[Rivet] Rivet plans and next monthly meeting -- Mon 3 Sep?

[Holger Schulz]

Hi all,

I re-ran the script with some better strucured output
and put the results in a webspace:

    https://users.hepforge.org/~holsch/HepDataRivetConsistency2.6.1/

Holger




On 14/08/18 12:33, Andy Buckley wrote:
> Thanks! And d'oh, i hadn't yet uploaded my own Docker image build!
> I'll put the Dockerfiles in the main repo.
>
> Andy
>
> On 14 Aug 2018 17:23, "Holger Schulz" <iamholger at googlemail.com
> <mailto:iamholger at googlemail.com>> wrote:
>
>
>
>     On 14/08/18 10:26, Andy Buckley wrote:
>>
>>     The main thing to be discussed at that is devising a concrete
>>     plan, with assigned tasks, for resyncing Rivet's ref data
>>     collection with HepData. As a reminder, Holger was going to
>>     provide his "old" Rivet/HD comparison scripts, and Christian &
>>     Louie were going to start building some notes (from the cmp
>>     script output?) on the known origins of discrepancies, e.g. where
>>     Rivet ref data was on-purpose submitted with different plot
>>     content. The triage process is most likely to involve focusing
>>     first on trivial cases where the dataset ID codes are different
>>     but semantically the same; then getting into the (much fewer?)
>>     more complex cases where the datasets actually have different
>>     meanings and maybe we need to port back improvements into HD with
>>     experiment approval.
>     Hi all,
>
>     the script is in the contrib repo of Rivet,
>     hepforge/hg/rivet/public/contrib/devscripts/HepDataConsistency
>
>     All one needs to do is to source the rivet one wants to test and run
>
>         bash mkReports.sh
>
>     I attached the generated reports using the fresh Rivet 2.6.1
>     release. As you know, the differences
>     are numerous. The script finds things like differing number of
>     bins, differing histo keys, whether there
>     are still spires ids being used and so on. Not sure how to proceed
>     here. The smaller problems
>     like spires ids and non-LHC experiments are porbably the easiest
>     ones to tackle but also the least relevant
>     problems.
>
>     The CI person at FNAL, Vito Benedetto, is on a much deserved
>     vacation break
>     at the moment but will be back soon. I am generating and copying
>     hepmc events
>     for the about 30 dedicated nightly tests one after the other (as I
>     find the time really).
>     I identified a small set (maybe not the smallest) of analyses that
>     cover all projections
>     as discussed.
>
>     Also, I took the liberty to build and upload a docker image
>
>         docker.io/hepstore/rivet:2.6.1
>         <http://docker.io/hepstore/rivet:2.6.1>
>
>     Holger
>
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://www.hepforge.org/lists-archive/rivet/attachments/20180815/d28ad106/attachment.html>