GalPaK v1.9



GalPaK 3D is a tool to extract the intrinsic (i.e. deconvolved) Galaxy Parameters and Kinematics from any 3-Dimensional data. The algorithm uses a disk parametric model with 10 free parameters (which can also be fixed independently) and a MCMC approach with non-traditional sampling laws in order to efficiently probe the parameter space.

More importantly, it uses the knowledge of the 3-dimensional spread-function to return the intrinsic galaxy properties and the intrinsic data-cube. The 3D spread-function class is flexible enough to handle any instrument.

One can use such an algorithm to constrain simultaneously the kinematics and morphological parameters of (non-merging, i.e. regular) galaxies observed in non-optimal seeing conditions. The algorithm can also be used on AO data or on high-quality, high-SNR data to look for non-axisymmetric structures in the residuals.


You can download the paper about galpak, authored by N. Bouché, H. Carfantan, I. Schroetter, L. Michel-Dansac, T. Contini.


2015-05-14 ∾ Easier setup and doit implementation

Since v1.5.0, galpak supports doit tasks and python install.

2014-12-08 ∾ Support for ALMA

Since v1.3.0, galpak can now work with ALMA data.

Set the lsf_fwhm to something less than 1 channel.

2014-11-01 ∾ Generate videos

Since v1.2.0, galpak can now generate a video of your run.

You'll probably need a very recent version of ffmpeg, or the video generation will fail after about 820 frames. This is for advanced users only, for now.

See the docstring of galpak.GalPaK3D.film_images for more informations.

2014-10-13 ∾ Galpak gains in flexibility

Since v1.1.1, galpak.GalaxyParameters are now much more flexible.

You can add new parameters and see them evolve through the MCMC. This flexibility comes at a small performance cost, which we tried to reduce as much as we could.


The following video was made using these parameters : from galpak import GalPaK3D, MUSEWFM g = GalPaK3D('hyperspectral_cube.fits', seeing=0.7, instrument=MUSEWFM(lsf_fwhm=2.519/1e4)) g.run_mcmc(max_iterations=10000)

Hyperspectral Cube Visualization