Size: 10014
Comment:
|
Size: 10426
Comment:
|
Deletions are marked like this. | Additions are marked like this. |
Line 123: | Line 123: |
* The output of Omicron is temporary stored in {{{k1sum1:/mnt/triggers/K1/}}}. By crontab every 30 minutes, this directory is copied in {{{k1det0fs0/mnt/detfs0/triggers/K1/}}}, which is mounted at k1sum0 as {{{~/triggers}}}. | * The output of Omicron is temporary stored in {{{k1sum1:/mnt/triggers/K1/}}}. By crontab every 30 minutes, this directory is copied in {{{k1detfs0/mnt/detfs0/triggers/K1/}}}, which is mounted at k1sum0 as {{{~/triggers}}}. |
Line 151: | Line 151: |
* If the following error message occurred, the cause might be list of ffl file. Check whether file list in cache is ordered or duplicated. [[https://klog.icrr.u-tokyo.ac.jp/osl/?r=18336|see k-log]] {{{ Segments::CheckConsistency: segments should be time sorted (GWOLLUM convention) -- 1315847008.00000>1315846912.0000 }}} * make correct ffl file {{{/users/DET/tools/Cache/Script/makeCache.sh 13158}}} |
Omicron
O3GK log
Some report on O3GK data processing in detail.
Installation
- The Omicron code is managed on the Virgo SVN and it should be able to be installed by CMT system.
- But when we use the CMT in order to install Omicron, dependency conflict is occurred and we cannot install it correctly.
F. Robinet helped us to install it on the k1sum0 and we can use copy of it on the another Debian 9 systems.
T. Yamamoto had already copied Omicron and the related packages on k1nfs0:/export/virgo/virgosoft.
- So Omicron can be used by mounting NFS area on the Debian 9 systems.
> grep virgsoft /etc/fstab 10.68.10.10:/export/virgo/virgosoft /home/controls/opt/virgosoft nfs rw,bf,soft 0 0
- So Omicron can be used by mounting NFS area on the Debian 9 systems.
How to use on Kamioka server
- Before using omicron, we must set the environmental variables by
> source ~controls/opt/virgosoft/environment.sh
- Omicron is available as a command line tool on k1sum0, k1sum1, and another test system of Debian 9.
- There are three ways to use omicron
case1> omicron start_gps stop_gps option_file case2> omicron segment_file option_file ### TY don't know much about this way. case3> omicron central_gps option_file
Some examples of a form of option_file are shown in [1] and [2].
- There are three ways to use omicron
- Omicron is available as a GUI launched from MEDM on the client workstations.
Sitemap -> Commissioning -> Glitch pipeline -> Omicron
How to use on Kashiwa server
- Before using omicron, we must set the environmental variables by
> source /gpfs/ligo/sw/conda/etc/profile.d/conda.sh > conda activate igwn-py38
If you want to use this environment on new machine, https://computing.docs.ligo.org/conda/ is helpful.
- Omicron is available as a command line tool on m31-01, m31-02.
- There are three ways to use omicron
case1> omicron start_gps stop_gps option_file case2> omicron segment_file option_file ### TY don't know much about this way. case3> omicron central_gps option_file
Some examples of a form of option_file are shown in [1] and [2].
- There are three ways to use omicron
Trigger generation for DailySummary
DailySummaryPage uses the trigger files using gwtrigfind for constructing the Omicron summary page.
T. Yamamoto modified gwtrigfind because our computer environment is slightly different from LIGO's one.
In LIGO convention [3], trigger files should be generated in the /home/detchar/triggers
We have only controls user on k1sum0 and another computers.
--- /home/controls/opt/summary-2.7/lib/python2.7/site-packages/gwtrigfind/core.py.bak 2018-10-19 09:12:45.375654000 +0900 +++ /home/controls/opt/summary-2.7/lib/python2.7/site-packages/gwtrigfind/core.py 2019-01-23 18:29:06.224172322 +0900 @@ -161,12 +161,12 @@ def find_detchar_files(channel, start, e ifo, name = _format_channel_name(channel).split('-', 1) # find base path relative to O1 or O2 formatting if start >= OMICRON_O2_EPOCH: - base = os.path.join(os.path.sep, 'home', 'detchar', 'triggers') + base = os.path.join(os.path.sep, 'home', 'controls', 'triggers') tag = etg.upper() dirtag = '%s_%s' % (name, tag) else: epoch = '*' - base = os.path.join(os.path.sep, 'home', 'detchar', 'triggers', '*') + base = os.path.join(os.path.sep, 'home', 'controls', 'triggers', '*') tag = etg.title() dirtag = '%s_%s' % (str(channel).split(':', 1)[1], tag)
(2021/06/17 yuzu) Because the conda environment was updated, I applied this patch for k1sum0 again. The difference is here.
--- /home/controls/bin/miniconda2/envs/ligo-summary-3.7/lib/python3.7/site-packages/gwtrigfind/core.py 2021-06-17 16:15:21.718561137 +0900 +++ /home/controls/bin/miniconda2/envs/ligo-summary-3.7/lib/python3.7/site-packages/gwtrigfind/core.py_bak 2021-06-17 16:12:06.961886182 +0900 @@ -161,12 +161,12 @@ def find_detchar_files(channel, start, e ifo, name = _format_channel_name(channel).split('-', 1) # find base path relative to O1 or O2 formatting if start >= OMICRON_O2_EPOCH: - base = os.path.join(os.path.sep, 'home', 'controls', 'triggers') + base = os.path.join(os.path.sep, 'home', 'detchar', 'triggers') tag = etg.upper() dirtag = '%s_%s' % (name, tag) else: epoch = '*' - base = os.path.join(os.path.sep, 'home', 'controls', 'triggers', '*') + base = os.path.join(os.path.sep, 'home', 'detchar', 'triggers', '*') tag = etg.title() dirtag = '%s_%s' % (str(channel).split(':', 1)[1], tag)
The Omicron job is automated using crontab on k1sum1. summary_*.txt is for * Hz sampling rate channels.
> crontab -l | grep Omicron 7-59/15 * * * * env - `cat /users/DET/tools/Omicron/etc/sum1_cron-omicron.env` /users/DET/tools/Omicron/Script/automation_segment.sh /users/DET/tools/Omicron/Parameter/summary_2048.txt > /tmp/auto-omicron-2048_segment.log 2>&1 9-59/15 * * * * env - `cat /users/DET/tools/Omicron/etc/sum1_cron-omicron.env` /users/DET/tools/Omicron/Script/automation_segment.sh /users/DET/tools/Omicron/Parameter/summary_1024.txt > /tmp/auto-omicron-1024_segment.log 2>&1 11-59/15 * * * * env - `cat /users/DET/tools/Omicron/etc/sum1_cron-omicron.env` /users/DET/tools/Omicron/Script/automation_segment.sh /users/DET/tools/Omicron/Parameter/summary_512.txt > /tmp/auto-omicron-512_segment.log 2>&1 13-59/15 * * * * env - `cat /users/DET/tools/Omicron/etc/sum1_cron-omicron.env` /users/DET/tools/Omicron/Script/automation_segment.sh /users/DET/tools/Omicron/Parameter/summary_256.txt > /tmp/auto-omicron-256_segment.log 2>&1 3-59/15 * * * * env - `cat /users/DET/tools/Omicron/etc/sum1_cron-omicron.env` /users/DET/tools/Omicron/Script/automation_segment.sh /users/DET/tools/Omicron/Parameter/summary_4096.txt > /tmp/auto-omicron-4096_segment.log 2>&1
- When you want to add some channels for the Omicron search,
1) add channel name to DATA CHANNELS line of /users/DET/tools/Omicron/Parameter/summary_*.txt. Please choose appropriate * referring sampling rate.
or
2) add new job to crontab> crontab -e 5-59/15 * * * * env - `cat /users/DET/tools/Omicron/etc/sum1_cron-omicron.env` /users/DET/tools/Omicron/Script/automation.sh /users/DET/tools/Omicron/Parameter/summary_test.txt > /tmp/auto-omicron.log 2>&1 5-59/15 * * * * env - `cat /users/DET/tools/Omicron/etc/sum1_cron-omicron.env` /users/DET/tools/Omicron/Script/automation.sh /path/to/your_parameter.txt > /tmp/auto-omicron2.log 2>&1
In order to use trigger results on DailySummaryPage, use same OUTPUT parameters as
OUTPUT DIRECTORY /home/controls/triggers/tmp OUTPUT PRODUCTS triggers OUTPUT FORMAT xml OUTPUT VERBOSITY 0
The output of Omicron is temporary stored in k1sum1:/mnt/triggers/K1/. By crontab every 30 minutes, this directory is copied in k1detfs0/mnt/detfs0/triggers/K1/, which is mounted at k1sum0 as ~/triggers.
Finally, we must modify the INI file of DailySummaryPage.
Trigger generation on Kashiwa server
It is not yet automated, only manual script is available. Command example:
>./manual_segment_kashiwa.sh /home/detchar/git/kagra-detchar/tools/Omicron/Parameter/O3_C20.txt /home/detchar/Segments/K1-DET_FOR_GRB200415A/2020/K1-DET_FOR_GRB200415A_UTC_2020-04-15.txt > log.dat 2>&1
Example of submission file to condor is Omicron/Script/condor.sdf For automation, cache file in parameter file (DATA FFL) has to be modified.
Now directory structure modification command is commented out. mvtoK1.sh (mvtoK1.sdf for condor)
Failure check
- Running omicron job as above
checkOmicron.py (checkOmicron.sdf for condor): please check the file name of L32 locked, L89 f. The output goes in Omicron/Script/checkLog/. succeeded* is OK, failed is due to too high trigger rate, unknown is failure by other reason.
- retry.sh (retry.sdf for condor): It will reprocess according to checkLog/failed* segment files. Now it runs for all the failed* files.
- Go back to 2. and check the result.
Note for developers
- Bug in automation_segment.sh should be fixed.
- Treatment of 0:00:00 : the segment including 0:00:00 maybe separated in segment files. If one of them is shorter than 64 sec, not processed. manual_segment.sh has a hint for this modification.
- MARGIN is different from the O3GK configuration. In O3GK configuration, GPS_END is set earlier by 30 sec. 30 sec is determined by burst analysis margin.
Please modify segment if needed.
- To share the trigger files with other machines,
rsync -av /mnt/triggers/K1/ /mnt/detfs0/triggers/K1/ > /tmp/cp-omicron.log 2>&1 is required on k1sum1. It is temporary commented out in crontab.
- If something went wrong, please check OUTPUT DIRECTORY in the parameter file. It should be /home/controls/triggers/tmp on Kamioka server and /data/detchar/triggers/tmp on Kashiwa server.
If the following error message occurred, the cause might be list of ffl file. Check whether file list in cache is ordered or duplicated. see k-log
Segments::CheckConsistency: segments should be time sorted (GWOLLUM convention) -- 1315847008.00000>1315846912.0000
make correct ffl file /users/DET/tools/Cache/Script/makeCache.sh 13158
Reference
[1] VIR-0135A-15, Omicron: a tool for detector characterization
[2] VIR-0545C-14, Omicron: an algorithm to detect and characterize transient events in gravitational-wave detectors
[3] LIGO-T1300468, DetChar Trigger Handling Conventions