GalSpec: Galaxy Spectrum Fitting

GalSpec is a Python package for fitting astronomical spectra, specifically designed for AGN and galaxy spectra with complex emission and absorption line features.

Key Features

• Astropy Integration: All models inherit from Astropy’s modeling framework, ensuring compatibility with standard astronomical tools

• Multiple Fitting Methods: Support for least-squares, MCMC (emcee), and nested sampling (dynesty; under development) algorithms

• Comprehensive Line Database: Built-in database of UV/optical emission lines for AGN and galaxy analysis

• Template Support: Iron (Fe II) templates and stellar population models included

• Flexible Model Composition: Easily combine continuum, line, and template components

• Professional Fitting: Parameter constraints, tying, and bounds for complex models

• Sample Fitting From a File: one TOML describing the data, the model and the sampler, then fit one source or a whole sample with python -m galspec.pipeline

Quick Start

Install GalSpec from source:

git clone https://github.com/yourusername/galspec.git
cd galspec
pip install -e .

Fit a simple emission line:

import numpy as np
from astropy.modeling.models import Linear1D
from galspec import Line_Gaussian
from astropy.modeling.fitting import LevMarLSQFitter

# Load your spectrum data
wave = np.linspace(4000, 7000, 1000)
flux = np.random.randn(1000) * 0.1  # Your flux data here

# Create a model: continuum + emission line
continuum = Linear1D(slope=0.0, intercept=1.0)
line = Line_Gaussian(amplitude=1.0, dv=0, sigma=200, wavec=6563, name='Halpha')
model = continuum + line

# Fit the model
fitter = LevMarLSQFitter()
fitted_model = fitter(model, wave, flux)

Contents

User Guide:

• Installation
  • Requirements
  • Installation Methods
  • Verification
  • Troubleshooting
  • Next Steps
• Quick Start Guide
  • Summary
  • Next Steps
  • Advanced Topics
• Priors
  • The four shapes
  • A centred prior, in detail
  • What the fitter does with them
  • The published recipe
  • Worked example: a prior centred on another parameter
  • Worked example: switching a component off
  • Saving priors with a fit
• JAX / GPU acceleration
  • What is available
  • API
  • Fitting a sample
  • Precision
  • Adding a new model
  • Benchmarks
  • Saving a fit, and what lives on the GPU
  • Caveats
  • A complete AGN fit: CPU and GPU, same answer
  • Prior questions: how many components, and how similar must two lines be?
• Fitting From a Configuration File
  • Starting from a template
  • Why a file
  • Requirements on the data
  • Preparing a file
  • The run section
  • The data section
  • The fit section
  • The output section
  • The model section
  • Output
  • Compiling once for a sample
  • Command line
  • The IRAS reproduction
  • A small example to copy from
  • Limitations

Examples:

• Example: fitting a strong AGN
  • The spectrum
  • 1. Read the spectrum
  • 2. Measure the narrow-line profile on [O III]
  • 3. Build the model, one component at a time
  • 4. The ties
  • 5. A prior for every parameter
  • 6. Run the sampler
  • 7. Present the results
  • 8. Save it, and read it back
  • 9. What the numbers say
  • Where to go next
• Example: measure the black hole and the host galaxy
  • The answer, first
  • 1. The spectrum
  • 2. The continuum windows
  • 3. The continuum model
  • 4. sigma_* and the 5100 A luminosity
  • 5. What the continuum posterior is, and is not
  • 6. The narrow-line template, measured on [S II]
  • 7. The H-alpha complex
  • 8. H-beta and [O III]
  • 9. The fluxes and the widths
  • 10. The BPT diagram
  • 11. The black hole and the host galaxy
  • 12. What the numbers say against the reference fit
  • 13. What to check before believing any of it
  • 14. Saving the fit, and where to go next

Indices and tables

• Index

• Module Index

• Search Page

Links

• GitHub Repository: yourusername/galspec

• Issue Tracker: yourusername/galspec#issues

• Documentation: https://yourusername.github.io/galspec/