Et tu, Brute?
brutus is a Pure Python package for "brute force" Bayesian inference to derive distances, reddenings, and stellar properties from photometry. The package is designed to be highly modular and user-friendly, with comprehensive support for modeling individual stars, star clusters, and 3-D dust mapping.
Please contact Josh Speagle ([email protected]) with any questions.
The most recent stable release can be installed via pip by running
pip install astro-brutusI am currently in the midst of a large-scale refactor and the main documentation and repository will be unstable. Please refer to the last stable release in the meantime and avoid installing from GitHub.
🌟 Individual Star Modeling: Fit distances, reddenings, and stellar properties for individual stars using Bayesian inference
🌟 Cluster Analysis: Model stellar clusters with consistent ages, metallicities, and distances
🌟 3D Dust Mapping: Integrate with 3D dust maps and model extinction along lines of sight
🌟 Modern Stellar Models: Built-in support for MIST isochrones and evolutionary tracks
🌟 Flexible & Fast: Optimized algorithms with numba acceleration and modular design
🌟 Publication Ready: Designed for ease of use in research workflows
- Python: 3.8 or higher
- Operating System: Linux, macOS, or Windows with WSL (see Windows note below)
For most users, install from PyPI:
pip install astro-brutusFor development or to get the latest features:
git clone https://github.com/joshspeagle/brutus.git
cd brutus
pip install -e ".[dev]"healpy dependency (required for dust mapping), brutus does not work reliably on native Windows. Windows users should install and run brutus in WSL (Windows Subsystem for Linux).
Alternative Windows installation options:
- WSL (Recommended): Install Ubuntu or another Linux distribution via WSL and use the standard installation
- Conda: Try
conda install -c conda-forge astro-brutuswhich may have pre-compiled Windows wheels - Docker: Use a Linux-based Docker container
If you use conda, you can install from conda-forge:
conda install -c conda-forge astro-brutusCore dependencies that will be automatically installed:
numpy(≥1.19) - Numerical computingscipy(≥1.6) - Scientific computingmatplotlib(≥3.3) - Plottingh5py(≥3.0) - HDF5 file supporthealpy(≥1.14) - HEALPix utilities for incorporating dust mapsnumba(≥0.53) - Just-in-time compilation for performancepooch(≥1.4) - Data downloading and management
import numpy as np
from brutus import BruteForce, StarGrid, load_models
# Load stellar models
models, labels = load_models('path/to/models.h5')
# Create a StarGrid
star_grid = StarGrid(models, labels)
# Set up the fitter
fitter = BruteForce(star_grid)
# Your photometry data (flux units)
photometry = np.array([1.2e-3, 0.8e-3, 0.6e-3]) # g, r, i bands
errors = np.array([1e-5, 1e-5, 1e-5])
# Fit the star
results = fitter.fit(photometry, errors, parallax=2.5, parallax_err=0.1)from brutus import Isochrone
# Create isochrone generator
iso = Isochrone()
# Generate stellar parameters for an isochrone
params = iso.get_predictions(
feh=0.0, # Solar metallicity [Fe/H]
afe=0.0, # Solar alpha enhancement [alpha/Fe]
loga=9.0 # 1 Gyr age (log10(age/yr))
)from brutus import fetch_grids, fetch_isos, fetch_dustmaps
# Download stellar evolution grids
fetch_grids()
# Download isochrone data
fetch_isos()
# Download 3D dust maps
fetch_dustmaps()📚 Full Documentation: https://github.com/joshspeagle/brutus (comprehensive docs coming soon)
🎓 Tutorials: The tutorials/ directory contains Jupyter notebooks demonstrating key workflows
Brutus requires stellar evolution models and other data files to function. These can be downloaded automatically using built-in utilities:
from brutus import fetch_grids, fetch_isos, fetch_dustmaps
# Download MIST stellar evolution grids
fetch_grids()
# Download MIST isochrones
fetch_isos()
# Download 3D dust maps
fetch_dustmaps()🔧 Major Refactoring: Complete reorganization for improved usability and maintainability
🐍 Modern Python: Updated to support Python 3.8+ (dropped Python 2.7 support)
📦 Modern Packaging: Migrated to pyproject.toml and modern build system
🧪 Testing Framework: Added comprehensive test suite with CI/CD
📁 Modular Structure: Reorganized into logical modules (core, analysis, dust, utils, data)
🔄 Backward Compatibility: Old import patterns continue to work during transition
See CHANGELOG.md for complete details.
We welcome contributions! Please see CONTRIBUTING.md for guidelines.
git clone https://github.com/joshspeagle/brutus.git
cd brutus
pip install -e ".[dev]"# Basic tests
pytest
# Include slow tests
RUN_SLOW_TESTS=1 pytest
# With coverage
pytest --cov=brutusThis project is licensed under the MIT License - see the LICENSE file for details.
- Author: Joshua S. Speagle
- Email: [email protected]
- GitHub: https://github.com/joshspeagle/brutus
- Issues: https://github.com/joshspeagle/brutus/issues
brutus builds on stellar evolution models from the MIST project and dust maps from the Bayestar project. We thank the developers of these projects for making their data publicly available.
