GALLUMI: A Galaxy Luminosity Function Pipeline for Cosmology and Astrophysics

Observations of high-redshift galaxies have provided us with a rich tool to study the physics at play during the epoch of reionisation. The luminosity function (LF) of these objects is an indirect tracer of the complex processes that govern galaxy formation, including those of the first dark-matter structures. In this work, we present an extensive analysis of the UV galaxy LF at high redshifts to extract cosmological and astrophysical parameters. We provide a number of phenomenological approaches in modelling the UV LF and take into account various sources of uncertainties and systematics in our analysis, including cosmic variance, dust extinction, scattering in the halo-galaxy connection, and the Alcock-Paczy\'{n}ski effect. Using UV LF measurements from the Hubble Space Telescope together with external data on the matter density, we derive the large-scale matter clustering amplitude to be $\sigma_8=0.76^{+0.12}_{-0.14}$, after marginalising over the unknown astrophysical parameters. We find that with current data this result is only weakly sensitive to our choice of astrophysical modelling, as well as the calibration of the underlying halo mass function. As a cross check, we run our analysis pipeline with mock data from the IllustrisTNG hydrodynamical simulations and find consistent results with their input cosmology. In addition, we perform a simple forecast for future space telescopes, where an improvement of roughly 30% upon our current result is expected. Finally, we obtain constraints on astrophysical parameters and the halo-galaxy connection for the models considered here. All methods discussed in this work are implemented in the form of a versatile likelihood code, GALLUMI, which we make public.