Introduction for Reviewers¶
This page describes how to test run the low-latency multi-messenger analysis pipeline. This pipeline currently focuses on gravitational wave (GW) and high-energy neutrino (HEN) searches. This pipeline listens GW events and upon a trigger initiates searching for HENs and calculates a coincidence significance for spatially and temporally correlated GW and HEN. It then creates a joint skymap for candidate multi-messenger neutrino events. If the significance is higher than the determined threshold for multi-messenger events it sends a multi-messenger event alert.
GWHEN-LLAMA (GWHEN search using Low-Latency Algorithm for Multi-messenger Astrophysics) is a low-latency GW+HEN analysis pipeline that ran during O2 (using previously reviewed offline analysis methods) with the goal of rapidly identifying joint GW+HEN sources. This review targets the upgrades made to the pipeline, the most major of which is an upgrade to the significance calculation algorithm (see documentation section below).
The up-to-date full software documentation for the pipeline can be found
on this site (use the sidebar to navigate) or can be downloaded as a
Testing on Review Server¶
During the test the pipeline will be fully functional (except for sending alerts for high significance events and uploading data products, which it will not do in this testing environment).
There are fake created edge test scenarios for this review whose significance magnitudes can be ordered easily with respect to each other.
Precise GW skymap temporally and spatially coincident with a neutrino event
Coarse GW skymap temporally and spatially coincident with a neutrino event
Precise GW skymap temporally coincident with a neutrino event
Precise GW skymap spatially coincident with a neutrino event
Coarse GW skymap temporally coincident with a neutrino event
Expectations for significances:
a > b > e > c, d ~ 0
Running Fake Cases¶
ssh-ing in to
reviewers.gwhen.com, you can run the fake test
Verbose output will print to the terminal. In particular, you’ll be able to see analysis outputs in ~/.local/share/llama/current_run/, which contains four test triggers (each in their own directory). You can also find these files at http://reviewers.gwhen.com/tests/, which displays the contents of ~/.local/share/llama/current_run/ in your web browser.
Contents of the Output Folder¶
icecube_neutrino_list.jsonThe list of IceCube neutrino triggers in the +/-500s window around the GW event.
lvc_gracedb_event_data.jsonMetadata about the GW trigger pulled from GraceDB. For the test events, it’s a minimal example with only a rho value, though in general this is just the JSON pulled from GraceDB.
lvc_initial_skymap.fits.gzThe LVC skymap selected for the analysis (an injection for the purpose of the review process).
skymap_info.jsonMetadata specifying which skymap to use for lvc_initial_skymap.fits.gz. Can be parsed from a GCN Notice, LVAlert, GraceDB, or manually created by a pipeline operator.
coinc_significance_IceCube-LVC.jsonThe immediate output of the new significance calculation. Contains information on all inputs, search parameters, and outputs. Not designed to be especially human-readable.
coinc_scatterplot_LVC-IceCube.pdfA joint skymap showing the neutrino triggers scattered over the GW skymap in PDF form.
coinc_scatterplot_LVC-IceCube.pngA joint skymap showing the neutrino triggers scattered over the GW skymap in PNG form for upload to GraceDB.
IceCubeNeutrinoList.texThe IceCube neutrinos in a LaTeX table together with their odds ratios calculated for the GW trigger.
IceCubeNeutrinoList.txtSame, but in an ascii format used for GCN circulars.
coinc_scatterplot_summary_LVC-IceCube.texLaTeX file that can later be used to manually generate a summary PDF, coinc_scatterplot_summary_LVC-IceCube.pdf.
lvc_initial_skymap.hdf5The same LVC skymap as above, but converted to an internal HDF5 representation.
The pipeline can also be tested against real cases. Instructions for testing against real cases will be added soon.