When the James Webb Space Telescope captures light from distant galaxies, it records a treasure trove of emission lines—the spectral signatures of hydrogen, oxygen, and other elements forged in stellar furnaces. But extracting meaningful measurements from these faint, often undersampled spectral features requires sophisticated analysis that can handle the complexities of real astronomical data. This is where Unite steps in as your computational ally.
Built on JAX and NumPyro for blazing-fast performance, Unite specializes in exact pixel integration of line profiles using multiple mathematical models (Gaussian, Voigt, Gauss-Hermite) while simultaneously fitting spectra across different gratings with shared kinematic parameters. The engine handles the messy realities of astronomical spectroscopy—calibration uncertainties, flexible continuum modeling, and multi-instrument data fusion—all wrapped in a YAML-serializable configuration system that makes complex analyses reproducible and shareable.
Originally designed for JWST’s NIRSpec instrument but extensible to any spectrograph, Unite empowers researchers to extract precise redshifts, line fluxes, and kinematic properties from their spectral observations. Whether you’re studying galaxy evolution, stellar atmospheres, or hunting for the most distant quasars, this turbo-charged inference engine transforms spectral analysis from a computational bottleneck into a streamlined discovery pipeline.
⭐ Stars: 6
💻 Language: Python
🔗 Repository: TheSkyentist/unite