Split the propagation-time correction into independent hit-time and time-window flags - #17
Conversation
|
Thanks @jburzy! Is there any strong reason not to do this? |
|
Fair point, I was thinking that it would be better to loudly break existing workflows rather than silently break them, since with Alternatively, if I go with your suggestion but default |
|
Then if you don't mind I'd go for keeping the name of the existing flag and defaulting CorrectTimeWindowForPropagation to True and keep existing workflows alive and unchanged. I see that you closed key4hep#57 in the meantime, so we don't have to worry about consistency of the PRs. |
…me window cut Split the time window cut's propagation time-of-flight correction out of CorrectTimesForPropagation into a new independent flag: - CorrectTimesForPropagation (default false, unchanged): subtract the propagation time-of-flight (radial distance/c) from the stored hit time - CorrectTimeWindowForPropagation (new, default true): subtract the propagation time-of-flight from the hit time when applying the time window cut This allows applying the time window cut on TOF-corrected times without altering the stored hit times. Configurations that set CorrectTimesForPropagation=True behave exactly as before without changes.
00c38f1 to
58eb687
Compare
|
Merged - the tests failures are unrelated |
2 participants
DDPlanarDigihad a singleCorrectTimesForPropagationflag that subtracted thepropagation time-of-flight (radial distance/c) from the hit time before the time
window cut. This coupled two independent choices: whether the time window is defined
relative to the expected arrival time, and whether the stored hit times are
TOF-corrected. In particular, it was impossible to apply the usual TOF-corrected time
window while keeping absolute hit times in the output — which is needed when the
downstream tracking wants to use the physical hit time (e.g. time-aware seeding and
4D track finding).
This PR changes the old behavior of
CorrectTimesForPropagationand adds an additional flag:CorrectTimesForPropagation— subtract the propagation TOF from thestored (digitized) hit time
CorrectTimeWindowForPropagation— subtract the propagation TOF from the hittime only when applying the time window cut, leaving the stored time untouched
Setting both reproduces the old
CorrectTimesForPropagation=Truebehaviour.@madbaron