Skip to main content

Interface to data produced by the Simulating eXtreme Spacetimes collaboration

Project description

Test Status Documentation Status PyPI Version Conda Version MIT License Binder

Simulating eXtreme Spacetimes python package

The sxs python package provides a high-level interface for using data produced by the SXS collaboration. In particular, the function sxs.load can automatically find, download, and load data, returning objects that provide common interfaces to the various types of data, without forcing the user to worry about details like data formats or where to find the data. It can also automatically select the newest or highest-resolution dataset for a given simulation, or return a range of versions or resolutions. Currently, the high-level objects encapsulate

  • Catalog — a listing of all data produced by the SXS collaboration
  • Metadata — data describing the simulation parameters
  • Horizons — time-series data describing the apparent horizons
  • Waveforms — time-series data describing the extrapolated gravitational-wave modes

Installation

Because this package is pure python code, installation is very simple. In particular, with a reasonably modern installation, you can just run a command like

conda install -c conda-forge sxs

or

python -m pip install sxs

Here, conda requires the conda installation of python, which is the most recommended approach for scientific python; the second command assumes that you have an appropriate python environment set up in some other way. Either of these commands will download and install the sxs package and its most vital requirements.

If you want to install all the goodies that enable things like jupyter notebooks with plots and interactive tables, you could run

conda install -c conda-forge sxs-ecosystem

or

python -m pip install sxs[ecosystem]

You will probably also want to set some sensible defaults to automatically download and cache data:

python -c "import sxs; sxs.write_config(download=True, cache=True)"

This will write a configuration file in the directory returned by sxs.sxs_directory("config"), and downloaded data will be cached in the directory returned by sxs.sxs_directory("cache"). See that function's documentation for details.

Usage

An extensive demonstration of this package's capabilities is available here, in the form of interactive jupyter notebooks that are actually running this code and some pre-downloaded data. The following is just a very brief overview of the sxs package's main components.

There are four important objects to understand in this package:

import sxs

catalog = sxs.load("catalog")
metadata = sxs.load("SXS:BBH:0123/Lev/metadata.json")
horizons = sxs.load("SXS:BBH:0123/Lev/Horizons.h5")
waveform = sxs.load("SXS:BBH:0123/Lev/rhOverM", extrapolation_order=2)

The catalog object contains information about every simulation in the catalog, including all available data files, and information about how to get them. You probably don't need to actually know about details like where to get the data, but catalog can help you find the simulations you care about. Most importantly, catalog.simulations is a dict object, where the keys are names of simulations (like "SXS:BBH:0123") and the values are the same types as the metadata object, which contains metadata about that simulation — things like mass ratio, spins, etc. This metadata reflects the actual output of the simulations, which leads to some inconsistencies in their formats. A more consistent interface (though it is biased toward returning NaNs where a human might glean more information) is provided by catalog.table, which returns a pandas DataFrame with specific data types for each column.

The actual data itself is primarily contained in the next two objects. The horizons object has three attributes — horizons.A, horizons.B, and horizons.C — typically representing the original two horizons of the black-hole binary and the common horizon that forms at merger. In matter simulations, one or more of these may be None. Otherwise, each of these three is a HorizonQuantities object, containing several timeseries relating to mass, spin, and position.

Finally, the waveform encapsulates the modes of the waveform and the corresponding time information, along with relevant metadata like data type, spin weight, etc., and useful features like numpy-array-style slicing.

Project details


Release history Release notifications | RSS feed

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

sxs-2020.9.20.tar.gz (147.4 kB view hashes)

Uploaded Source

Built Distribution

sxs-2020.9.20-py3-none-any.whl (164.6 kB view hashes)

Uploaded Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page