Reconstruction with redshift evolution

[epaillas]

Hi Arnaud,

Here is my implementation of the redshift evolution for the growth rate and bias, as well as the optimal weighting scheme.

For the redshift evolution, the idea is that the user needs to provide a callable that returns the growth rate or bias as a function of the coming radial distance to redshift z, when initializing the reconstruction algorithm.

def interpolate_growth_bias():
    z = np.linspace(0.0, 5.0, 10000)
    growth_rate = cosmo.growth_rate(z)
    bias = bias_evolution(z)
    growth_at_dist = InterpolatedUnivariateSpline(distance(z), growth_rate, k=3)
    bias_at_dist = InterpolatedUnivariateSpline(distance(z), bias, k=3)
    return growth_at_dist, bias_at_dist
    
    
recon = IterativeFFTReconstruction(f=growth_at_dist, bias=bias_at_dist, ...)

This creates internal meshes that encode the growth rate, bias, and optimal weights, for any given cell in the survey volume.

Feel free to tweak this as you wish or let me know if there's any modification you'd want me to implement.

Enrique