feat: add solar wind dm gp

[jeremy-baier]

Copying over from enterprise extensions.

Adds Fourier basis solar wind dm gp and time domain sw gp with a linear interpolation.

Bug fix (taken from Pat) for deterministic solar wind.

[jeremy-baier]

preliminary testing show that the Fourier basis SWGP is >2x faster overall than the time independent deterministic solar wind gp.
(comparison is single likelihood evaluation time using a fixed white noise + varied dmgp model + varied solar wind model)

[meyers-academic]

I don't think this is necessary. I think the np.sinc in make_solardm should be faster than using a jnp.where. Note that the definition of np.sinc is sin(pi * x) / (pi * x) hence the 1 - theta / pi in the argument. But make_solardm above should give back the You 2007 function.

[davecwright3]

@jeremy-baier @meyers-academic I've been testing this PR out with some slight modifications to fourierbasis_solar_dm:

def fourierbasis_solar_dm(psr,
                          components,
                          T=None):
    """
    From enterprise_extions: construct DM-Solar Model Fourier design matrix.
    :param psr: Pulsar object
    :param components: Number of Fourier components in the model
    :param T: Total timespan of the data
    :return: F: SW DM-variation fourier design matrix
    :return: f: Sampling frequencies
    """

    # get base Fourier design matrix and frequencies
    f, df, fmat = fourierbasis(psr, components, T)
    dm_sol_wind = make_solardm(psr)(1.0)

    return f, df, fmat * dm_sol_wind[:, None]

Solar wind can now be put in commongp for the ArrayLikelihood. Example:

commongp = [                                                                           
    ds.makecommongp_fourier(                    
        psrs,                                   
        ds.powerlaw,                            
        30,                                     
        fourierbasis=ds.fourierbasis_solar_dm,  
        T=Tspan ,
        common=["sw_gp_log10_A", "sw_gp_gamma"],
        name="sw_gp",                           
    ),                                          
]                                               

[jeremy-baier]

i think this branch was on my old fork of discovery. i am going to close this and open a new PR. see #98