A unified toolkit for extracting cosmic signals from noisy sphere data

CosmoForge unifies power spectrum estimation and likelihood analysis—the two main ways cosmologists squeeze information from CMB maps—into a single Python toolkit. The framework intelligently switches between harmonic and pixel-space calculations depending on which is cheaper for a given problem, and includes algorithmic tricks that cut computational cost to O(ℓ_max^4). It reproduces Planck's official Fortran results exactly, works with arbitrary noise patterns, and scales via MPI parallelization.