Large Scale Data & Computing Resources
LIGO, Virgo, KAGRA and GEO data centers use tools from Open Science Grid (OSG) (OSG) to share their data and computing resources. See here for more information.
Data access
The Open Science Data Federation (OSDF) is a service provided by OSG that acts as a storage resource broker and optimizes data access through a worldwide network of data centers.
Open data are stored under the /gwdata
namespace. It can be accessed freely without any identification step.
Similarly, LVK members can access proprietary data under /igwn
namespace after having registered a SCiToken.
OSDF access point servers:
- Open data discovery:
datafind.gwosc.org
- Proprietary data:
datafind.igwn.org
Software
To access data on the OSG, you need specialized software. We will use:
- GWDataFind: this is a command-line executable and a python library to query for the location of Gravitational-Wave Frame (GWF) files containing data from the LVK detectors.
- Pelican command-line client: this is the official command-line interface for OSDF.
- requests-pelican: a python library to provide simple interface to Pelican.
requests-pelican can be installed with pip.
Examples
On the command-line
Data discovery consists in the identification of frame types and time segments.
This is done with the gw_data_find
command.
Available frame types among all detectors can be queried as:
gw_data_find -r datafind.gwosc.org --show-types
Find available segments for the L1_GWOSC_O3a_16KHZ_R1 frame type from the Livingston observatory (L):
gw_data_find -r datafind.gwosc.org -o L -t L1_GWOSC_O3a_16KHZ_R1 --show-times
The list of OSDF data file paths for the Handford observatory H1_LOSC_4_V1 frame type with times between GPS 1126259460 and 1126259464 is given by:
gw_data_find -r datafind.gwosc.org -o H -t H1_LOSC_4_V1 -s 1126259460 -e 1126259464 -u osdf
Once the wanted file OSDF path is known from data discovery, it can be downloaded directly on your computer using the pelican
command.
As an example, to download the data file from the H1 observatory
with frame type H1_LOSC_4_V1 at GPS start time 1126256640:
pelican object get osdf:///gwdata/O1/strain.4k/frame.v1/H1/1126170624/H-H1_LOSC_4_V1-1126256640-4096.gwf my_local_copy.gwf
In Python
Data can be queried and downloaded in a similar manner inside a python script:
import os
import requests_pelican as rp
from gwdatafind import find_urls
url = find_urls( "H", "H1_LOSC_4_V1", 1126259460, 1126259464, urltype='osdf', host="https://datafind.gwosc.org")
with open(os.path.basename(url[0]), mode="wb") as file:
file.write(rp.get(url[0]).content)
Computing Resources
Some researchers may want to make use of the OSG service OSPool to submit jobs to shared computing resources.
- Create an account at osgconnect.net
- Use the OSG Help Desk to find documentation, support, and office hours
- Read the IGWN Guide to HTCondor to learn how to submit jobs to the OSPool using HTCondor