The machine learning pipeline CADET (CAvity DEtection Tool) finds and size-estimates arbitrary surface brightness depressions (X-ray cavities) on noisy Chandra images of galaxies. The pipeline is a self-standing Python script and inputs either raw Chandra images in units of counts (numbers of captured photons) or normalized background-subtracted and/or exposure-corrected images. CADET saves corresponding pixel-wise as well as decomposed cavity predictions in FITS format and also preserves the WCS coordinates; it also outputs a PNG file showing decomposed predictions for individual scales.

Idefix solves non-relativistic HD and MHD equations on various grid geometries. Based on a Godunov finite-volume method, this astrophysical flows code includes a wide choice of solvers and several modules, including constrained transport, orbital advection, and non-ideal MHD, to address complex astrophysical and fluid dynamics applications. Written in C++, Idefix relies on the Kokkos meta-programming library to guarantee performance portability on a wide variety of architectures.

CONCEPT (COsmological N-body CodE in PyThon) simulates cosmological structure formation. It can simulate matter particles evolving under self-gravity in an expanding background. The code offers multiple gravitational solvers and has adaptive time integration built in. In addition to particles, CONCEPT also evolves fluids at various levels of non-linearity, providing the means for the inclusion of more exotic species such as massive neutrinos, as well as for simulations consistent with general relativistic perturbation theory. Various non-standard species, such as decaying cold dark matter, are fully supported. CONCEPT includes a sophisticated initial condition generator and can output snapshots, power spectra, bispectra ,and several kinds of renders.

COLT (Cosmic Lyman-alpha Transfer) is a Monte Carlo radiative transfer (MCRT) solver for post-processing hydrodynamical simulations on arbitrary grids. These include a plane parallel slabs, spherical geometry, 3D Cartesian grids, adaptive resolution octrees, unstructured Voronoi tessellations, and secondary outputs. COLT also includes several visualization and analysis tools that exploit the underlying ray-tracing algorithms or otherwise benefit from an efficient hybrid MPI + OpenMP parallelization strategy within a flexible C++ framework.

lasso_spectra fits Lasso regression models to data, specifically galaxy spectra. It contains two classes for performing the actual model fitting. GeneralizedLasso is a tensorflow implementation of Lasso regression, which includes the ability to use link functions. SKLasso is a wrapper around the scikit-learn Lasso implementation intended to give the same syntax as GeneralizedLasso. It is much faster and more reliable, but does not support generalized linear models.

CosmoGraphNet infers cosmological parameters or the galaxy power spectrum. It creates a graph from a galaxy catalog with information the 3D position and intrinsic galactic properties. A Graph Neural Network is then applied to predict the cosmological parameters or the galaxy power spectrum.

ECLIPSE (Efficient Cmb poLarization and Intensity Power Spectra Estimator) implements an optimized version of the Quadratic Maximum Likelihood (QML) method for the estimation of the power spectra of the Cosmic Microwave Background (CMB) from masked skies. Written in Fortran, ECLIPSE can be used in a personal computer but also benefits from the capabilities of a supercomputer to tackle large scale problems; it is designed to run parallel on many MPI tasks. ECLIPSE analyzes masked CMB maps in which the signal can be affected by the beam and pixel window functions. The masks of intensity and polarization can be different and the noise can be isotropic or anisotropic. The program can estimate auto and cross-correlation power spectrum, that can be binned or unbinned.

Butterpy simulates star spot emergence, evolution, decay, and stellar rotational light curves. It tests the recovery of stellar rotation periods using different frequency analysis techniques. Butterpy can simulate light curves of stars with variable activity level, rotation period, spot lifetime, magnetic cycle duration and overlap, spot emergence latitudes, and latitudinal differential rotation shear.

Mixclask combines Cloudy (ascl:9910.001) and SKIRT (ascl:1109.003) to predict spectra and gas properties in astrophysical contexts, such as galaxies and HII regions. The main output is the mean intensity of a region filled with stars, gas and dust at different positions, assuming axial symmetry. The inputs for Mixclask are the stellar and ISM data for each region and an file for the positions (x,y,z) that will be output.

rfast ingests tables of opacities and generates synthetic spectra of worlds and retrieves real or simulated spectral observations. It can add noise, perform inverse modeling, and plot results. The tool can be applied to simulated and real observations spanning reflected-light, thermal emission, and transit transmission. Retrieval parameters can be toggled and parameters can be retrieved in log or linear space and adopt a Gaussian or flat prior.

Planetary Ephemeris Program (PEP) computes numerical ephemerides and simultaneously analyzes a heterogeneous collection of astrometric data. Written in Fortran, it is a general-purpose astrometric data-analysis program and models orbital motion in the solar system, determines orbital initial conditions and planetary masses, and has been used to, for example, measure general relativistic effects and test physics theories beyond the standard model. PEP also models pulsar motions and distant radio sources, and can solve for sky coordinates for radio sources, plasma densities, and the second harmonic of the Sun's gravitational field.

The Parthenon framework, derived from Athena++ (ascl:1912.005), handles massively-parallel, device-accelerated adaptive mesh refinement. It provides a device first/device resident approach, transparent packing of data across blocks (to reduce/hide kernel launch latency), and direct device-to-device communication via asynchronous, one-sided MPI communication to enable high performance. Parthenon uses an intermediate abstraction layer to hide complexity of device kernel launches, offers support for particles and abstract variable control via metadata tags, and has a flexible plug-in package system.

ALminer queries, analyzes, and visualizes the ALMA Science Archive. Users can programmatically query the archive for positions, target names, or other keywords in the archive metadata (such as proposal title, abstract, or scientific category). ALminer's plotting routines allow the query results to be visualized, and its analysis functions allow users to filter the results and check whether certain frequencies of interest are covered in the queried observations. The code also allows users to directly download ALMA data products in FITS format and/or the raw data that can be used for manual image processing. ALminer has been designed to make mining the ALMA archive as simple as possible, while being flexible to be customized according to the user's scientific interests. The code is released with a detailed tutorial Jupyter notebook, introducing ALminer's common functions as well as some of its more advanced options.

COpops computes semi-analytically the CO flux of a disc (given initial conditions and age) under the assumption of LTE and optically thick emission. It then runs disc population synthesis using observationally-informed initial conditions. CO fluxes is one of the most easily accessible observables for studying disc evolution; COpops is a faster alternative to running computationally-expensive thermochemical models for hundreds of discs and is accurate, recovering agreement within a factor of three.

RELAGN creates spectral models for the calculation of AGN SEDs, ranging from the Optical/UV (outer accretion disc) to the Hard X-ray (Innermost X-ray Corona). The code is available in two languages, Python and Fortran. The Fortran version is written to be used with the spectral fitting software XSPEC (ascl:9910.005), and is the preferred version for analyzing X-ray spectral data. The Python version provides more flexibility for modeling. Whereas the Fortran version produces only a spectrum, the Python implementation can extract the physical properties of the system (such as the physical mass accretion rate, disc size, and efficiency parameters) since these are all stored as attributes within the model. Both versions require a working installation of HEASOFT (ascl:1408.004).

apollinaire provides functions and a framework for helioseismic and asteroseismic instruments data managing and analysis, and includes all the tools necessary to analyze the acoustic oscillations of solar-like stars. The core of the package is the peakbagging library, which provides a full framework to extract oscillation modes parameters from solar and stellar power spectra.

pipes_vis is an interactive graphical user interface for visualizing SPS spectra. Powered by Bagpipes (ascl:2104.017), it provides real-time manipulation of a model galaxy's star formation history, dust, and other relevant properties through sliders and text boxes.

mockFRBhosts estimates the fraction of FRB hosts that can be cataloged with redshifts by existing and future optical surveys. The package uses frbpoppy (ascl:1911.009) to generate a population of FRBs for a given radio telescope. For each FRB, a host galaxy is drawn from a data base generated by GALFORM (ascl:1510.005). The galaxies' magnitudes in different photometric surveys are calculated as are the number of bands in which they are detected. mockFRBhosts also calculates the follow-up time in a 10-m optical telescope required to do photometry or spectroscopy and provides a simple interface to Bayesian inference methods via MCMC simulations provided in the FRB package (ascl:2306.018).

The transient search pipeline realfast integrates with the real-time environment at the Very Large Array (VLA) to look for fast radio bursts, pulsars, and other rare astrophysical transients. The software monitors multicast messages, catches visibility data, and defines a fast transient search pipeline with rfpipe (ascl:1710.002). It indexes candidate transients and other metadata for the search interface, and writes and archives new visibility files for candidate transients. realfast provides support for GPU algorithms, manages distributed futures, and performs blind injection and management of mock transients, among other tasks, and rapidly distributes data products and transient alerts to the public.

FRB performs calculations, estimations, analysis, and Bayesian inferences for Fast Radio Bursts, including dispersion measure and emission measure calculations, derived properties and spectrums, and Galactic RM.

Zeus21 (Zippy Early-Universe Solver for 21-cm) captures the nonlocal and nonlinear physics of cosmic dawn to create an effective model for the 21-cm power spectrum and global signal. The code takes advantage of the approximate log-normality of the star-formation rate density (SFRD) during cosmic dawn to compute the 21-cm power spectrum analytically. It agrees with more expensive semi-numerical simulations to roughly 10% precision, but has comparably negligible computational cost (~ s) and memory requirements. Zeus21 pairs well with data from HERA, but can be used for any 21-cm inference or prediction. Its capabilities include finding the 21-cm power spectrum (at a broad range of k and z), the global signal, IGM temperatures (Tk, Ts, Tcolor), neutral fraction xHI, Lyman-alpha fluxes, and the evolution of the SFRD; all across cosmic dawn z=5-35. It can also predict UVLFs for HST and JWST. Zeus21 can use three different astrophysical models, one of which emulates 21cmFAST (ascl:1102.023), and can vary the cosmology through CLASS (ascl:1106.020).

SuperRad models ultralight boson clouds that arise through black hole superradiance. It uses numerical results in the relativistic regime combined with analytic estimates to describe the dynamics and gravitational wave signals of ultralight scalar or vector clouds. Written in Python, SuperRad includes a set of testing routines that check the internal consistency of the package; these tests mainly serve the purpose of ensuring functionality of the waveform model but can also be utilized to check that SuperRad works as intended.

Mangrove uses Graph Neural Networks to regress baryonic properties directly from full dark matter merger trees to infer galaxy properties. The package includes code for preprocessing the merger tree, and training the model can be done either as single experiments or as a sweep. Mangrove provides loss functions, learning rate schedulers, models, and a script for doing the training on a GPU.

AIOLOS solves differential equations for hydrodynamics, friction, (thermal) radiation transport and (photo)chemistry for simulating accretion onto, and hydrodynamic escape from, planetary atmospheres. The 1-D multispecies, multiphysics hydrodynamics code, written in C++, compiles in a flexible mode that runs problems with any number of input species, and can be sped up by setting the number of species at compile time, and allows the user to provide initial conditions or boundary conditions if desired. AIOLOS provides output and diagnostic files that give snapshots in time of the state of the simulation. Output files are specific to each species, and diagnostic files contain summary as well as detailed information for, for example, the radiation transport, opacities for all species, and optical cell depths per band, in addition to other information.

SCONCE-SCMS detects cosmic web structures, primarily cosmic filaments and the associated cosmic nodes, from a collection of discrete observations with the extended subspace constrained mean shift (SCMS) algorithms on the unit (hyper)sphere (in most cases, the 2D (RA,DEC) celestial sphere), and the directional-linear products space (most commonly, the 3D (RA,DEC,redshift) light cone).

The subspace constrained mean shift (SCMS) algorithm is a gradient ascent typed method dealing with the estimation of local principal curves, more widely known as density ridges. The one-dimensional density ridge traces over the curves where observational data are highly concentrated and thus serves as a natural model for cosmic filaments in our Universe. Modeling cosmic filaments as density ridges enables efficient estimation by the kernel density estimator (KDE) and the subsequent SCMS algorithm in a statistically consistent way. While the standard SCMS algorithm can identify the density ridges in any "flat" Euclidean space, it exhibits large bias in estimating the density ridges on the data space with a non-linear curvature. The extended SCMS algorithms used in SCONCE-SCMS are adaptive to the spherical and conic geometries and resolve the estimation bias of the standard SCMS algorithm on a 2D (RA,DEC) celestial sphere or 3D (RA,DEC,redshift) light cone.

ZodiPy simulates the zodiacal emission in intensity that an arbitrary solar system observer is predicted to see given an interplanetary dust model, either in the form of timestreams or full-sky HEALPix maps. Written in Python, the code makes zodiacal emission simulations more accessible by providing a simple interface to existing models.

Margarine computes marginal bayesian statistics given a set of samples from an MCMC or nested sampling run. Specifically, the code calculates marginal Kullback-Leibler divergences and Bayesian dimensionalities using Masked Autoregressive Flows and Kernel Density Estimators to learn and sample posterior distributions of signal subspaces in high dimensional data models, and determines the properties of cosmological subspaces, such as their log-probability densities and how well constrained they are, independent of nuisance parameters. Margarine thus allows for direct and specific comparison of the constraining ability of different experimental approaches, which can in turn lead to improvements in experimental design.

MOBSE investigates the demography of merging BHBs. A customized version of the binary stellar evolution code BSE (ascl:1303.014), MOBSE includes metallicity-dependent prescriptions for mass-loss of massive hot stars and upgrades for the evolution of single and binary massive stars.

Albatross analyzes Milky Way stellar streams. This Simulation-Based Inference (SBI) library is built on top of swyft (ascl:2302.016), which implements neural ratio estimation to efficiently access marginal posteriors for all parameters of interest. Using swyft for its internal Truncated Marginal Neural Ratio Estimation (TMNRE) algorithm and sstrax (ascl:2306.008) for fast simulation and modeling, Albatross provides a modular inference pipeline to support parameter inference on all relevant parts of stellar stream models.

sstrax provides fast simulations of Milky Way stellar stream formation. Using JAX (ascl:2111.002) acceleration to support code compilation, sstrax forward models all aspects of stream formation, including evolution in gravitational potentials, tidal disruption and observational models, in a fully modular way. Although sstrax is a standalone python package, it was also developed to integrate directly with the Albatross (ascl:2306.009) inference pipeline, which performs inference on all relevant aspects of the stream model.

PhotoParallax calculates photometric parallaxes for distant stars in the Gaia TGAS catalog without any use of physical stellar models or stellar density models of the Milky Way. It uses the geometric parallaxes to calibrate a photometric model that is purely statistical, which is a model of the data rather than a model of stars per se.

β-SGP deconvolves an astronomical image with a known Point Spread Function, providing a means for restoration of telescopic images due to issues ranging from atmospheric turbulence to instrumental aberrations. The code supports improved astrometry, deblending of overlapping sources, faint source detection, and identification of point sources near bright extended objects, and other tasks. β-SGP generalizes the Scaled Gradient Projection (SGP) image deconvolution algorithm using β-divergence as a loss function to restore distorted stellar shapes.

Delight infers photometric redshifts in deep galaxy and quasar surveys. It uses a data-driven model of latent spectral energy distributions (SEDs) and a physical model of photometric fluxes as a function of redshift, thus leveraging the advantages of both machine- learning and template-fitting methods by building template SEDs directly from the training data. Delight obtains accurate redshift point estimates and probability distributions and can also be used to predict missing photometric fluxes or to simulate populations of galaxies with realistic fluxes and redshifts.

Latest version of TS (Turbospectrum), with NLTE capabilities.
Computation of stellar spectra (flux and intensities) in 1D or average stellar atmosphere models.
In order to compute NLTE stellar spectra, additional data is needed, downloadable outside GitHub.
See documentation in DOC folder

Python wrappers are available at https://github.com/EkaterinaSe/TurboSpectrum-Wrapper/ and https://github.com/JGerbs13/TSFitPy
They allow interpolation between models and fitting of spectra to derive stellar parameters.

The N-body code TIDYMESS (TIdal DYnamics of Multi-body ExtraSolar Systems) can describe the mass distribution of each body its inertia tensor and directly and self-consistently integrates orbit, spin, and inertia tensors. It manages the deformation of a body follows the tidal Creep model and includes the centrifugal force and tidal force. Written in C++, TIDYMESS is available as a standalone package and also through the AMUSE framework (ascl:1107.007).

SAVED21cm extracts the 21cm signal from the simulated mock observation for the Radio Experiment for the Analysis of Cosmic Hydrogen (REACH). Though built for the REACH experiment, this 21cm signal extraction pipeline can in principle can be utilized for any global 21cm experiment. The toolkit is based on a pattern recognition framework using the Singular Value Decomposition (SVD) of the 21cm and foreground training set. SAVED21cm finds the patterns in the training sets and properly models the chromatic distortions with a better basis than the polynomials.

Simulation-based inference is the process of finding parameters of a simulator from observations. The PyTorch package sbi performs simulation-based inference by taking a Bayesian approach to return a full posterior distribution over the parameters, conditional on the observations. This posterior can be amortized (i.e. useful for any observation) or focused (i.e.tailored to a particular observation), with different computational trade-offs. The code offers a simple interface for one-line posterior inference.

HAFFET (Hybrid Analytic Flux FittEr for Transients) analyzes supernovae photometric and spectroscopic data. It handles observational data for a set of targets, estimates their physical parameters, and visualizes the population of inferred parameters. HAFFET defines two classes, snobject for data and fittings for one specific object, and snelist to organize the overall running for a list of objects. The HAFFET package includes utilities for downloading SN data from online sources, intepolating multi band lightcurves, characterizing the first light and rising of SNe with power law fits, and matching epochs of different bands. It can also calculate colors, and/or construct the spectral energy distribution (SED), estimate bolometric LCs and host galaxy extinction, fit the constructed bolometric lightcurves to different models, and identify and fit the absorption minima of spectral lines, in addition to performing other tasks. In addition to utilizing the built-in models, users can add their own models or import models from other python packages.

The Australian Square Kilometre Array Pathfinder (ASKAP) has been enabled by the Commensal Real-time ASKAP Fast Transients Collaboration (CRAFT) to detect Fast Radio Bursts (FRBs) in real-time and save raw antenna voltages containing FRB detections. CELEBI, the CRAFT Effortless Localization and Enhanced Burst Inspection pipeline, extends CRAFT’s existing software to process ASKAP voltages to produce sub-arcsecond precision localizations and polarimetric data at time resolutions as fine as 3 ns of FRB events. CELEBI uses Nextflow (ascl:2305.024) to link together Bash and Python code to perform software correlation, interferometric imaging, and beamforming, thereby making use of common astronomical software packages.

Nextflow enables scalable and reproducible scientific workflows using software containers. It allows the adaptation of pipelines written in the most common scripting languages. Its fluent DSL simplifies the implementation and the deployment of complex parallel and reactive workflows on clouds and clusters. Nextflow supports deploying workflows on a variety of execution platforms including local, HPC schedulers, AWS Batch, Google Cloud Life Sciences, and Kubernetes. Additionally, it provides support for workflow dependencies through built-in support for, for example, Conda, Spack, Docker, Podman, Singularity, and Modules.

GLASS (Generator for Large Scale Structure) produces cosmological simulations on the sphere. The full, three-dimensional past light cone of the observer is discretized into a sequence of nested shells, which are further discretized in the angular dimensions into maps of the sphere. GLASS was originally designed to simulate cosmic matter, weak gravitational lensing, and galaxy positions, but its flexible design and open architecture allows it to be used for a wide range of cosmological and astrophysical simulations on the sphere.

GrGadget merges the Particle-Mesh (PM) relativistic GEVOLUTION code (ascl:1608.014) with the TreePM GADGET-4 code (ascl:2204.014) to create a TreePM simulation code that represents metric perturbations at the scales where they are relevant while resolving non-linear structures. The better resolution of the highly non-linear regime improves the representation of the relativistic fields sampled on the mesh with respect to PM-only simulations.

COLIBRÌ (which roughly stands for “Cosmological Libraries”) computes cosmological quantities such as ages, distances, power spectra, and correlation functions. It supports Lambda-CDM cosmologies plus extensions including massive neutrinos, non-flat geometries, evolving dark energy (w0-wa) models, and numerical recipes for f(R) gravity. COLIBRÌ is built especially for large-scale structure purposes and can interact with the Boltzmann solvers CAMB (ascl:1102.026) and CLASS (ascl:1106.020).

JEDI searches for and characterizes coronal dimming in light curves produced from the Solar Dynamics Observatory (SDO) Extreme Ultraviolet (EUV) Variability Experiment (EVE). The suite has a wrapper script that calls other functions, which can also be run independently assuming needed inputs from prior functions are provided. JEDI's functions fit light curves and return the best fit, compute precision for iron light curves, and find the biggest dimming depth and its time in a given light curve. JEDI also includes functions for finding the duration of the dimming, minimum, maximum, and mean slope of dimming of a light curve, and for identifying the biggest peak in two light curves, time shifting them so the peaks are concurrent, scaling them so the peaks are the same magnitude, and then subtracting them, among other useful functions.

The sterile neutrino production code sterile-dm incorporates new elements to the calculations of the neutrino opacity at temperatures 10 MeV ≤ T ≤ 10 GeV and folds the asymmetry redistribution and opacity calculations into the sterile neutrino production computation, providing updated PSDs for the range of parameters relevant to the X-ray excess. The code requires several data files, which are included. With each run, sterile-dm creates a new output sub-directory that contains a parameter file listing the mass, mixing angle, initial lepton asymmetry and other information, a state file, which includes, among other states, the temperature and FRW coordinate time, and a set of snapshot files, one for each line in the state file.

GWSurrogate provides an easy to use interface to gravitational wave surrogate models. Surrogates provide a fast and accurate evaluation mechanism for gravitational waveforms which would otherwise be found through solving differential equations. These equations must be solved in the “building” phase, which was performed using other codes.

Made-to-measure (M2M) is a standard technique for modeling the dynamics of astrophysical systems in which the system is modeled with a set of N particles with weights that are slowly optimized to fit a set of constraints while integrating these particles forward in the gravitational potential. Simple-m2m extends this standard technique to allow parameters of the system other than the particle weights to be fit as well, including nuisance parameters that describe the observer's relation to the dynamical system (e.g., the inclination) or parameters describing an external potential.

The gw_pta_emulator reads in gravitational wave (GW) characteristic strain spectra from black-hole population simulations, re-bins for the user's observing baseline, and constructs new spectra. The user can train a Gaussian process to emulate the spectral behavior at all frequencies across the astrophysical parameter space of supermassive black-hole binary environments.

EIDOS models the primary beam of radio astronomy antennas. The code can be used to create MeerKAT L-band beams from both holographic (AH) observations and EM simulations within a maximum diameter of 10 degrees. The beam model is less accurate at higher frequencies, and performs much better below 1400 MHz. The diagonal terms of the model beam Jones matrix are much better known than the off-diagonal terms. The performance of EIDOS is dependent on the quality of the given AH and EM datasets; as more accurate AH models and EM simulations become available, this pipeline can be used to create more accurate sparse representation of primary beams using Zernike polynomials.

DP3 (the Default Preprocessing Pipeline) is the LOFAR data pipeline processing program and is the successor to DPPP (ascl:1804.003). It performs many kinds of operations on the data in a pipelined way so the data are read and written only once. DP3 preprocesses the data of a LOFAR observation by executing steps such as flagging or averaging. Such steps can be used for the raw data as well as the calibrated data by defining the data column to use. One or more of the following steps can be defined as a pipeline. DP3 has an implicit input and output step. It is also possible to have intermediate output steps. DP3 comes with predefined steps, but also allows the user to plug in arbitrary steps implemented in either C++ or Python.