adwasser / slomo

Python bindings for Slomo.jl

slomo

Python bindings for Slomo.jl

Dependencies

You'll have to install pyjulia to get Python and Julia to play nice with each other.

Installation

1. Download and install Julia

   Link to Julia site

2. Clone this (slomo) repository

   We need to also copy over the Slomo.jl submodule by using the --recurse-submodules option.

   git clone --recurse-submodules https://github.com/adwasser/slomo.git

3. Download and install Julia dependencies

   From the shell:

   cd slomo/slomo/Slomo.jl
   julia

   From the Julia REPL:

   using Pkg
   Pkg.activate(".")
   Pkg.instantiate()
   Pkg.resolve()

4. Build a Julia system image for faster load times (optional)

   This will take a long time. Go make a cup of tea and read a book while you wait. The benefit is that the startup time should be much faster if you have a precompiled Julia system image.

   Run the build script in the Slomo.jl directory:

   cd slomo/slomo/Slomo.jl
   mkdir build
   julia build.jl

5. Install the python package

   From the top-level package directory:

   pip install .

If all goes well, you should be able to do the following from a Python shell:

import numpy as np
from slomo import *
model = JeansModel(halos.NFWModel(), SersicModel(), ConstantBetaModel())
R = np.logspace(-1, 1)
sigma_profile = sigma_los(model, R)

The first call of sigma_los will be incredibly slow due to the JIT compilation from Julia; subsequent calls should be much faster.

Multi-threading

To set up multi-threading with Julia, set the JULIA_NUM_THREADS environmental variable to the desired number of threads prior to importing slomo, e.g. for bash

export JULIA_NUM_THREADS=4

To do the Jeans integration for a single model structure with multiple parameter sets, pass a list of dictionaries to sigma_los_parallel, where each dictionary contains parameter updates for the model.

parameter_sets = [
	dict(beta=0.5, Re=12.3),
	dict(beta=0.1, Re=12.3),
	dict(beta=0.5, Re=10.9),
	dict(beta=0.1, Re=10.9)
]
sigma_profiles = sigma_los_parallel(model, R, parameter_sets)