[Rivet] rivet 2.1.0a0

[Andy Buckley]

On 05/02/14 18:18, Dmitri Konstantinov wrote:
> Hello Andy,
> 
> We have compiled rivet 2.1.0a0 succesfully on all SLC platforms.(slc5,6
> + gcc)

Hi Dima, thanks!

> But our tests require some adjustments.
> We have a quite simple test allowing to verify if we are still able to
> run rivet and it is run with some same arbitrary hepmc file.
> And now it throws the following message:
> 
> Rivet.Analysis.Handler: WARN  Analysis 'ATLAS_2012_I1083318' is incompatible with the provided beams: removing
> Rivet 2.1.0a0 running on machine lxbsp0516.cern.ch (x86_64)
> All analyses were incompatible with the first event's beams
> Exiting, since this probably wasn't intentional!
> 
> But it is good that there is new flag --ignore-beams  ;)

Actually that has been there for a while! Why not run an analysis which
is valid for the beams in your HepMC sample, or beams suitable for the
analyses?

> Then our validation with rivet is ruined for the moment as yoda.read
> returns str but not yoda object as before.
> we hope to fix it asap.

It now returns a dict, not a list by default. The strings you are seeing
are the keys of the dict. You can either do

histo_list = yoda.read("foo.yoda", asdict=False)
for h in histo_list:
    h.foo
    ...

or

histo_dict = yoda.read("foo.yoda")
for path, h in histo_dict:
    h.foo
    ...
for path in histo_dict.keys():
    histo_dict[path].foo
    ...
for h in histo_dict.values():
    h.foo
    ...

(hope that's clear)

> Therefor I would say that rivet is working.

Excellent! We got a few more minor fixes/requests, and implemented a
couple more analyses, but there shouldn't be anything that affects your
systems beyond what's mentioned above. I'll do a bit of testing on
lxplus and elsewhere to make sure it still builds from the tarballs, and
probably we'll release tomorrow.

Thanks again,
Andy & all

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