[Rivet] Boosts in FinalStates/ParticleFinders

[Andy Buckley]

On 08/04/16 09:58, Leif Lönnblad wrote:
> I think the TransformationFinder idea sounds like the most reasonable
> idea. Can we find another use case for manipulating particles in the
> ParticleFinder?

Yes, I think another use-case is needed... at least to justify immediate 
action. But thanks for the input -- I also think it is the way forward.

In the 2.5.0beta1 that is about ready for release, I've just provided a 
few more helper functions to make it easy to get the beam CMS transform 
and apply it to Particle and Jet objects. For example, if you wanted to 
be inefficient and recompute the transform for every particle, you could 
do this:

Particle pprime = p.transformBy(beamTransform(event))

It's obvious how to make this more efficient. And hopefully it'll be 
enough for Tanguy & co for now, until we've engineered something a bit 
more joined-up.

> The DISKinematics class was something I added in the very early days of
> Rivet since I wanted to compare some stuff with HZTOOL. As far as I can
> see it is not used at all, and should maybe be removed completely.

I've had a look and think this can probably be re-engineered a bit, to 
make it and DISLepton more flexible. This would be best tied in with a 
push to get the main HERA analyses (finally) Rivetised -- I'll prod some 
DESY people about that again ;-)  I don't suppose anyone has a 
convenient stash of e+-p+ events lying around for testing purposes?

Andy



> On 2016-04-07 22:59, Andy Buckley wrote:
>> One of the things that we were asked to provide in Rivet recently was
>> automatic frame transformation of projections, e.g. to return particles
>> or jets computed in the beam centre-of-mass.
>>
>> I've added some tools to the Beam projection to compute (I hope
>> correctly...) the appropriate boost for that case, but it's not clear
>> how to propagate it to the projections. I can't just make a
>> BoostedFinalState, then pass that (maybe wrapped in another FinalState)
>> to FastJets, for example, because then the tagging ghosts mode via an
>> internal HeavyHadrons projection will be in a different frame from the
>> jets they're being clustered into: not good.
>>
>> Any ideas how we should go about doing this? We could maybe go down the
>> rabbit-hole above, but I feel it gets very complex very quickly. Another
>> way is to add an optional TransformationFinder projection argument to
>> the ParticleFinder base class, so that the boost is computed
>> event-by-event by that projection before being applied to the found
>> particles. BeamBoostFinder would then be one of several implementations
>> of TransformationFinder, and perhaps the existing but little-used
>> DISKinematics projection would also be retrofitted that way.
>>
>> This feels quite neat to me, but isn't quite as minor a bit of surgery
>> as I'd hoped for. Any thoughts / much easier ways that I've not seen?
>>
>> Cheers!
>> Andy
>>
>
>
> _______________________________________________
> Rivet mailing list
> Rivet at projects.hepforge.org
> https://www.hepforge.org/lists/listinfo/rivet


-- 
Dr Andy Buckley, Lecturer / Royal Society University Research Fellow
Particle Physics Expt Group, University of Glasgow