Speed up Numba-fied Point Source analytic Fourier Transform

[ronniyjoseph]

Beam-Perturbations/gain_variance_simulation.py

Line 271 in 2786c96

def get_observations_numba(source_population, baseline_table, frequency_range):

Currently Numba iterates over the source, not optimally using the memory I have available. It does one source, per baseline per frequency. It might be worthwile to investigate more optimal was of doing this and properly utilising 60 GB RAM available.