MGLensing

Repository containing a likelihood (initially based on this pipeline) for photometric probes with emulated nonlinear power spectra and various other emulators.

The purpose of this code:

• Perform a quick and light MCMC analysis for Stage IV surveys, which is useful for forecasts, scale-cuts, and studies of projection effects.
• Without inclusion of systematic effects on large scales, we focus on systematics on small/nonlinear scales for standard and extended cosmologies: nonlinear galaxy bias expansion, nonlinearities in the matter power spectrum, baryons.
• Quick analysis of N-body simulations.

This code is not complete to perform a real data analysis (yet), but is a fairly realistic and good-enough in-between solution.

Available Models

• BACCO-emulator and HMcode2020-emulator:
  • $\Lambda$ CDM, $w$ CDM and $w_0w_a$ CDM with neutrinos.
• ReACT-based emulators:
  • normal branch of DGP gravity (nDGP);
  • Hu-Sawicki f(R) gravity;
  • growth index parametrisation and time-dependent growth index with Screening;
  • mu-Sigma (scale-independent and scale-dependent) with DE time-evolution with Screening;
  • Interacting Dark Energy also known as Dark Scattering.
• Simulation-based emulators:
  • eMANTIS for f(R) gravity;
  • nDGPemu for nDGP.
• Galaxy bias:
  • Linear;
  • Phenomenalogical $b(z, k) = b_1(z) + b_2(z) k^2$;
  • Hybrid Lagrangian bias epxansion (HEFT) with BACCO-emulator.
• Baryons:
  • HMcode2020;
  • BCemu;
  • BACCO-emulator.
• Intrinsic alignment:
  • extended redshift-dependent nonlinear alignment (e-zNLA);
  • tidal alignment and tidal torquing (TATT).
• Photo-z uncertainties:
  • additative;
  • multiplicative.

How to run

Modify "config.yaml" file and specify

• Observables;
• Specifics of your survey;
• Modelling for your mock synthetic data;
• Modelling for your theoretical predictions;
• File-paths with fiducial data points (e.g., "params_data.yaml") and priors (e.g., "params_model_hmcode.yaml");
• Output-file name.

You can then run a test likelihood computation:

import MGLensing

folder = os.path.dirname(os.path.abspath(__file__))
os.chdir(folder)

MGL = MGLensing.MGL("config.yaml")
MGL.test() 

Following files provide examples of using MGL:

• test_mcmc_run.py: run an MCMC chain with Nautilus sampler;
• notebooks in the tutorials-folder

Documentation

The documentation can be found at mglensing.readthedocs.io.