PyCBC offline multi ifo coincident search

PyCBC offline multi ifo coincident search

To Do Table

Issue	description	Assigned to	Status
Coincidence finding	link to code	Marton	Done (#2449)
Foreground censor	Remove coincs from 'closed box' single-ifo results plotted / followed up	Marton	Done #2568
Default multi-ifo statistic	Calculate simple ranking stat and IFAR for each type of coinc, combine IFAR via trials factor	Gareth	Done (part of #2473)
Improved multi-ifo statistic:noise & network sensitivity model	Calculate and implement noise element of multi-ifo ranking statistic	Gareth	see Multiifo-Ranking-Statistic-Development for components of this
Improved multi-ifo statistic:signal model	Calculate and implement signal element of multi-ifo ranking statistic	Gareth/+?
Statmap	Combine Coincs to map between stat and fap and hierarchical removal	Marton	Done
Statmap Inj	equivalent of statmap, but for injinj/fullinj/injfull	Marton	Done (#2474)
Combine_statmap	Combine coincidences from different combinations	Gareth	Done #2473
Add combine_statmap to workflow	above code is currently not included in workflow	Marton	Done #2539
hdf_injfind	match injections to times of coincidences	Gareth	Done #2499
Add hdf_injfind to workflow	make it possible to give --ifos as an input to hdf_injfind	Marton	Done #2539
Plotting/result display codes - list incomplete page_foundmissed page_segtable page_coinc_snrchi	Codes may be currently hard-coded for H1L1 - see 'Failing plotting codes' below	Gareth + Ian	foundmissed merged (#2516) segtable Done (#2519) Remaining section 2&3 plots DONE (#2557) Ian working on Section 4 & 7 integration (see list of failures below) Section 5 and 6 plots not yet integrated Summary page not yet integrated
Add plots/tables to workflow	?	Some plots present, in progress
Fix storage format of coinc segments in coinc code and jobs downstream	Don't have redundant/misleading information duplicated between different datasets	Marton (Gareth to check implications for downstream code)	Done, PR #2517

---

Failing plotting codes

The following plotting codes are currently known to fail.

• pycbc_foreground_minifollowup - This runs once and sets up the minifollowups. Fixing this is priority amongst these as there will likely also be failures within the minifollowups dax. Looks like it's not understanding the new time format. Started by Marton #2579
• pycbc_page_foreground - This runs once and dumps the list of triggers in a table format. Looks like it's not understanding the new time format.
• pycbc_page_ifar - This runs a bunch of times to make various counts vs FAR plots. This appears hardcoded for H1L1, so fails on the individual STATMAP files. It also fails on the combined STATMAP job as a bunch of the background stuff is not stored there. Not sure what plot we might want to make on the combined file.
• pycbc_page_snrifar - This also runs a bunch of times and makes various statistic vs FAR plots. Some instances of this are running, but it fails when running on the combined STATMAP file. Might need to think about what we actually want in that case as not all of the background information is now stored.
• pycbc_page_snrratehist - Same situation as pycbc_page_snrifar

To see failures, if you have the current version of the code installed in CIT go to this directory:

/home/spxiwh/aLIGO/O3/devel/offline_multiifo/DEBUG_JOBS

and run . ${EXECUTABLE_OF_CHOICE}_failure.sh. The error message can be viewed from within this directory, but you can also copy the whole thing away to aid debugging.