Automation

OSETI AUTOMATION

This note discusses the automation of the oseti observations. It uses similar software to that of the BAIT project.

REQUEST FILE

The core to the scheme is a request file which specifies the nature of the observation. The request looks like a FITS header and specifies a typical observation.

  OBSERVER='robot'
  EXPTIME=100
  OBJECT='HR6806'
  PROCEDUR='oseti'
  RA='18:09:37.4'
  DEC='+38:27:27'
  EPOCH=2000
  MAG=6.40
  END

The RA, DEC, EPOCH keywords specify where the telescope will point. The PROCEDUR (the bizarre spelling results from the 8 character limitation of the FITS standard) keyword specifies the name of the program to be executed. The EXPTIME (or OBSTIME) keyword specifies how long the observation will take to execute (for purposes of scheduling it). The OBJECT string is stuck into the FITS header of the resulting file. Other keywords such as MAG might be used to set things in the procedure.

A very crude program mkrequest.c to generate request files from a target list can be found in ~treffers/seti/lists.

PROCEDURE

The procedure is the program or script that will be executed for each object. There are several things to bear in mind when developing procedures.

Execute the openup each time
They should exit with a non-zero exit code if something goes wrong
All keywords in the request file will be placed into the environment
A unique id will be generated as the DATID environment variable

Each observation is assigned a unique id code specified by the DATID enviroment variable. This id is of derived from the time of observation and looks like May0vxmy or Mar0ejvp. Details can be found in the description of the mkid command.

A very crude draft of the the oseti procedure written as a C-Shell script follows: It opens the observatory, points the telescope, finds and centers on the star and takes the data. Ultimately writing the data into the file with a name specified by the DATID environment variable.

  #!/bin/csh
  # procedure to make an optical seti observation
  echo running oseti
  openup
  if ($status ) exit 1
  tx point ra=$RA dec=$DEC epoch=$EPOCH | grep done
  if ($status ) exit 1
  tx dac dac0 dac0=100 dac1=100
  spiral
  if ($status ) exit 1
  center back=300 
  if ($status ) exit 1
  adjust
  if ($status ) exit 1
  qseti outfile=$DATID nsamples=1000 rate=20
  if ($status ) exit 1
  putsym file=$DATID.fits object=\'$OBJECT\'

Notice: that the keywords from the request file are transferred into action via the environment (e.g. $RA means the value of the RA environment variable).

EXECUTING A REQUEST FILE

A request file can be executed using the telco command. For example to execute the HR6806.rqs file type.

telco datid verbose HR6806.rqs

SCHEDULING

The group of request files can be automatically scheduled using the insgen scheduling program.

To run the scheduler you must do the following:

All requests must be placed into the directory ~thirty/targets
As user thirty type: startins

Note: The startins script is somewhat bizarre and I I recommend reading the source script before executing it. Once fired up the show ins command can be used to look at what is happening.

The startins script starts insgen as a program running as a separate process writing its output into the file insgen.log. To modify its behavior commands are transmitted to it via the tin command. With it you can:

NOTE: some tin requests may take 5-10 seconds to execute.

add new requests to the queue: tin newreqs xxx.rqs yyy.rqs
interrupt the observations: tin interrupt
end the scheduler: tin shutdown

SIMULATIONS

You can simulate the insgen scheduler (without moving the telescope or anything in the observatory) in your own directory (user in this example) by the following:

create targets directory under your root
copy (or link) ~thirty/ait.config to your home directory
setenv HOME_DIR ~user in all windows for your simulation
in one window in targets directory: tin wait newreqs *.rqs
in another window type: insgen sim
be patient and watch the output of insgen

There is a syntax checker called accepter for request files that points out some of the more obvious errors. Try it with no worry of moving anything by typing.

accepter myname.rqs

You can do some simple tests of the request file without moving anything by typing

telco shell datid myname.rqs

This command will pop up an xterm window with the environment loaded with the stuff from the request file. Type env to see what variables are in the environment. You may then type the program name of the command to be executed (watch out this might move the telescope). Or type other procedures that might use this environment.

DATA FILES AND LOGS

Each days observation is traditionally kept in the directory with the date (e.g. /data/May21 or /data/Apr18). In automatic operation the FITS files will be found there as well as the files

insgen.log - the log of the insgen scheduler
tserver.log - the log of the tserver server program

All output written by scripts whether written to stdout or stderr will be put into the file insgen.log.

You may want to inspect them with the show command.

To look watch the output of insgen: show ins
To look watch the output of tserver: show log
To see what requests are scheduled: show list

[ Seti index | Leuschner Observatory ]

Last Revised June 26, 2000