| Version 11 (modified by , 14 years ago) ( diff ) |
|---|
How to rerun stuff for Mops
Mops likes their data to appear on their datastore, and they like quads to be reprocessed. This is Heather's list of steps:
- queue with chiptool. defaults usually work fine, but use your judgement here.
- queue the quads. They preferably like 1 - 2 and 3 - 4. You can trick out difftool to do it in 2 commands:
difftool -definewarpwarp -input_comment '%visit 2' -template_comment '%visit 1' -input_label 'czw.mopstest.rerun' -template_label 'czw.mopstest.rerun' -dbname gpc1 -set_workdir neb://@HOST@.0/czw/mopstest.rerun -set_label 'czw.mopstest.rerun' difftool -definewarpwarp -input_comment '%visit 4' -template_comment '%visit 3' -input_label 'czw.mopstest.rerun' -template_label 'czw.mopstest.rerun' -dbname gpc1 -set_workdir neb://@HOST@.0/czw/mopstest.rerun -set_label 'czw.mopstest.rerun'
ran into problem with diffs on stacks: I didn't have enough good quality skycells (need 60) to do it on more than half of the exposures, so need -available to queue up what's there.
difftool -definewarpstack -stack_label MD03.refstack -warp_label haf.mops.rerun -dbname gpc1 -set_workdir neb://@HOST@.0/haf.mops.rerun -set_label haf.mops.rerun -set_reduction SWEETSPOT -available
SCH edit, I use for ThreePi data:
difftool -dbname gpc1 -definewarpstack -good_frac 0.2 -warp_label czw.mopstest.rerun -data_group czw.mopstest.rerun -stack_label ecliptic.rp -set_label sch.mops.rerun -set_data_group sch.mops.rerun.20130115 -set_workdir neb://@HOST@.0/sch/mops.rerun.20130115 -available -set_reduction SWEETSPOT_WS -set_dist_group NODIST -rerun -simple -pretend
Remove pretend when you're satisfied
See progress using either http://ippmonitor.ipp.ifa.hawaii.edu/ippMonitor/diffSummary.php with the right lable/dfata group, or
difftool -dbname gpc1 -listrun -label sch.mops.rerun -simple
- publishing- they like stuff to be put in client_id 14, ie, ipp_mops_dev
pubtool -dbname gpc1 -set_label czw.mopstest.rerun -client_id 14 -definerun -label czw.mopstest.rerun
or
pubtool -dbname gpc1 -set_label sch.mops.rerun -client_id 14 -definerun -data_group sch.mops.rerun.20130115
and add that label to publishing pantasks. note that if you omit -label it queues everything to be published. do not do that.
That should be it!
