GravImage is a tool to determine the mass distribution in one-dimensional disc-like or spherical systems. It takes as input a tracer density distribution, a line-of-sight velocity dispersion, and possibly the velocity's fourth moment as a function of radius. It then generates a highly-dimensional parameter space for tracer density, overall density distribution, and a velocity anisotropy profile in bins and samples it with MultiNest, which is a specialized high-dimensional parameter space sampling algorithm. The Jeans equations for the systems under study is then used to calculate a goodness-of-fit from the surface density and velocity dispersion along the line of sight. The accepted models are visualized in a later step.
Following packages need to be installed on your system:
- python3
- matplotlib/pylab
- scipy
- ipdb, pdb
Then execute
git clone https://github.com/PascalSteger/darcoda $DARCODA_DIR
cd $DARCODA_DIR
or unzip the file darcoda.zip to $DARCODA_DIR . Then set the environment variables
export PYTHONPATH=$PYTHONPATH:$DARCODA_DIR/gravimage/programs/ export PYTHONPATH=$PYTHONPATH:$DARCODA_DIR/gravimage/programs/plotting/
Adapt the path specifications to your needs in
gl_params.py gl_class_files.py import_path.py
and run
python3 gravimage.py
Sample parameter files are stored in the subfolders disc/ and sphere/. The file ./gl_params.py is a soft link to one of them. Following mass modelling methods have been implemented so far:
- hern: spherical mock data taken from a Hernquist profile
- gaia: spherical mock data from the Gaia challenge catalogue, 1 population
- walk: spherical Walker mock data from the Gaia challenge catalogue, 2 populations
- obs: observations of 4 dwarf spheroidals
- discmock: disk-like mock data, generated on the fly
- discsim: disk-like mock data, from a simulation by S. Garbari
Additional documentation on profile representation, priors, runtime parameters and more can be found in the paper draft folder, and in the code itself.
A HTML documentation of all files and functions can be generated using
doxygen Doxyfile
and then browsing to doc/html/.
The code is under constant development, and might show bugs. Feel free to branch the code, correct them, and send a patch!
January 2015, Pascal Steger psteger@phys.ethz.ch http://steger.aero