Gravitational Lens Modeling of HST Observed Doubly Imaged Quasars

This repository stores the modeling and analysis notebooks for the paper, TDCOSMO XXIX: Uniform Lens Modeling of Eight Doubly Imaged Quasars (link TBD). This repository is organized as follows:

1. run_files: Contains the .sh slurm submission scripts used to run the Jupyter notebooks. The script run_joint.sh runs the three band joint fit via joint_model.ipynb.

   run_photometry.sh and multi_band_photometry.ipynb handle photometric calculations and posterior sampling.

   For cosmological analysis, run_cosmo.sh submits a job to run cosmology.ipynb, which calculates the Fermat potential and time delay differences with the posteriors generated from the modeling routine.

   run_conjugate.sh and run_conj_cosmo.sh submit jobs for the conjugate point analysis.

   continue_mcmc.ipynb allows one to continue the MCMC fitting using the final walker positions as the new starting point. run_mcmc.sh submits a job to continue the MCMC, with the parameter n_additional_steps controlling the number of additional steps passed into the Jupyter notebook.

   The jobs make use of papermill to execute the Jupyter notebooks.

2. yaml_files: Contains the input papermill .yaml files utilized in the joint modeling and cosmology analysis.

3. jupyter_notebooks. This directory stores all modeling notebooks.

To access the data products of Brady et al. 2026 (link TBD), please visit the corresponding Zenodo repository (link TBD). Upon downloading and opening the zipped directory, move brady_et_al_2026 into this directory, and move each individual folder out of brady_et_al_2026 and into this directory. The data products include:

1. joint_modeling: Stores the outputs for the FULL IMAGE JOINT MODELING, including the fitting notebook, photometry notebook, and cosmology notebook.
2. conjugate_point: This directory stores the outputs for the CONJUGATE POINT MODELING.
3. no_source: Stores the model results for full image modeling without the inclusion of a source profile.
4. cutout_data: PSF outputs for all systems in the sample, alongside old image cutouts and old files that modeled the systems via single band fits.

Useful analysis notebooks can be found in analysis. These include:

1) Producing the science image cutouts from the drizzled HST data
2) Generating a PSF with STARRED
3) Making a PSF variance map with psfr
4) Making LaTeX tables from the data
5) Creating a posterior comparison plot for full image vs. conjugate point
6) Mass-Light Parameter plot by band
7) Astrometry comparison with Gaia DR2
8) Fermat Potential Precision w/ Lensed Arc Magnitude
9) Visualize MCMC chains for convergence

Please contact Ryan Brady [ryan.m.brady@stonybrook.edu] for any questions, comments, or suggestions!