piXedfit

piXedfit provides a compehensive set of tools for analyzing spatially resolved spectral energy distributions (SEDs) of galaxies and dissecting the spatially resolved properties of the stellar populations and dust in the galaxies. First, it can produce a pixel-matched 3D data cube from an input of a set of mutliband imaging data alone or in combination with an integral field spectroscopy (IFS) data. When IFS data is provided, it can produce a 3D spectrophotometric data cube in which spectra and photometric SEDs are combined on pixel level. Second, it has a unique pixel binning feature that can optimize the S/N ratio of SEDs on spatially resolved scales while retaining the spatial and spectral variations of the SEDs by accounting the similarity of SED shape of pixels in the binning process. This can be expected to reduce biases introduced by the binning process that combines pixels regardless of the variations in their SED shapes. Finally, piXedfit also provides a stand-alone SED fitting capability. It has two options of fitting methods: MCMC and random dense sampling of parameter space (RDSPS). Most of the modules in piXedfit have implemented MPI for parallel computation. A detailed description of piXedfit is presented in Abdurro'uf et al. (2021). The documentation is given here. Some examples of practical usages and tutorials can be found at folder examples and recent analysis with JWST + HST images.

Features

piXedfit has six modules that work independently with each other. For instance, it is possible to use the SED fitting module for fitting either global (integrated) or spatially resolved SEDs of galaxies. Those modules include:

piXedfit_images: image processing.
piXedfit_spectrophotometric: spatial and spectral matching between multiband imaging data and IFS data.
piXedfit_bin: pixel binning to optimize S/N of SEDs on spatially resolved scales.
piXedfit_model: generating model SEDs.
piXedfit_fitting: SED fitting on spatially resolved scales or global (integrated) scales.
piXedfit_analysis: Analysis of SED fitting result, including visualization plots and retrieving best-fitting parameters.

Installation

Please see installation for dependencies and requirements. To install piXedfit, please follow the instruction below.

cd to a desired installation directory, clone piXedfit, and install.

cd <install_dir>
git clone https://github.com/aabdurrouf/piXedfit.git
cd piXedfit
python -m pip install .

Set an environmental variable called PIXEDFIT_HOME that point to piXedfit parent directory.
```
export PIXEDFIT_HOME="$PWD"
```
The issue with the above command is that we need to do it every time we open a new terminal. Alternatively, we can add this environmental variable to .bashrc (add export PIXEDFIT_HOME="path_to_piXedfit" to the last line in the .bashrc file).
```
vi ~/.bashrc
```
Add the above line and then do
```
source  ~/.bashrc
```
To add the environmental variable permanently, we can add the same line to ~/.bash_profile or ~/.profile and do source command.

The above instruction is for first installation, for upgrading piXedfit, please see installation.

Convolution Kernels

Because of the large sizes of the kernel files, we do not upload them to this repository. We put the kernel files on this google drive link. To be able to use image processing feature (piXedfit_images), you need to download the necessary kernel files and copy them to /data/kernels ($PIXEDFIT_HOME/data/kernels). List of the kernel files needed for image processing would depend on the imaging data that will be analyzed.

Citation

If you use this code for your research, please reference Abdurro'uf et al. (2021):

@ARTICLE{2021ApJS..254...15A,
       author = {{Abdurro'uf} and {Lin}, Yen-Ting and {Wu}, Po-Feng and {Akiyama}, Masayuki},
        title = "{Introducing piXedfit: A Spectral Energy Distribution Fitting Code Designed for Resolved Sources}",
      journal = {\apjs},
     keywords = {Astronomical methods, Bayesian statistics, Galaxy evolution, Posterior distribution, 1043, 1900, 594, 1926, Astrophysics - Astrophysics of Galaxies},
         year = 2021,
        month = may,
       volume = {254},
       number = {1},
          eid = {15},
        pages = {15},
          doi = {10.3847/1538-4365/abebe2},
archivePrefix = {arXiv},
       eprint = {2101.09717},
 primaryClass = {astro-ph.GA},
       adsurl = {https://ui.adsabs.harvard.edu/abs/2021ApJS..254...15A},
      adsnote = {Provided by the SAO/NASA Astrophysics Data System}
}

If you use the pixel binning module (piXedfit_bin), please also reference Abdurro'uf & Akiyama (2017).

Reference

A list of some projects piXedfit is benefitted from:

Astropy
Photutils
Aniano et al. (2011) who provides convolution kernels
FSPS and Python-FSPS stellar population synthesis model
emcee package for the Affine Invariant Markov Chain Monte Carlo (MCMC) Ensemble sampler

aabdurrouf / piXedfit

readme