MEGS: Morphological Evaluation of Galactic Structure

This project, developed by Ufuk Çakır, introduces the code generate the dataset and evaluate galaxy morphology using Principal Component Analysis. The dataset offers detailed 2D maps and 3D cubes of 11 960 galaxies, capturing essential attributes: stellar age, metallicity, and mass.

Interdisciplinary Center for Scientific Computing (IWR), Heidelberg University, 09/2023

Table of Contents

1. Download Dataset
2. Installation
3. Configuration
4. Generation
5. Data Structure
6. Loading Data
7. PCA Benchmark
8. Interactive UMAP

Download Dataset

The dataset can be downloaded on Zenodo.

Installation

The code loads galaxies from the IllustrisTNG suite. For that, the respective python package should be installed:

$ cd ~
$ git clone https://github.com/illustristng/illustris_python.git
$ pip install illustris_python/

Check the Starting Guide on the TNG webpage for more information.

For installation of the code in a seperate virtual environment (either conda or venv), run
source setup.sh

Configuration

The config.json file contains all the settings nedded to run the data generation. All the configuration should be made there. The required fields are:

• simulation: Specifies the simulation from which the data will be generated. Currently, only IllustrisTNG is supported.
• particle_types: Defines the particle type(s) for which the images will be calculated.
• galaxy_parameters: Lists additional parameters to be saved for each galaxy.
• img_res: Sets the image resolution in each dimension.
• path: Indicates the output path for the created HDF5 file.
• halo_ids: Specifies the halo IDs for selection. If set to none, the program will automatically select galaxies.
• dim: Determines the dimension of the images, either 2D, 3D, or both.
• log_M_min: Defines the lower mass cut in $\log_{10}(M_\odot/h)$.
• log_M_max: Defines the upper mass cut in $\log_{10}(M_\odot/h)$.
• fields: Specifies the fields for which images will be calculated. For each field, the attributes mass_weighted and normed determine whether to calculate a mass-weighted image and whether or not to normalize it.
• GalaxyArgs: Contains arguments specified for loading galaxies, as defined in the Galaxy Class.

Generation

To generate the dataset run source generate_data.sh

This will select galaxies using the select_galaxies function and save the halo ids in a numpy array. Finaly the code runs the generate.py code to generate the dataset from the selected galaxies.

Data Structure

The data will be stored in a HDF5 File in the following way:

Loading Data

You can use the Gammaclass defined in load.py:

>>> from megs.data import Gamma
>>> path = "GAMMA.hdf5"

data = Gamma(path)

To get specific data from the Attributes group you can simply call

>>> data = Gamma("GAMMA.hdf5")
>>> data.get_attribute("mass")  # Get the mass of all galaxies in the dataset
>>> data.get_attribute("mass", 10)  # Get the mass of the 10th galaxy in the dataset

To get the images you can use:

>>> image = data.get_image("stars", "Masses", 10) # Get the stars masses image of the 10th galaxy in the dataset
>>> all_images = data.get_image("stars", "Masses") # Get all stars masses images in the dataset

PCA Benchmark

>>> from megs.data import Gamma
>>> path = "GAMMA.hdf5"
>>> data = Gamma(path)
>>> from megs.model import mPCA
>>> model = mPCA(data, dim=2) # Initialize PCA model for two dimensional data
Creating datamatrix with the following fields:
 ===============================================
 Particle type:  stars
 Fields:  ['GFM_Metallicity', 'GFM_StellarFormationTime', 'Masses']
 Dimension:  dim2
 Default arguments are used for the fields that are not specified in the norm_function_kwargs
 ===============================================
 Created datamatrix with shape:  (11727, 12288) 
>>> model.fit(n_components = 60, show_results = True)

Interactive UMAP

We calculate UMAP on the lower dimensional PCA scores to visualize the eigen space. An interactive version of this plot can be accessed through an online Dashboard.