Currently, the completion of n_images written to the drive (e.g. 3 for triple imaging) triggers the watchdog to associate a run_id. The time used to validate the association is the trigger time (local) and the run_id generation time (breadboard server). The difference in these times is currently ~3-8s, with a cutoff at 10s.
The latency may be improved by changing the timestamp to be registered on the writing of any image (e.g. the first of three in triple imaging). See line 106, monitor_watchfolder method.
Currently, the completion of n_images written to the drive (e.g. 3 for triple imaging) triggers the watchdog to associate a run_id. The time used to validate the association is the trigger time (local) and the run_id generation time (breadboard server). The difference in these times is currently ~3-8s, with a cutoff at 10s.
The latency may be improved by changing the timestamp to be registered on the writing of any image (e.g. the first of three in triple imaging). See line 106, monitor_watchfolder method.