Halofit: fnu correction

[jecampagne]

The present halofit code (power.py) does not introduce any fnu correction. Notice that this correction is coded differently in CCL and CLASS. Here is the CLASS comment

     pk_halo = a*pow(y,f1*3.)/(1.+b*pow(y,f2)+pow(f3*c*y,3.-gam));
      pk_halo=pk_halo/(1+xmu*pow(y,-1)+xnu*pow(y,-2))*(1+fnu*0.977);

      /* until v2.9.3 pk_halo did contain an additional correction
         coming from Simeon Bird: the last factor was
         (1+fnu*(0.977-18.015*(pba->Omega0_m-0.3))). It seems that Bird
         gave it up later in his CAMB implementation and thus we also
         removed it. */
      // rk is in 1/Mpc, 47.48and 1.5 in Mpc**-2, so we need an h**2 here (Credits Antonio J. Cuesta)

CCL implements the (1+fnu*(0.977-18.015*(pba->Omega0_m-0.3))) factor.

[jecampagne]

This fnu correction is related to the question of introducing neutrino contributions is jax-cosmo.