Author

Sami-Matias Niemi

Contact

s.niemi@icloud.com

issue

1.2

version

1.4

date

5th of March 2015

This Python package VIS-PP provides subpackages and methods related to generating mock data and reducing it, the main consideration being the visible instrument (VIS) that is being developed for the Euclid telescope. The subpackages include methods to e.g. generate object catalogues, simulate VIS images, study radiation damage effects and fit new trap species, reduce and analyse data, and to include instrumental characteristics such as readout noise and CTI to "pristine" images generated with e.g. the GREAT10 photon shooting code. In addition, an algorithm to measure ellipticities of galaxies is also provided. Thus, this package tries to provide an end-to-end simulation chain for the VIS instrument.

The documentation is also available in PDF format, please download the VIS Python Package Manual.

The VIS Python Package (VIS-PP) is held in a GIT repository. The package contains a mixture of classes and scripts divided in subpackages based on the usage. Unfortunately, there is no official or preferred installation instructions yet. To get most scripts working you should place the path to the root directory of the package to your PYTHONPATH environment variable. In addition, it is useful to compile the Fortran code available in the fortran subdirectory with the following command:

f2py -c -m cdm03 cdm03.f90

and then copy the .so file to the CTI directory. Please note that f2py is available in the NumPy package, but you still need for example gFortran compiler to actually compile the Fortran source.

The VIS Python package depends heavily on other Python packages such as NumPy, SciPy, PyFITS, and matplotlib.

The support subpackage contains functions that define the VIS instrument model. This model contains information about noise, dark, cosmic rays, radiation damage parameters, sky background, and pixel scale.

The postproc subpackage contains methods related to either generating a CCD mosaics from simulated data that is in quadrants like the VIS reference simulator produces or including instrument characteristics to simulated images that contain only Poisson noise and background.

The package provides a simple exposure time calculator (ETC) that allows to estimate a signal-to-noise ratio of an average galaxy or star with a given number of VIS exposures. The ETC also allows to calculate limiting magnitude or an exposure time for a given magnitude.

The sources subpackage contains a script to generate object catalogs with random x and y positions for stars and galaxies. The magnitudes of stars and galaxies are drawn from input distributions that are based on observations. As the number of stars depends on the galactic latitude, the script allows the user to use three different (30, 60, 90 degrees) angles when generating the magnitude distribution for stars (see the example plot below).

Another way of creating a source catalog is to use generateGalaxies script in the simulator subpackage. This script depends on IRAF and uses the makeart IRAF package. There are many more options in this script, which basically just calls the makeart's gallist and starlist. For options, see IRAF's documentation.

The simulator subpackage contains scripts to generate simulated VIS images. Two different methods of generating mock images is provided. One which takes observed images (say from HST) as an input and another in which analytical profiles are used for galaxies. The former code is custom made while the latter relies hevily on IRAF's artdata package and mkobjects task.

The VIS reference simulator is the custom made with real observed galaxies as an input. The IRAF based simulator can be used, for example, to train algorithms to derive elliptiticy of an object. For more detailed documentation, please see:

The reduction subpackage contains a simple script to reduce VIS data.

The analysis subpackage contains classes and scripts related to data analysis. A simple source finder and shape measuring classes are provided together with a wrapper to analyse reduced VIS data.

The data subfolder contains the supporting data, such as cosmic ray distributions, cosmetics maps, flat fielding files, PSFs, and an example configuration file.

The fitting subpackage contains a simple script that can be used to fit trap species so that the Charge Transfer Inefficiency (CTI) trails forming behind charge injection lines agree with measured data.

The fortran folder contains a CDM03 CTI model Fortran code. For speed the CDM03 model has been written in Fortran because it contains several nested loops. One can use f2py to compile the code to a format that can be imported directly to Python.

A few postage stamps showing observed galaxies have been placed to the objects directory. These FITS files can be used for, e.g., testing the shape measurement code.

The support subpackage contains some support classes and methods related to generating log files and read in data.

The reference image simulator has been tested against photometric accuracy (without aperture correction). A simulated image was generated with the reference simulator (using three times over sampled PSF) after which the different quadrants were combiled to form a single CCD. These data were then reduced using the reduction script provided in the package. Finally, sources were identified from the reduced data and photometry performed using SExtractor, after which the extracted magnitudes were compared against the input catalog.

The following figures show that the photometric accuracy with realistic noise and the end-of-life radiation damage is about 0.08 mag. Please note, however, that the derived magnitudes are based on a single 565 second exposure, and therefore the scatter is large in case of fainter objects.

The support subpackage contains classes and scripts that support other codes.