Draft: proof-of-concept for a new tupling algorithm
Preliminary proof-of-concept version of a new nTupling algorithm independent of DecayTreeTuple.
Starts to address https://gitlab.cern.ch/lhcb-dpa/project/-/issues/6.
Replaced by Analysis!767 (merged) in Analysis.
Merge request reports
Activity
added Tuples label
I have a few comments on this, but perhaps that would be premature on this MR, I haven't been following the plan.
Is there a design document, or something like that, for how you see the DecayTreeTuple replacement? I noticed that the stuff here is based on
IDVAlgorithm
, which is a bit of a beast, making lots of assumptions on how things work. Maybe this is the way you'd like to go, I'm just wondering about what's been considered./cc @olupton, who developed an ntuple dumper for use in HLT1.
If I may add: discussions are being held at WP3 meetings, and RTA collab is more than welcome :-). BTW, the fact that we have strong overlaps and inter-dependencies is the main reason why we should have joint meetings every now and then, and we will have one on July 24th where things such as the new tupling algo framework should certainly be presented to the wider RTA community ...
Yes you're right that many bits of the software will have to evolve as we settle on the event model, selectors, etc. But we have to move in parallel, hence this very welcome step
- Resolved by Eduardo Rodrigues
Just to note that I did point out that
DumpContainer
exists in one of the DPA WP3 meetings (It's even in the minutes ).I don't know if anyone tried it out after that. Nobody contacted me, which I guess means either "the documentation is perfect / it was trivial to use", or "nobody tried"
FWIW I think it is compulsory to have tupling code that is extremely closely aligned with the selection code. A tuple whose columns do not match the quantities that selections cut on is only useful in a rather narrow set of circumstances.
mentioned in issue Moore#198 (closed)
mentioned in issue Moore#200 (closed)
added 1 commit
- 3d509302 - Moving FunTuple to FunTupleModern i.e usage of Gaudi Functional
- Resolved by Abhijit Mathad
- Resolved by Abhijit Mathad
- Resolved by Abhijit Mathad
- Resolved by Abhijit Mathad
- Resolved by Abhijit Mathad
- Resolved by Abhijit Mathad
- Resolved by Abhijit Mathad