This code base is used to search for ADC/DAC overflows in LIGO instruments. Developed by TJ Massinger for use in the LVC.
The purpose of this code is to generate an appropriate set of condor dag and sub files to search for ADC/DAC overflows and generate sngl_burst XML files containing triggers. The generated DAG will call the executable gen_overflow_acc_trigs.sh, which is a wrapper around the python script gen_single_channel_acc_trigs.py.
*** This code does not currently perform a segment database query due to the fact that the average user doesn't have a robot cert to perform segment database query from a Condor process. The user must instead run a segment database query and provide the results as an argument to the program. ***
Instructions for running on an LDAS cluster:
source /home/detchar/opt/gwpysoft/etc/gwpy-user-env.sh
Results of the segment database query should be piped through ligolw_print and stored in a whitespace delimited text file.
Example call:
ligolw_segment_query_dqsegdb --segment-url https://dqsegdb5.phy.syr.edu --query-segments --include-segments H1:DMT-DC_READOUT_LOCKED --gps-start-time 1117742416 --gps-end-time 1117747816 | ligolw_print -t segment -c start_time -c end_time -d ' ' > segment_list.txt
The resulting segment_list.txt should look like:
1117742416 1117743020
1117745521 1117747816
This script will go through the entire process of generating an appropriate condor DAG that should be submit-ready. It will generate a directory structure containing log files and executables in the designated output directory. All directories and segment lists should be full paths. The padding argument will remove a number of seconds from the end of each science segment. The example below uses a 10-second pre-lockloss padding to avoid seeing overflows that result from lock loss transients. If you do not want any padding, enter a zero as the last argument.
Syntax:
./gen_overflow_acc_condor.sh {start_gps} {end_gps} {L,H} {build directory} {trigger directory} {segment list} {padding}
Example call:
./gen_overflow_acc_condor.sh 1117742416 1117747816 H /home/tjmassin/git_repos/adc-overflow/dev/acc/ /home/tjmassin/temp_triggers/ /home/tjmassin/git_repos/testing-adc-overflow/adc-overflow/dev/acc/segment_list.txt 10
The process is as follows:
a. Set up a directory structure to store all information regarding models and overflow channels, as well as condor dag, sub, and log files. The location of this structure is specified by the {build directory} argument
b. Grab model-wide cumulative overflow channels to see if any channel in a given model has overflowed during the time period of interest. This is done using plot_overflow_accum.py.
c. Generate a list of models that have at least one overflow.
d. Generate a list of channels contained in those overflowing models and mark them for further follow-up.
e. Generate condor DAG and sub files that allow the search code to farm the computation out channel-by-channel.
f. Copy over the search executables to the condor run directory and create log file directory.
The condor files will be stored in {build directory}/condor_dag/
Example submit:
condor_submit_dag ADC_1117742416_1117747816.dag
Logging information will be stored in the directory "log_{start_gps}_ADC"