[Rivet] HepData and Rivet growing apart

[Christian Gutschow]

Hi,

I’ve mentioned this before, but one motivation for us to put a reference data file into Rivet, that is different from what HEPData would give us, is the lack of flexibility when it comes to the identifiers. The d01-x01-y01 type naming (of which x01 is actually just a dummy identifier that cannot be changed, as I had to learn the hard way…) is too restrictive when you want to put things like ratios into Rivet and be able to juggle them alongside numerators and denominators in Rivet. An additional field in the HEPData yaml input file that would allow us to specify a custom Rivet identifier (instead of the d01-x01-y01 bit) that is used when HEPData converts the data into the yoda format would be ace.

Further, new analyses don't seem to make it into hepdata even after half a
year or so (e.g. ATLAS_2017_I1509919).

I think you’ll find that the data from this analysis were put on HEPData around the same time as the Rivet routine was submitted, i.e. January 2017:

https://hepdata.net/record/ins1509919

If you know of any actual missing HEPData entries for ATLAS measurements that have a corresponding Rivet routine, please do let me know and I’m happy to follow up.

Cheers,
Chris

On 16 Aug 2017, at 10:59, Andy Buckley <andy.buckley at cern.ch<mailto:andy.buckley at cern.ch>> wrote:

Hi Holger,

Thanks for highlighting this again.

It's a disaster for Rivet if the data submitted to us in Rivet analyses does not align with that in HepData. HD should be our primary data resource, from which we can resync our data: I knew there was some divergence due to a few "hacked" ref data files, but this sounds like there has been a much greater break of synchronisation since the HD user-submission system was introduced.

Graeme, what do you suggest that we do? I hope we can also help you to identify and import analysis data that exists only in Rivet: this should not have happened, and please feed back to the experiments that the data repository is HD, not Rivet. I would be more than happy to co-sign.

Andy

PS. Holger, this reminds me: was the Rivet analyses json file for HepData ever created for the Rivet 2.5.4 release? Could you pull the scripts to generate that thing into the Rivet doc directory on the default branch, so people other than you can run it in future? Thanks.


On 16/08/17 09:20, Holger Schulz wrote:
Hi riveteers and Graeme,
I have noticed that the data stored in hepdata and their equivalent in rivet
are diverging quite a bit. This mail is just a report I do not have any suggestions
but would like to remind everyone that this problem exists.
Not only are the histokeys quite different in the sample I checked but also the
binning and the numerical precision of central values at times can differ quite
a bit. Further, new analyses don't seem to make it into hepdata even after half a
year or so (e.g. ATLAS_2017_I1509919).
I was toying with the in principle very helpful yodastat download link that gives
you the yoda files with statistical errors only and is hence desirable for
more in depth statistical analysis. But due to the inconsistencies it cannot be
used without great pain.
If you want to have a look, I checked this entry:
   http://hepdata.cedar.ac.uk/view/irn8994773/yoda
against
   https://rivet.hepforge.org/hg/rivet/file/6375cc481dd0/data/refdata/ATLAS_2011_S8994773.yoda
Any thoughts and discussions are welcome.
Thanks,
Holger
_______________________________________________
Rivet mailing list
Rivet at projects.hepforge.org<mailto: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
_______________________________________________
Rivet mailing list
Rivet at projects.hepforge.org<mailto:Rivet at projects.hepforge.org>
https://www.hepforge.org/lists/listinfo/rivet




 —

 Dr. Christian Gütschow

 Department of Physics and Astronomy
 University College London
 Gower Street
 London WC1E 6BT

 > D10 Physics Building
 > +44 (0)20 7679 3775
 > chris.g at cern.ch<mailto:chris.g at cern.ch>




-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://www.hepforge.org/lists-archive/rivet/attachments/20170816/eeac7304/attachment.html>