Installation¶
CCL can be installed from pip
, conda
, or directly from source.
It is configured to install most of its requirements automatically. However, if
you want to use CCL with Boltzmann codes like CLASS
or CAMB
, you will
need to make sure the Python
wrappers for these packages are installed
as well. See the instructions for Getting a Boltzmann Code below.
Once installed, you can take CCL for a spin by following the Quickstart instructions.
With conda¶
Installing CCL with conda
will also install a copy of pycamb
, the
Python
wrapper for CAMB
in your environment.
$ conda install -c conda-forge pyccl
With pip¶
The PyPi installation will actually build a new copy of CCL from source as
the code is installed. In order to do this, you will need CMake
installed
on your system. See the instructions below for Getting CMake.
Once you have CMake
, simply run:
$ pip install pyccl
Getting a Boltzmann Code¶
In order to use CCL with a Boltzmann code, you will need the Python
wrappers
for either CLASS
or CAMB
.
CLASS¶
In order to use CCL with CLASS
, you need to install the CLASS
Python
wrapper classy
. See their instructions
here.
Note that you may need to edit the Makefile
in the CLASS
repo to work
with your system. Please report any installation issues to the CLASS
developers
on their issues tracker.
Getting CMake¶
The pip
installation of CCL requires that CMake
is installed on your
system. CMake
is available from package managers like apt-get
and
homebrew
. You need version 3.2 or higher.
Ubuntu¶
$ sudo apt-get install cmake
Known Installation Issues¶
For some Mac OSX versions, the standard
C
headers are not in the usual spot, resulting in an error offatal error: 'stdio.h' file not found
. This can be resolved with the command:$ sudo installer -pkg /Library/Developer/CommandLineTools/Packages/macOS_SDK_headers_for_macOS_10.14.pkg -target /
which will install all the required headers into
/usr/include
.
Uninstalling CCL
¶
CCL
can be uninstalled using the uninstallation functionality of the
package manager (i.e., conda
or pip
) you used to install it. When in doubt,
first try with conda
and then try with pip
. In either case, the command is
$ [conda|pip] uninstall pyccl