Data-discovery on the LIGO Data Grid¶

Data discovery¶

The LIGO Data Grid consists of a trio of large-scale high-throughput computing (HTC) facilities operated by the LIGO Laboratory, and serves as the primary host of the data recorded from the observatories.

The observatories record ~30 MBytes/second of ‘raw’ data that are continuously written to GWF-format files (see Gravitational-wave frames (GWF) for more details on the GWF format), as well as the processed strain data produced in real-time for low-latency gravitational-wave searches. A data discovery service named datafind is available allowing any authenticated user to query for the locations of GWF files by specifying the observatory, the frame type, and a time interval.

The TimeSeries.find method leverages that discovery service to enable users to automatically locate and read data from these files in a convenient manner. For example:

>>> data = TimeSeries.find('L1:ISI-GND_STS_ITMY_Z_DQ', 'Jan 1 2016', 'Jan 1 2016 01:00')

This method will search through all available data to find the correct files to read, so this may take a while. If you know the frametype - the tag associated with files containing your data - you can pass that via the frametype keyword argument to significantly speed up the search:

>>> data = TimeSeries.find('L1:ISI-GND_STS_ITMY_Z_DQ', 'Jan 1 2016', 'Jan 1 2016 01:00', frametype='L1_R')

Frametypes¶

All data recorded by LIGO are identified by a frametype tag, which identifies which data are contained in a given gwf file. The following table is an incomplete, but probably OK, reference to which frametype you want to use for auxiliary data access:

The above frametypes refer to the H1 (LIGO-Hanford) instrument, the same are available for LIGO-Livingston by substituting the L1 prefix.