Apply a simple cut of triggers before uploading them to GraceDB. In implementation, check if any event has a FAR less than 1e-6Hz. If so, examine if its coherent network SNR is less than a defined value (default is 8). If so, do not upload it to GraceDB.

postcoh_finalsink.py: turn off skymap generation from inside pipeline; do it using lvalert from gracedb

this will immedialy return and will not introduce latency for following triggers.
between gracedb event appearance and spiir.fits.gz showup, there is about 17 seconds for several test events.
also change the chisq_ratio_thresh to 10.

add constraint on single far veto test, the single fars need to be > 0.0 to be valid
more robust on time extraction from combine_stats filename

if we want to annote offline to a gracedb upload, set this option;
postcoh_finalsink.py: change the subprocesses of ls and grep to os.listdir

postcoh_finalsink.py: remove logging when there is no events as this causes too many outputs, 12MB for half day

gstlal_inspiral_postcohspiir_online: add options --finalsink-singlefar-veto-thresh for single far vetoes

postcoh_finalsink.py: change back that the single FAR vetoes will only apply to events that satisfy superevent thresh

postcoh, cohfar_accumbackground, postcoh_finalsink: add the first entry to the output table received by postcoh_finalsink to indicate participating ifos. This entry will not be output to the final pipeline trigger table.

gstlal_inspiral_postcohspiir_online, postcoh_finalsink.py: set the default value of fapupdater_output_fname = cohfar_assignfar_input_fname
the trigger channel name to be set from options --channel-name instead of hard-coded names

spiirparts.py: rename uni_datasource to snglrate_datasource

replace gracedb new interface

add SegmentDocument to snapshot segment information used for daily plot;
add combine_stats function in FAPUpdater, so we can remove the combine_stats.sh program;
gstlal_inspiral_postcohspiir_online: rename finalsink-background options to be finalsink-fapupdater to be clearer