated BAO errors are actually quite similar.