Installing via pip¶
juliet can be easily installed using pip:
pip install juliet
The core of
juliet is comprised of the transit (batman,
starry), radial-velocity (radvel)
and Gaussian Process (george,
celerite) modelling tools, as well as
of the Nested Sampling algorithms (MultiNest via pymultinest,
dynesty) that it uses. However, by default the ``juliet`` installation will
force `dynesty` as the main sampler to be installed, and will not install `pymultinest`. This is because
pymultinest installation can involve a couple of extra steps, which we really recommend following, as
pymultinest might be faster for problems involving less than about 20 free parameters (see below).
Installing from source¶
The source code for
juliet can be downloaded from GitHub by running
git clone https://github.com/nespinoza/juliet.git
Once cloned, simply enter the
juliet folder and do
python setup.py install
To install the latest version of the code.
As described above, we really recommend installyng
pymultinest. The full instructions on how to install
pymultinest can be found in the project’s documentation.
We repeat here the main steps. First, install it via pip:
pip install pymultinest
Then, you need to build and compile MultiNest. For this, do:
git clone https://github.com/JohannesBuchner/MultiNest cd MultiNest/build cmake .. make
This will create a file
MultiNest/lib: that is the one that will allow us
pymultinest. To not move that file around in your system, you can include the
MultiNest/lib folder in your
LD_LIBRARY_PATH (e.g., in your
~/.bashrc file). In my case, the library is under
/Users/nespinoza/github/MultiNest/lib, so I added the following line to my
The above installation instructuins for
juliet assume you have a Python installation.
juliet, in turn,
depends on the following libraries/packages, all of which will be installed automatically if you follow the instructions
- pymultinest (optional),
- matplotlib (optional), and
- seaborn (optional).
The last are only needed for certain plotting functions within
pymultinest installation is optional, but highly recommended.