Node to Joy - Modeling the external convergence from photometric catalogs

This package contains functionality to

• postprocess the coarse convergence values of an existing simulation to introduce finer fluctuations at galaxy-galaxy lensing scales
• train a Bayesian graph neural network to infer convergence given photometric measurements of galaxies around a line of sight
• hierarchically infer the mean and standard deviation of convergence in the population

Installation

0. Virtual environments are strongly recommended, to prevent dependencies with conflicting versions. Create a conda virtual environment and activate it:

$conda create -n n2j python=3.8 -y
$conda activate n2j

1. Clone the repo and install.

$git clone https://github.com/jiwoncpark/node-to-joy.git
$cd node-to-joy
$pip install -e . -r requirements.txt

2. (Optional) To run the notebooks, add the Jupyter kernel.

$python -m ipykernel install --user --name n2j --display-name "Python (n2j)"