planck_py

Lightweight python-based Planck likelihood code

File	likelihood/planck_py/planck_py_interface.py
Attribution	Heather Prince, Jo Dunkley (Library)
	CosmoSIS Team (Interface)
URL	https://github.com/heatherprince/planck-lite-py
Citations	Phys. Rev. D 100, 083502
	https://doi.org/10.1051/0004-6361/201936386
	https://doi.org/10.1051/0004-6361/201526926

The lite version of the Planck likelihoods can be more useful than the main version both because it is purely Gaussian and so can be easily used in Fisher matrix calculations, and because it is much cleaner and easier to understand.

Assumptions

• Planck 2015 or 2018 data

• Compresed form of Planck likelihoods

Setup Parameters

Name	Type	Default	Description
year	int		Whether to use 2015 or 2018 data
spectra	str		Which set of spectra to use, TT or TTTEEE
use_low_ell_bins	bool		Whether to include ell < 30 data points
data_directory	[module dir]/data		Directory to search for data
use_data_from_test	str		Set to a directory generated by the CosmoSIS test sampler to use it as mock observations instead of true Planck data
x_name	str	ell	Datablock name for CMB ell values
y_section	str	cmb_cl	Datablock section for CMB spectra
like_name	str	planck	Name for the saved output likelihood
likelihood_only	bool	False	Skip saving everything except the likelihood. This prevents you from using e.g. the Fisher matrix sampler but can be faster for quick likelihoods
include_norm	bool	False	Include the normalizing constant at the start of the likelihood. May be needed when comparing models.

Input values

Section	Name	Type	Default	Description
cmb_cl	ell	real 1d		CMB power spectrum ell values of other inputs
	TT	real 1d		CMB temperature power spectrum in ell(ell+1)/2pi units
	TE	real 1d		CMB temperature-E-mode polarization cross power spectrum in ell(ell+1)/2pi units
	EE	real 1d		CMB E-mode polaration power spectrum in ell(ell+1)/2pi units

Output values

Output values

Section	Name	Type	Description
likelihoods	planck_like	real	Planck likelihood for given selection
data_vector	planck_covariance	real 2d	Fixed covariance matrix, only if likelihood_only=F
	planck_data	real 1d	Concatenated D_ell, only if likelihood_only=F
	planck_simulation	real 1d	Simulated D_ell data vector including simulated noise for e.g. ABC, only if likelihood_only=F
	planck_theory	real 1d	Predicted theory values D_ell only if likelihood_only=F
	planck_chi2	real	chi^2 value, only if likelihood_only=F