Erica, Wes, Eric, Liz, Jim K, Brian, Herb, Gianluca present
Erica’s LArSoft Indico talk
scope of changes to address issues of data products
just agree on the abstractions needed for
raw::RawDigit, recob::Wire,Hit,Cluster,EndPoint2D, Track
(1) Brian says ArgoNeut should be contacted for raw::RawDigit about Pedestal-related, else removing it is non controversial.
will also change fADC from private to public
(2) recob::Wire eliminate art::Ptr RawDigit, store wireID and channelNum, eliminate SignalType().
wireId will become short, short, short, int.
(3) recob::HitSignal() will be kept; LBNE might need it since they may want to drop wires. Charge() has a bool arg that will be dumped now. Replace with PeakAmplitude()/SummedADC()/IntegratedADC(). Sigma assumes Gaussian, will be replaced by FWHM, time means TDC tick, StartTick and EndTick define the region for hit/hits not start/end for individual hits necessarily, these will replace StartTime and EndTime: will thus dump SigmaStartTime and SigmaEndTime. Will add localIndex which is a short, so one does a for loop, picking up localIndex values 0 through Multiplicity. Drop art::Ptr. HitSignal() to go from private to public; should be a const ref …. or maybe dropped, not clear anyone uses it
(4) recob::Cluster
Lotsa added things, inclduing NHit/NWire, 1/what was suggested, width filled by some alg, replace dt/dW_start, end for dQ/dW, still ongoing… Herb appropriately wonders if all Bruce’s additions are in (Wesley confirmed they are).
(5) recob::EndPoint2D
Introduce a new class Vertex2D, as a reconstructed 2D vertex; move Strength() method here, Multiplicty() to indicate how many clusters/tracks involved. Drop Charge. Besides, define EndPoint2D as a purely geometric entity, representing a (wire x time) coordinate.
(6) recob::Track
Bezier is a continuous trajectory and derives from Track.
Proposal is to add a trajectory class to represent continuous trajectory
use Track + trajectory class for Bezier
use Track + created-on-the-fly linear interp of traj points for Kalman
Add a Momentum object
Remove fit momentum from Track
Keep trajectory points in Track
Decision cleared by those (remaining) in the room to allow 1-3,5 above. 4,6 face continued discussion.
Jim K asks for class diagrams to understand scope and functionality, e.g., for Assns.
adjourned.
There are minutes attached to this event.
Show them.