COLIBRÌ (which roughly stands for 'Cosmological Libraries') is a set of Python files containing useful routines to compute cosmological quantities such as ages, distances, power spectra and correlation functions. It supports Lambda-CDM cosmologies plus extensions including massive neutrinos, non-flat geometries, evolving dark energy (w0-wa) models and numerical recipes for f(R) gravity. These files are built especially for large-scale structure purposes.
It is compatible with both Python 2 and 3 (soon Python 2 compatibility will be deprecated). Furthermore, it can interact with the Boltzmann solvers CAMB (by Antony Lewis and Anthony Challinor) and Class (by Julien Lesgourgues).
Please, cite this repository ( https://github.com/GabrieleParimbelli/COLIBRI ) if you use this library.
FYI, "colibrì" is the Italian for "hummingbird".
COLIBRÌ works properly provided that the usual Python packages such as NumPy and SciPy are installed. However, some of the routines require some external libraries and Boltzmann solvers to be installed. The list that follows itemizes all the packages needed to let all the routines work.
- FFTLog by Dieter Werthmüller. This package is a Python wrapper for the FFTLog by Andrew Hamilton and allows to compute Fourier transforms of a log-spaced array. Download FFTLog and unzip the folder. When in the folder compile with
python setup.py install --user(the system may not recognize the ü on the name of the author, if so just remove it). - Cython: C compiler for Python. Its presence is needed for the compilation of CLASS (see below). The installation can be done both using
pip install Cythonor from source, downloading it from here, unzipping the folder, go into it and then runningpython setup.py install --user. Important: if one follows the latter way, the folder cython-master contains another folder named in the same way. This must not happen, since the paths may not be recognized. Just move everything in the outer of the two folders and delete the inner one. - the Python wrapper of the Cosmic Linear Anisotropy Solving System (Class) by Julien Lesgourgues. Download Class, untar and compile with
makeonly (notmake classotherwise the Python wrapper will not work). Go now to the python folder inside Class and runpython setup.py install --user. This will build the classy file: check that everything has gone fine by typing in terminalpython -c "from classy import Class". If nothing is returned then everything is ok. - the Python wrapper of the Code for Anisotropies in the Microwave Background (CAMB) by Antony Lewis and Anthony Challinor. Clone CAMB or install it using pip. In order for COLIBRÌ to work, a version > 1.0 of CAMB is required. Notice that to install this, a gfortran compiler > 6 is needed.
The importation of these packages is always conditional in COLIBRÌ: if they are not found they are not imported. Thus, no error will be returned until a function that requires some of them is called. However, we strongly recommend to install all these 4 packages (for convenience and future employment of the user).
Here we report briefly what the various packages are needed for.
- FFTLog is needed to use the Fourier and Hankel transforms for log-spaced arrays in
fourier.py. - CAMB and/or Class are needed to compute power spectra given a cosmological model. If you wish, you can install only one of the two, although it is a good thing to have them both also for comparisons.
- Cython is just needed to compile the Python wrapper of Class: thus if you do not want to install Class, Cython is not necessary as well.
The best way to install COLIBRÌ is to clone the GitHub repository https://github.com/GabrieleParimbelli/COLIBRI , enter in the COLIBRI directory and run:
python setup.py install --user
A second way consists of cloning the repository and add the libraries to your PYTHONPATH editing your .bashrc file.:
export PYTHONPATH="${PYTHONPATH}:/path/to/COLIBRI/"
So, set your preferences and get started!
COLIBRÌ consists of simple Python files: one, named constants.py is just a collection of physical constants, useful numbers and conversion factors between different units of measurement. A second one, useful_functions.py contains some routines of everyday life for a cosmologist. The remaining ones each contain one or more Python classes. Moreover two folders (tests and notebooks) containing Python files and notebooks are provided with the code.
In the following sections we briefly present the codes. The documentation for both classes and built-in methods is inside the code as well as on ReadTheDocs .
The main file of COLIBRÌ is without doubt the cosmology.py file.
This file contains one Python class, named cosmology.cosmo(), in whose initialization the cosmological parameters must be provided (otherwise the default values will be loaded).
This class has all the routines necessary to compute distances, ages, density parameters, halo mass functions, void size functions and power spectra.
For the latter, both Boltzmann solvers CAMB and Class can be used, provided their Python wrapper is correctly compiled, as well as the Eisenstein-Hu formula (see arXiv:9710252).
Included here are several different ways to compute the non-linear matter power spectrum from a linear one.
These methods include HMcode2016 by Mead et al. (see arXiv:1505.07833 ), HMcode2020 by Mead et al. (see arXiv:2009.01858 ) Takahashi and Bird following the prescription by Takahashi et al. for the former (see arXiv:1208.2701 ) and the correction for the presence of massive neutrinos by Bird et al. for the latter (see arXiv:1109.4416 ).
The limber class inside this file is finalized to compute angular power spectra and correlation functions in the flat sky and Limber's approximations. In this file are provided the routines to compute example functions for galaxy distributions (although different ones can be defined by the user outside the class) and window functions.
This three files are linked with each other. The basis is halo.py : it contains the class halo which computes the non-linear matter power spectrum according to the pure halo model (see for instance Cooray & Sheth (2001)).
While this is known to return a poor description of the matter clustering, the class has routines able to compute properly halo mass functions and halo biases.
In the file galaxy.py the class galaxy is implemented, which uses the Halo Occupation Distribution (see e.g. arXiv:0408564 ) prescription to predict the galaxy power spectrum in real space.
Conversely, the redshift-space power spectrum is provided by the class RSD in the file RSD.py: currently the dispersion model is implemented (with both Gaussian and Lorentzian dampings) as well as a halo model based prescription.
This file contains routines to compute Fourier and Hankel Transforms. They employ the NumPy FFT libraries as well as FFTlog in some cases. They return sorted frequencies for an immediate interpretation of the outcomes. In particular, these routines can be useful to compute two-point correlation functions starting from a power spectrum.
This file is just a compilation of physical constants and does not contain any class or method. While typing help(constants) will provide the list of quantities, it will not be documented. To obtain a full description of the quantities, type in a Python session or program:
import constants constants.explanatory()
The file contains (as is obvious) useful functions such as extrapolation of arrays and top-hat window functions.
Together with the files, a folder named tests and another called notebooks containing some useful and explanatory tests are provided. Each of them is adequately commented, so check them out and run them!