DM-41717: Prompt Processing observing support for AuxTel week 46 #98
Conversation
The HTMs are difficult to derive by hand from the next_visit metadata, but are essential to debugging refcat loading problems.
The tract and patch are difficult to derive by hand from the nextVisit metadata, but are essential to debugging template loading problems.
kfindeisen
force-pushed
the
tickets/DM-41717b
branch
from
df2aec3 to
d9905bd
Compare
| private_seqNum: int = 0 # queue-specific counter. Not the same as Butler/RubinTV seq_num | ||
| private_rcvStamp: float = 0.0 | ||
| private_sndStamp: float = 0.0 | ||
| private_sndStamp: float = 0.0 # time of visit publication; TAI in unix seconds |
There was a problem hiding this comment.
Trying to dig out again what these private_* fields mean and thinking whether adding more info would be useful. Just dumping them here
SAL timestamps are added automatically by SAL and have the private_ prefix. In particular, the private_sndStamp timestamp is the time at which the message was published by the CSC in TAI, which should be very close to the time at which the telemetry was measured. In fact, many telemetry topics rely on the private_sndStamp timestamp for that.
The private_sndStamp timestamp is also converted to UTC and recorded in the private_efdStamp field, which is used as the time index in the EFD.
private_sndStamp, which should be the TAI instant when the SAL message was issued from the CSC, presumably only a short time after the actual event.
private_efdStamp should be the same value in UTC
And slack thread
private_sndStamp ... is the time the sample was published.. this is TAI but is converted to UTC (and stored in private_efdStamp
most topics will rely on private_sndStamp as also being the time the measurement was performed..
There was a problem hiding this comment.
After reading them again, I think what you've put in the comments are perfect summary for prompt processing context.
There was a problem hiding this comment.
Yes, I used the Sasquatch page as my source.
Though given the example of private_seqNum, we should probably avoid using the private_ fields if possible. 😬
Many fields have different meanings from Butler or RubinTV fields of the same names; explain the known differences.
kfindeisen
force-pushed
the
tickets/DM-41717b
branch
from
d9905bd to
dab60ae
Compare
|
(Oops, apparently you can review a PR that's "not ready for review"!) |
This is a running PR to test Prompt Processing fixes and the
BASE_TAG_LISTsystem. It uses a different branch name from #97 to ensure thatDM-41717-d_2023_11_14is not overwritten.