Installation

Installation#

Requirements#

Before installing pyRadtran, make sure you have:

  • Python 3.9 or higher

  • libRadtran installed on your system

  • Common scientific Python packages (numpy, pandas, xarray) — installed automatically with pyRadtran

Install from Source#

# Using SSH (if you have SSH keys set up)
git clone git@github.com:FranzFlink/pyRadtran.git
# or standard HTTPS
# git clone https://github.com/FranzFlink/pyRadtran.git

cd pyRadtran
pip install -e .

Installing libRadtran#

pyRadtran requires a working libRadtran installation. Follow these steps:

  1. Download libRadtran from http://www.libradtran.org

  2. Extract the archive and build:

tar -xzvf libradtran-x.yy.tar.gz
cd libradtran-x.yy
./configure
make
make check

  1. Note the installation path — you’ll need it to configure pyRadtran.

Setting Up Your Master Configuration#

pyRadtran uses a layered configuration system. The most important first step after installation is creating your personal master configuration file. This tells pyRadtran where to find libRadtran on your machine.

The quickest way is from Python:

import pyradtran

pyradtran.save_master_config(
    libradtran_bin="/opt/libradtran/bin/uvspec",
    libradtran_data="/opt/libradtran/share/libRadtran/data",
)
# Master config saved to: ~/.pyradtran/config.yaml

Or write ~/.pyradtran/config.yaml by hand, adjusting paths to match your libRadtran installation:

# ~/.pyradtran/config.yaml — User-specific configuration
# This file is loaded automatically and overrides built-in defaults.
# Adjust the paths below to match your local libRadtran installation.

paths:
  libradtran_bin: /opt/libradtran/bin/uvspec
  libradtran_data: /opt/libradtran/share/libRadtran/data
  atmosphere_profile: afglms        # short name or full path
  solar_spectrum: NewGuey2003       # short name or full path

atmosphere_profile and solar_spectrum accept catalogue short names (afglus, afglms, …, kurudz_1.0nm, NewGuey2003, …) that resolve inside libradtran_data — see pyradtran.list_atmosphere_profiles() and pyradtran.list_solar_spectra().

Tip

Once your master config is set, you typically don’t need to specify libRadtran paths in individual simulation configs — they inherit from your master config automatically.

Optional Dependencies#

Some pyRadtran features require additional packages:

Package

Purpose

Install

siphon

Radiosonde data retrieval from IGRA

pip install siphon

gcsfs

ERA5 data access from Google Cloud

pip install gcsfs

scikit-learn

Statistical metrics for validation notebooks

pip install scikit-learn

seaborn

Enhanced plotting in some notebooks

pip install seaborn

 
# Install all optional dependencies at once
pip install siphon gcsfs scikit-learn seaborn

Verifying Your Installation#

After installation, verify everything works:

import pyradtran
print(pyradtran.__version__)

Then try running the albedo quickstart notebook to confirm that pyRadtran can communicate with libRadtran.

Contents