FrankenBlast-Host 🚀

Code for rapid transient host association, photometry, and stellar population modeling.

If you use FrankenBlast, please cite:

• Nugent et al. 2026 (FrankenBlast) ADS link
• Jones et al. 2024 (BLAST web application)
  ADS link

The exact conda environment is specified in requirements.txt.

FrankenBlast requires Python 3.10.15.

1. Clone the repository:

   git clone https://github.com/anugent96/frankenblast-host.git
   cd frankenblast-host

2. Create the environment:

   conda create --name frankenblast python=3.10
   conda activate frankenblast
   pip install -r requirements.txt

3. Add custom paths from settings.sh to your shell configuration file (.zshrc, .bashrc, or .bash_profile).

Make sure the following packages are installed via pip or conda:

numpy, scipy, astro-prospector, python-fsps (and FSPS), sedpy, astro-prost, astroquery, 
corner, dustmaps, extinction, jupyter, pandas, sbi, seaborn, tensorflow, matplotlib

The SBI++ trained models for host SED fitting are too large for GitHub.
Please download them from:

Zenodo link

Files to download:

• sbipp_phot.zip
• sbi_training_sets.zip

Place these in the ./data/ directory and unzip.

• FrankenBlast Tutorial.ipynb — step-by-step guide to using FrankenBlast.
• SBI++ host SED fitting with LSST data — tutorial coming soon, although all trained models are currently available.

We are currently experiencing issues with the astro_prost PanSTARRS host association.

1. Edit astro_prost
   Comment out lines 2829–2830 in helpers.py:

   # galaxies["redshift_mean"] = np.nan
   # galaxies["redshift_std"] = np.nan

   ⚠️ This will return "0" as the redshift of the host.

2. Disable PanSTARRS
   In get_host.py (line 278), remove "panstarrs" from the catalogs dictionary.

Questions or comments?
Reach out to Anya Nugent → anya.nugent[at]cfa.harvard.edu