Note

Before reading further, if you do not wish to install PandExo, there is an online version of the code here at STScI’s ExoCTK.

Pre-installation Data Download

PandExo requires downloading reference data for 1) JWST instrument info, 2) JWST PSFs, 3) stellar SEDs, and 4) normalization bandpasses. It also requires setting up the pandeia_refdata, PSF_DIR, and PYSYN_CDBS environment variables.

JWST Reference Data

JWST reference data has been updated to 2026.7.

Warning

Reference data must match the software version. For example, Pandeia reference data 2026.7 should match the Pandeia software release 2026.7, and that should match PandExo 2026.7.

The 2026.7 JWST reference data is available here:

More information is available in the Pandeia Engine installation guide.

After downloading the reference data, add the environment variables to your shell startup file so they persist across terminal sessions. For Bash, use ~/.bashrc or ~/.bash_profile; for Zsh, use ~/.zshrc. Add these lines to the appropriate file, replacing the example paths:

export pandeia_refdata=/path/to/pandeia_data-2026.7-jwst
export PSF_DIR=/path/to/pandeia_psfs-2026.7-jwst

Source the startup file to make the changes available in your current terminal, for example:

source ~/.bashrc

Use the path to the file you edited, such as source ~/.bash_profile or source ~/.zshrc. Opening a new terminal will also load the saved variables.

Stellar SEDs

PandExo uses synphot/stsynphot for stellar spectra and PHOENIX model interpolation. The PHOENIX reference atlas can be downloaded here.

Once untarred, the files will produce a directory tree of grp/redcat/trds. The pandeia.engine uses the contents of the trds directory.

Environment variable: $PYSYN_CDBS must point to the trds directory (NOT grp)

Add this environment variable to the same shell startup file and source that file again:

export PYSYN_CDBS=/path/to/grp/redcat/trds

Normalization Files

PandExo also needs the STScI/CDBS-style throughput files used for J/H/K normalization bandpasses.

Download the file here

Once untarred, this archive also produces a grp/redcat/trds directory with comp and mtab subdirectories. Merge their contents into the existing comp and mtab directories under $PYSYN_CDBS. Do not create a second grp/redcat/trds tree inside $PYSYN_CDBS.

For example, if NORMALIZATION_TRDS is the trds directory extracted from the normalization archive:

NORMALIZATION_TRDS=/path/to/extracted/grp/redcat/trds
rsync -a "$NORMALIZATION_TRDS/comp/" "$PYSYN_CDBS/comp/"
rsync -a "$NORMALIZATION_TRDS/mtab/" "$PYSYN_CDBS/mtab/"

The $PYSYN_CDBS directory should now contain comp, grid, and mtab (among other synphot directories). Verify that layout and the files PandExo uses for J/H/K normalization and PHOENIX spectra:

ls "$PYSYN_CDBS"
ls "$PYSYN_CDBS/comp/nonhst/bessell_j_003_syn.fits"
ls "$PYSYN_CDBS/comp/nonhst/bessell_h_004_syn.fits"
ls "$PYSYN_CDBS/comp/nonhst/bessell_k_003_syn.fits"
ls "$PYSYN_CDBS/grid/phoenix/catalog.fits"

Verify the complete reference-data setup with:

python -c "import pandeia.engine; pandeia.engine.pandeia_version()"

If properly installed and configured, it should show the matching versions and stellar reference-data directory, like this:

Pandeia Engine version:  2026.7
Pandeia RefData version: 2026.7
Pandeia PSFs version:    2026.7
Synphot Data:            /path/to/grp/redcat/trds

Fortney+ 2010 Planet Grid (Optional)

A user may want to install a grid of atmospheric models to simulate planet atmospheres. Some of the example notebooks use an atmospheric model grid. This grid can be obtained from:

The required fortney_models.db database is included in the ExoCTK Fortney data archive. Download and extract that archive, or install ExoCTK and run exoctk.utils.download_exoctk_data('fortney').

After downloading the Fortney files, add this optional environment variable to the same shell startup file and source that file again:

export FORTGRID_DIR=/path/to/exoctk_data/fortney/fortney_models.db

Installation with Pip or Git

Install PandExo in a dedicated virtual environment so its dependencies do not affect other projects. For example, with Python’s built-in venv:

python -m venv pandexo-env
source pandexo-env/bin/activate

On Windows, use the conda-based instructions in Windows Installation Guide for PandExo.

Install with pip:

python -m pip install pandexo.engine

Alternatively, clone PandExo from GitHub. The repository includes notebooks for getting started:

git clone https://github.com/natashabatalha/PandExo.git
cd PandExo
python -m pip install -e .

Final Test for Success

For a normal pip installation, run this installed-package smoke test. It loads PandExo’s bundled NIRSpec configuration and does not download data, start the web application, or run a simulation:

python -c "import pandexo.engine.justdoit as jdi; mode = jdi.load_mode_dict('NIRSpec Prism'); assert mode['configuration']['instrument']['disperser'] == 'prism'; print('PandExo configuration smoke test passed.')"

If you cloned the PandExo source repository and want to run its regression test, install the test dependency and run the test from the repository root:

python -m pip install -e ".[test]"
python -m pytest tests/test_import.py -q

The full tests/test_run.py suite runs Pandeia calculations and requires the reference data configured above. Run it only when validating a source checkout with a complete local data installation.

The Importance of Upgrading PandExo

It is crucial that your version of PandExo is up to date. There were many critical updates in the reference files after launch and as a result of commissioning work. Updating PandExo requires three crucial steps:

1) Verify pandexo.engine is Current

python -m pip install --upgrade pandexo.engine

2) Verify pandeia.engine version compatible

Currently PandExo requires pandeia.engine==2026.7.

python -m pip install pandeia.engine==2026.7

3) Grab pandeia.engine data 2026.7

The 2026.7 JWST reference data is available here: