Codes from the KASI research

PyVI: visual inspection of astronomical data

This Python code is for the visual inspection of image files and leaving related logs for the images in a file. For a given list of images, it shows images and ask people to leave any logs that can be saved in a file. The Python Image Library and Tkinter are required with this code. A user can put comments on images through an interface based on Python curses module.
Download >> and documents >>.

Python module example:
>>from PyVI import image_db
>>from PyVI import curses_interface
>>infn="list.txt"
>>outfn="log.txt"
>>dir_images="./Images/"
>>dir_texts="./Texts/"
>>db = image_db(infn, outfn, dir_images, dir_texts)
>>use_fits = 1 # if you want to use FITS images with DS9.
>>use_fits = 2 # if you want to use FITS images with Aladin.
>>use_fits = 3 # if you want to use FITS images with both DS9 and Aladin.
>>import sampy
>>db_inf = curses_interface(db, infn)

Command-line options:
PyVI.py [list filename] [log filename] 
        --ds9  --aladin
        --image_dir [the directory name of image files]
        --text_dir [the directory name of text files]
Example) PyVI.py sel.list log.list --ds9 --image_dir ./Gallery --text_dir ./Extra_text

        --ds9 : PyVI will use the DS9 as an image viewer which supports FITS files. 
        You must check the DS9 first whether it is available with the SAMP.

        --aladin : PyVI will use the Aladin as an image viewer which supports 
        FITS and other file formats such as JPEG and PNG. You must check the 
        Aladin first whether it is available with the SAMP. If both --ds9 and 
        --aladin options are selected, both viewers are used in the PyVI.

        --general : PyVI will use a general FITS image viewer including JSky, 
        DS9, and Aladin. You must check the viewer first whether it is available 
        with the SAMP. However, do not use --general with either --ds9 or --aladin 
        together. This option is specially for FITS viwer showing only a single frame. 
        Therefore, this option is appropriate for tools like JSky.

        --spec : PyVI will use a spectrum viewer (e.g. SPLAT-VO, VOSpec). 
        You must check the viewer first whether it is available with the SAMP.

        --image_dir, --text_dir, --spec_dir : the directory names are optional
        if the directory names are not given, the current working directory 
        is a default directory.

MS_Period: Multi-step approach to find periods of time-series data

We developed an algorithm to identify and determine periods of variable sources. With its robustness and high speed, it is expected to become an useful tool for surveys with large volume of data. This new scheme consists of an initial coarse process of finding several candidate periods followed by a secondary process of much finer period search. With this multi-step approach, best candidates among statistically possible periods are produced without human supervision and also without any prior assumption on the nature of the variable star in question.
Download >> and documents >>.

Virtual Observatory tools

Severl tools have been developed and used by KASI scientists to access the VO data and services. The following is the list of the tools and their links for downloading.

Galaxy mask generator

The codes presented in this link are developed to produce a mask that represent pixel dominated by signals from galaxies above a certain threshould pixel level. A newly proposed method to define the threshould level uses the Gaussian Mixture Model (GMM) with the assumption that thre most dominant mixture component corresponds to the distribution of not source but background pixels. A final object mask is defined to be the union of masks-per-band (i.e., the outcome of OR operation among the masks-per-band). The proposed method is implemented for the Pan-STARRS1 stack images (https://outerspace.stsci.edu/display/PANSTARRS/PS1+Stack+images), and it also assumes that a target galaxy is centered in given images.

MBRNN (Multiple-Bin Regression with Neural Networks) for photometric redshifts

We developed a machine learning model that uses neural networks to estimate photometric redshifts of galaxies. Depending on the estimation range of redshifts, this model based on neural networks can handle the difficulty for inferring photometric redshifts. Moreover, to reduce bias induced by the new model's ability to deal with estimation difficulty, it exploits the power of ensemble learning.
Visit the code repository and check the relevant paper.

MBRNN OOD evaluation for photometric redshifts

We release codes that implements a multi-stage training strategy of the MBRNN for out-of-distribution (OOD) evaluation.
Visit the code repository and check the relevant paper.

LaRT (Lyman-alpha Radiative Transfer)

LaRT is a three-dimensional Monte Carlo Lyα radiative transfer code, which is developed to study the Lyα radiative transfer. LaRT is capable of predicting (1) the Lyα spectrum and surface brightness profile, (2) the Wouthuysen-Field effect, and (3) the polarization of Lyα radiation. The code is capable of treating arbitrary geometries, density distributions and source distributions. LaRT uses the "peeling-off" (next event estimation) technique to obtain high signal-to-noise spectro-polarimetric images in a detector plane. LaRT is superb, compared to the preexisting codes, in that it uses a smoothly and seamlessly varying phase function as frequency changes to deal with the polarization of Lyα radiation. LaRT can place an observer in an arbitrary location and make a detector plane to have an arbitrary orientation in the sky. The current version of LaRT uses a Cartesian grid to model the density distribution of hydrogen gas. LaRT has been benchmarked for a number of standard cases.
Visit the code repository and check the relevant papers: 1) Lyα Radiative Transfer: Monte Carlo Simulation of the Wouthuysen-Field Effect, 2) Ly-alpha Radiative Transfer: A Stokes Vector Approach to Ly-alpha Polarization.