This is a python re-implementation of the Jeremy Sanders contour binning algorithm, which was originally in c++. You can install it with pip install pycontbin, or upgrade it with pip install --upgrade pycontbin. Please note that this python build is under development and is subject to updates.
You can see the original files here: https://github.com/jeremysanders/contbin
Original Copyright Jeremy Sanders [email protected] (2002-2016) The reference paper is Sanders (2006), MNRAS, 371, 829, http://adsabs.harvard.edu/abs/2006MNRAS.371..829S
This version was translated to python by me, and might still contain some minor bugs, although the current implementation results seem to reproduce the original code results.
I have additionally added a routine that automatically produces and outputs polygon region files for each bin in sexagesimal coordinates, this is an improvement on the previous work that output many box regions instead.
The jupyter notebook contains the testing code, also an example use case.
This software is licensed under the GNU Public License.
To do:
- I want to implement PSF information in the code. I will be trying to work on this on another branch.
- Adding GPU optimization would be wonderful.
Done:
- Added polygon regions as output in sexagesimal coordinates, however this requires ciao and pyds9 installed.
- Created pip package
- numpy
- astropy
- ciao
- pyds9
# from contour_binning import *
from pycontbin import ContourBin
os.chdir("/Users/jpbreuer/Scripts/contbin-python")
inputfile = "scaled_fluxed_sps_filth_fov.fits"
sn_ratio = 50
smooth = 30
constrain_val = 1.5
reg_bin = 1
ContourBin(inputfile, sn_ratio, smooth, constrain_val, reg_bin, make_region_files=True)