MPDAF: MUSE Python Data Analysis Framework

• MUSE Framework with MPDAF is a specialized Python package that manages three-dimensional astronomical datacubes and metadata for rigorous data analysis.
• It integrates with NumPy, SciPy, Matplotlib, and Astropy to enable efficient numerical, statistical, and visualization operations essential for research.
• Its modular design supports advanced features such as emission line detection, subcube extraction, and automated spectral fitting to facilitate innovative astronomical studies.

The MUSE (Multi Unit Spectroscopic Explorer) Python Data Analysis Framework (MPDAF) is a domain-specific software package developed to facilitate the analysis and scientific exploitation of data produced by the MUSE integral-field spectrograph on the ESO VLT. MPDAF provides an integrated environment for manipulating, calibrating, and extracting information from MUSE data products, which are three-dimensional datacubes containing spatial and spectral dimensions.

1. Scope and Objectives

MPDAF addresses the complexities inherent to the MUSE instrument, whose datacubes encode $x$, $y$ (spatial) and $\lambda$ (spectral) information for each pixel, accompanied by extensive world coordinate system (WCS), variance, and bad pixel mask metadata. The framework is engineered to:

• Load, manipulate, and analyze MUSE data cubes, images, and spectra while rigorously maintaining associated variances, WCS metadata, and masking information.
• Facilitate astronomical workflows such as subcube extraction, aperture and region-based extraction, image generation by axis summation, automated source detection, and quantitative spectral analysis.
• Support advanced instrument calibration and reduction steps via targeted manipulation of raw or reduced pixel tables.
• Ensure all processing steps preserve the integrity and astrophysical context of the data, in compliance with professional standards for reproducible research.

2. Integration with the Python Scientific Stack

MPDAF is layered atop established Python scientific libraries:

Dependency	Role in MPDAF
Numpy	Core array storage (including masked arrays for variance/masking) and fast numerical computations
Scipy	Provides routines for interpolation, resampling, smoothing, and non-linear optimization in spectral fitting
Matplotlib	Publication-grade visualization of cubes, images, and fitted spectra
Astropy	High-level FITS I/O, WCS manipulation, unit handling, and sky coordinate propagation

This approach ensures MPDAF packages domain-specific MUSE logic while leveraging robust, well-maintained scientific infrastructure.

3. Architecture and Key Modules

MPDAF is organized into three principal modules, each targeting a critical stage of the MUSE data workflow:

Module Name	Core Classes or Tools	Functionality
mpdaf.obj	Cube, Image	Cube/image loading with WCS, extraction (subcube, aperture), masking, spectral analysis
mpdaf.sdetect	MUSELET	Emission line / object detection, source database management, FITS output packaging
mpdaf.drs	PixTable	Raw/reduced pixel table handling, filtering for calibration and diagnostic applications

mpdaf.obj

• Provides Cube and Image classes for high-level manipulation, always preserving essential WCS and mask/variance information.
• Subcube and aperture extraction with arbitrary center and dimension; narrow-band/white-light image creation via summation over arbitrary wavelength intervals.
• Aperture and region-based spectral extraction.
• In-built spectral model fitting (e.g., asymmetric Gaussian fitting of emission lines with extraction of peak, FWHM).

Example Code Snippet

from mpdaf.obj import Cube cube = Cube('DATACUBE-HDFS-v1.34.fits') center = (-60.56183, 338.2168) # RA/Dec in degrees subcube = cube.subcube(center, 10) # 10" square region sp = subcube.aperture(center, 1) # 1" radius spectrum sp.plot(title='integrated spectrum within 1" radius aperture') lfit, rfit = sp.gauss_asymfit(lmin=5180, lmax=5260, plot=True)

Spectral fitting yields quantitative measurements directly on extracted features (e.g., Lyα line).

mpdaf.sdetect

• Implements MUSELET (built atop SExtractor) for line emitter detection.
• Aggregates source data (sub-cubes, images, integrated and zoomed spectra, and line parameters) in a single FITS package for synthesis.
• FITS storage schema unifies subcubes, spectral extractions, and derived astrophysical properties for each detected source.

mpdaf.drs

• Provides access to foundational pixel tables where each row encodes a pixel’s location (CCD, slice, spatial, spectral).
• Facilitates filtering by spatial, spectral, or instrumental criteria to, e.g., diagnose slice-to-slice offsets, crucial for calibration prior to datacube synthesis.

4. Use Cases and Workflow Examples

MPDAF is tailored to both scientific investigation and instrument calibration/validation. Notable examples include:

• Extracting a subregion of the Hubble Deep Field South MUSE cube, producing a narrow-band image, masking regions, and calculating integrated Lyα flux and error as a function of aperture radius.
• Assembling a FITS “source package” containing all relevant data products for a detected Lyα emitter: subcube, full/zoomed spectra, emission line parameters (e.g., peak, FWHM), and multiple image modalities (white-light, continuum, narrow-band, and supporting HST images).
• Systematic investigation and error diagnosis of reduced pixel tables for instrument commissioning and calibration (pixel filtering, slice registration, etc.)

5. Community, Distribution, and Development

MPDAF was originally developed for the MUSE consortium and is now open-source, with the following support mechanisms:

• Source code and release management via a public git repository.
• Standard installation with pip.
• Comprehensive, example-driven online documentation targeting new and advanced users.
• Community engagement for bug reporting, discussion, and contributions.
• Ongoing, active development—including refinement and extension of key modules such as mpdaf.sdetect—maintains alignment with the evolving needs of MUSE users and the astronomical community.

6. Significance and Impact

MPDAF establishes a specialized, technically robust, and community-driven framework for MUSE data analysis. Unlike generic toolkits, its tightly integrated modules maintain astrophysical metadata (WCS, variance, bad pixel information) across all processing stages and support both routine and advanced analysis (from cube manipulation to emission line fitting and automated source detection). Its evolution from internal consortium tool to public resource illustrates its stability and scientific significance. MPDAF is now a cornerstone resource for astronomers pursuing research with MUSE and stands as a model for specialized, open-source data analysis frameworks in precision astronomy.