# H1Trig1 -> Coincidence [label="Trigs sub bank 1"];
# L1Trig1 -> Coincidence [label="Trigs sub bank 1"];
# V1Trig1 -> Coincidence [label="Trigs sub bank 1"];
# H1Trig2 -> Coincidence [label="Trigs sub bank 2"];
# L1Trig2 -> Coincidence [label="Trigs sub bank 2"];
# V1Trig2 -> Coincidence [label="Trigs sub bank 2"];
# H1TrigN -> Coincidence [label="Trigs sub bank N"];
# L1TrigN -> Coincidence [label="Trigs sub bank N"];
# V1TrigN -> Coincidence [label="Trigs sub bank N"];
#
# Coincidence -> SigEst -> Thresh -> EventGen;
#
# EventGen -> gracedb;
#
# }
# @enddot
#
# ### Command line interface
#
# + `--local-frame-caching`
# + `--psd-fft-length` [seconds] (int): The length of the FFT used to used to whiten the data (default is 32 s).
# + `--veto-segments-file` [filename]: Set the name of the LIGO light-weight XML file from which to load vetoes (optional).
# + `--veto-segments-name` [name]: Set the name of the segments to extract from the segment tables and use as the veto list, default = "vetoes".
# + `--nxydump-segment` [start:stop]: Set the time interval to dump from nxydump elments (optional). The default is \":\", i.e. dump all time."
# + `--output` [filename]: Set the name of the LIGO light-weight XML output file *.{xml,xml.gz} or an SQLite database *.sqlite (required). Can be given multiple times. Exactly as many output files must be specified as svd-bank files will be processed (see --svd-bank)
# + `--output-cache` [filename]: Provide a cache of output files. This can be used instead of giving multiple --output options. Cannot be combined with --output.
# + `--reference-psd` [filename]: Instead of measuring the noise spectrum, load the spectrum from this LIGO light-weight XML file (optional).
# + `--track-psd`: Enable dynamic PSD tracking. Always enabled if --reference-psd is not given.
# + `--svd-bank` [filename]: Set the name of the LIGO light-weight XML file from which to load the svd bank for a given instrument in the form ifo:file. These can be given as a comma separated list such as H1:file1,H2:file2,L1:file3 to analyze multiple instruments. This option can be given multiple times, unless --data-source is lvshm or framexmit in which case it must be given exactly once. If given multiple times, the banks will be processed one-by-one, in order. At least one svd bank for at least 2 detectors is required, but see also --svd-bank-cache.
# + `--svd-bank-cache` [filename]: Provide a cache file of svd-bank files. This can be used instead of giving multiple --svd-bank options. Cannot be combined with --svd-bank options.
# + `--time-slide-file` [filename]: Set the name of the xml file to get time slide offsets (required).
# + `--control-peak-time` [seconds] (int): Set a time window in seconds to find peaks in the control signal (optional, default is to disable composite detection statistic)).
# + `--fir-stride` [seconds] (int): Set the length of the fir filter stride in seconds, default = 8.
# + `--ht-gate-threshold` [sigma] (float): Set the threshold on whitened h(t) to excise a glitches in units of standard deviations (optional). If given, exactly as many h(t) thresholds must be set as svd-bank files will be processed (see --svd-bank).
# + `--chisq-type" [type]: Choose the type of chisq computation to perform. Must be one of (autochisq|timeslicechisq). The default is autochisq.
# + `--coincidence-threshold` [seconds] (float): Set the coincidence window in seconds (default = 0.005 s). The light-travel time between instruments will be added automatically in the coincidence test.
# + `--min-instruments` [count] (int): Set the minimum number of instruments that must contribute triggers to form a candidate (default = 2).
# + `--min-log-L` [log likelihood ratio] (float): Discard candidates that get assigned log likelihood ratios below this threshold (default = keep all). When used without --ranking-stat-input, the cut is decided based on an internal approximate ranking statistic.
# + `--write-pipeline` [filename]: Write a DOT graph description of the as-built pipeline to this file (optional). The environment variable GST_DEBUG_DUMP_DOT_DIR must be set for this option to work.
# + `--comment`: Set the string to be recorded in comment and tag columns in various places in the output file (optional).
# + `--check-time-stamps`: Turn on time stamp checking.
# + `--verbose`: Be verbose (optional).
# + `--tmp-space` [path]: Path to a directory suitable for use as a work area while manipulating the database file. The database file will be worked on in this directory, and then moved to the final location when complete. This option is intended to improve performance when running in a networked environment, where there might be a local disk with higher bandwidth than is available to the filesystem on which the final output will reside.
# + `--blind-injections` [filename]: Set the name of an injection file that will be added to the data without saving the sim_inspiral_table or otherwise processing the data differently. Has the effect of having hidden signals in the input data. --injections must not be specified in this case.
# + `--job-tag`: Set the string to identify this job and register the resources it provides on a node. Should be 4 digits of the form 0001, 0002, etc..
# + `--ranking-stat-output` [filename]: Set the name of the file to which to write ranking statistic data collected from triggers (optional). Can be given more than once. If given, exactly as many must be provided as there are --svd-bank options and they will be writen to in order.
# + `--ranking-stat-output-cache` [filename]: Provide a cache of ranking statistic output files. This can be used instead of giving multiple --ranking-stat-output options. Cannot be combined with --ranking-stat-output.
# + `--ranking-stat-input` [filename]: Set the URL from which to load a ranking statistic definition. When this is enabled, signal candidates will have ranking statistic values assigned on-the-fly, and the --min-log-L cut will be applied based on the assigned values. Required when --data-source is lvshm or framexmit; must also set --likelihood-snapshot-interval.
# + `--zerolag-rankingstat-pdf` [url]: Record a histogram of the likelihood ratio ranking statistic values assigned to zero-lag candidates in this XML file. This is used to construct the extinction model and set the overall false-alarm rate normalization during online running. If it does not exist at start-up, a new file will be initialized, otherwise the counts will be added to the file's contents. Required when --data-source is lvshm or framexmit; optional otherwise. If given, exactly as many must be provided as there are --svd-bank options and they will be used in order.
# + `--likelihood-snapshot-interval` [seconds] (float): How often to snapshot candidate and ranking statistic data to disk when running online.
# + `--ranking-stat-pdf` [url]: Set the URL from which to load the ranking statistic PDF. This is used to compute false-alarm probabilities and false-alarm rates and is required for online operation (when --data-source is framexmit or lvshm). It is forbidden for offline operation (all other data sources).
# + `--gracedb-far-threshold` (float): False-alarm rate threshold for gracedb uploads in Hertz (default = do not upload to gracedb).
# + `--gracedb-search`: Name of search to provide in GracedB uploads (default is LowMass).
# + `--gracedb-pipeline`: Name of pipeline to provide in GracedB uploads (default is gstlal).
# + `--gracedb-group`: Gracedb group to which to upload events (default is Test).
# + `--gracedb-service-url`: Override default gracedb service url (optional).
# + `--thinca-interval` [seconds] (float): Set the thinca interval (default = 30 s).
# + `--singles-threshold` [SNR] (float): Set the SNR threshold at which to record single-instrument events in the output (default = +inf, i.e., don't retain singles).
#
# ### Review Status
#
# | Names | Hash | Date | Diff to Head of Master |
