Installation
============
Install the latest stable release using::
pip install randomfield
Alternatively, you can install the latest developer version from github::
github clone https://github.com/dkirkby/randomfield.git
cd randomfield
python setup.py install
Optional Dependency: matplotlib
-------------------------------
The ``matplotlib`` package enables optional plotting capabilities and should
already be installed if you are using the recommended `anaconda scientific
python distribution `__.
Optional Dependency: classy
---------------------------
The ``classy`` package provides python bindings to the C++ `CLASS package
`__ and enables power spectra to be calculated
for arbitrary cosmologies. If you do not install ``classy``, you can still
generate fields for arbitrary cosmology using an external program such as
`CAMB `__ to write tabulated power spectra to files.
Optional Dependency: pyFFTW
---------------------------
The `pyFFTW package `_ provides a wrapper around the `FFTW C library `_. Installation of
pyFFTW is straightforward but requires that multiple versions of the FFTW library are already installed.
Install the float/double/quad and single/threaded versions of the FFTW library using each of the following configurations::
# Single precision
./configure --enable-sse --enable-float --enable-shared
./configure --enable-sse --enable-float --enable-shared --enable-threads
# Double precision
./configure --enable-sse2 --enable-shared
./configure --enable-sse2 --enable-shared --enable-threads
# Long-double precision
./configure --enable-long-double --enable-shared
./configure --enable-long-double --enable-shared --enable-threads
After entering each `configure` command above, build and install the corresponding library with::
make
sudo make install
Install the `pyFFTW package`_ using::
pip install pyfftw