User Parameters - Data Location & Beam SelectionΒΆ

The following parameters determine where the measurement sets are that hold the data to be processed. The default behaviour is to read the MS from the given Scheduling Block (SB) directory. If SB numbers are given, this is where to look. The location for these SB directories is given by DIR_SB. If SB numbers are not provided, however, the MS_INPUT variables give the filename for the relevant MS. The full path must be specified here.

There are two modes in which ingest has been run to create measurement sets in the SB directory. The usual way has been to create a single file for the entire observation. Recently, ingest has started producing a single MS per beam. The pipeline is now compatible with both these modes. For the file-per-beam mode, it assumes there is the same number of beams as in the footprint. If there is not, the pipeline will currently exit with an error message.

In the file-per-beam mode, the metadata for the observation is taken from a single MS (although they should all be the same for a given observation).

If no channel or scan selection is being made (that is, using CHANNEL_RANGE_SCIENCE or SCAN_SELECTION_SCIENCE), there is only a single field in the dataset, and the observation used the MS-per-beam mode, then instead of splitting the dataset the pipeline uses a parallel copy (using the dcp utility) to copy the archive MS to the output directory. This should run faster than using mssplit.

To use this mode, you will need to have your SSH keys set appropriately in your ~/.ssh directory. The copying is run on hpc-data, under an SSH connection, and you need to be able to connect to it without providing a password. If there is an issue with your SSH keys, or you know hpc-data is unavailable, you can use regular cp by setting USE_DCP_TO_COPY_MS=false. Note that this will be slower than the dcp option.

The bandpass calibration datasets are treated the same way, although mssplit is always used (since the relevant scan needs to be split out).

It may be that the SB you wish to process has been removed from /astro. You can follow the instructions on Commissioning Archive Platform to request it be reinstated. If you have already started processing it and wish to continue, you can specify the measurement set name it would have (this forms the metadata/mslist_ filename), using the parameters MS_INPUT_1934 or MS_INPUT_SCIENCE, and the scripts will recognise the metadata file and continue. For anything to be done, it requires you to have already run the splitting, so that there is a local copy of the individual beam data.

Variable Default Description
DIR_SB /astro/askaprt/askapops/askap-scheduling-blocks Location (on galaxy) of the scheduling blocks. This is used when specifying SB numbers - the default is the standard location for ASKAP operation. If blank (“”), the MS used will be given by MS_INPUT_SCIENCE and MS_INPUT_1934.
SB_1934 no default SB number for the 1934-638 calibration observation.
MS_INPUT_1934 no default MS for the 1934-638 observation. Ignored if the SB number is provided.
SB_SCIENCE no default SB number for the science field observation.
MS_INPUT_SCIENCE no default MS for the science field. Ignored if the SB number is provided.
USE_DCP_TO_COPY_MS true Whether to use dcp via ssh to hpc-data to run the copying of the input data. If false, ordinary cp is used without ssh.

These parameters determine which beams in the data to process. The beams can be listed explicitly via the BEAMLIST parameter, given as a comma-separated list of beams and beam ranges, such as “”0,1,4,7,8-10”. Alternatively (and if BEAMLIST is not provided), the BEAM_MIN and BEAM_MAX parameters are used to specify an inclusive range. Note that in all cases the beams are numbered from zero (ie. beam 0 is the first beam in the measurement set).

The beams used for the bandpass calibrator will be all beams up to the maximum beam requested for the science dataset (either BEAM_MAX or the largest number in BEAMLIST).

Variable Default Description
BEAMLIST “” A comma-separated list of beams and beam ranges (for instance “0,1,4,7,8-10”), to specify the set of beams than should be processed from the science observation. Beam numbers are 0-based.
BEAM_MIN 0 First beam number (0-based). Used when BEAMLIST is not given. All beams from BEAM_MIN to BEAM_MAX inclusive are used for the science processing.
BEAM_MAX 35 Final beam number (0-based), to go with BEAM_MIN.

Previous topic

User Parameters - Pipeline & job control

Next topic

User Parameters - Bandpass Calibration

This Page