[Rivet] Rivet 2.4 and xsec/event counter information

[Andy Buckley]

On 05/11/15 13:11, Daniel Rauch wrote:
> Dear Andy,
>
> I was quite excited to see the second-to-last bullet point in the
> release announcement of Rivet 2.4.0:
>
>     Cross-section and event counter are now written out by default --
>     these should be very useful for run merging of multiparton
>     generation with separate partonic multiplicity samples, cf. MadGraph
>     and Alpgen. The Rivet scripts do not yet use these, so please get in
>     touch if it sounds useful! You'll need the latest YODA to read them
>     correctly.
>
> This sounds very useful to me indeed. During my master's thesis I
> modified ThePEG/Herwig++ and Rivet 1.9.0 to achieve an exact combination
> of 1D histos for arbitrary statistics of the individual parallel jobs -
> which precisely needed information about the 'number of attempts' ( =
> number of events + number of times the xsec is zero ).

Fantastic. Glad that we're doing something that should finally be able 
to help... and you can let us know if what we've provided is sufficient.

> Now Radek (in cc)
> wanted to look into the 2.4 series and see how everything is handled
> there. When trying out 2.4.0 (with YODA 1.5.5) for the first time we
> couldn't spot any new information (xsec, number of events) in the
> (attached) yoda file. Is there anything else to be considered except the
> YODA version?

Check the last two entries in the file you attached: the /_EVTCOUNT and 
/_XSEC (Counter and Scatter1D) objects. (The leading underscore is used 
by the plotting scripts and other places in Rivet to avoid doomed 
attempts to plot those objects.)

> And how would we be able to access the cross section and event counter
> information from inside the C++ Rivet source code?

This has always been possible with the crossSection() and 
numEvents()/sumOfWeights() methods... maybe I misunderstood something? 
The new feature is that this information is present in the .yoda file, 
so you can use it in post-processing scripts.

We're avoiding building any special interpretation of such variables 
into YODA (and especially yodamerge), so that it remains a generic data 
analysis package with capabilities useful for these studies, rather than 
a one-trick package specifically for MC analysis, so at a later point we 
plan to have a rivet-merge or similar script which will use the logic of 
yodamerge with added MC-specific use of this information.

Andy

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