You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
kafka_event_aggregator: do wiring table mapping #41
It's sort of being done in DSGs rust-data-stream-processor at the moment - but not clear that that's the best place because:
It makes it impossible to configure a new wiring table from IBEX
SANDALS-2 or other instruments may not have a rust-data-stream-processor at all
I'd suggest a more suitable place might be the event_aggregator, we would need some schema to pass through the wiring table from kdae_control to event_aggregator.
Discussed with DSG: there are two separate things needed, wiring-table mapping and spectrum-mapping.
For wiring table mapping, agreed to try to do "as close to the hardware as possible". This means in the detector FPGA for Sandals-2, and in data-stream-processor (aka udp2kafka) for HRPD-X. This will be configured by a .csv or similar format table that DSG will propose a format for.
TODO:
DSG to propose a .csv or similar format for wiring tables
ECG to propose a flatbuffers schema to transport that information
For spectra table mapping, we had a meeting and the consensus was that each consumer may need it's own spectrum-table mapping - Potentially independently configured.
TODO:
ECG to figure out how this works and how to pipe spectra-mapping information down to each consumer that needs it
Activity
[-]kafka event aggregator: do wiring table mapping[/-][+]`kafka_event_aggregator`: do wiring table mapping[/+]Tom-Willemsen commentedon Apr 10, 2026
It's sort of being done in DSGs
rust-data-stream-processorat the moment - but not clear that that's the best place because:rust-data-stream-processorat allI'd suggest a more suitable place might be the
event_aggregator, we would need some schema to pass through the wiring table fromkdae_controltoevent_aggregator.Tom-Willemsen commentedon Apr 16, 2026
Discussed with DSG: there are two separate things needed, wiring-table mapping and spectrum-mapping.
For wiring table mapping, agreed to try to do "as close to the hardware as possible". This means in the detector FPGA for Sandals-2, and in
data-stream-processor(aka udp2kafka) for HRPD-X. This will be configured by a.csvor similar format table that DSG will propose a format for.TODO:
.csvor similar format for wiring tablesFor spectra table mapping, we had a meeting and the consensus was that each consumer may need it's own spectrum-table mapping - Potentially independently configured.
TODO:
Tom-Willemsen commentedon Jun 9, 2026
Latest discussion with DSG:
event_udp_to_kafkato do mapping/decoding of events based on that identifier rather than a wiring table / IP address combinationDefer this ticket until after DSG have implemented this in the frame header, to see to what extent we still need wiring tables at all.
Tom-Willemsen commentedon Aug 20, 2026
To keep all discussion in one place I'll close this in favour of #27 while the whole mechanism is still in flux.