[Rivet] The rivet Python module could not be loaded

[Andy Buckley]

Hi Hao,

Glad that it works! That's definitely the easiest way.

You will need to rebuild your analysis plugin for the new platform and
Rivet version, but this particular problem is because the Athena Rivet_i
interface changed a little, and the attribute to set is now
rivet.HistoFile rather than rivet.File.

Hope that helps
Andy


On 28/04/14 19:45, Hao Liu wrote:
> Thank you so much Andy, I think I have another quick question here.
> 
> I finally choose to run it on SLC6, since the setup is easy and straight
> forward.
> After I asetup 19.0.1, I tried rivet —help and athena —help. They both
> works perfectly. 
> 
> I was running athena and old version rivet on SLC5 before. So I have an
> executable after compilation from rivet 1.8 on SLC5. 
> Just now I tried to run athena by doing:
> 
> *athena jobconf_rivet.py (this one works fine on SLC5).*
> *
> *
> While it gives me error saying below:
> 
> *Shortened traceback (most recent user call last):
>   File "./jobconf_rivet.py", line 21, in <module>
>     rivet.File = lpath+'/'+file2+'.aida'
> AttributeError: 'Rivet_i' object has no attribute ‘File'*
> *
> *
> I know that the aida would be cause some problem because we don’t use
> aida anymore in Rivet 2.X. While here I don’t think it’s caused by the
> aida thing. 
> I am wondering is this because I am using the file compiled on SLC5
> using an old version Rivet? I attached the jobconf_rivet.py below. The
> azh_lltautau is from a rivet code I wrote for the analysis.
> 
> *theApp.EvtMax = -1
> import AthenaPoolCnvSvc.ReadAthenaPool
> import os
> lpath = os.path.abspath('.')
> input = lpath+'.txt'
> file = open(input)
> file2 = input.split('/')[-1][:-4] 
> svcMgr.EventSelector.InputCollections = []
> for line in file:
> if line[-1] == '\n':
> sline = line[:-1]
> svcMgr.EventSelector.InputCollections.append(sline)
> from AthenaCommon.AlgSequence import AlgSequence
> job = AlgSequence()
> from Rivet_i.Rivet_iConf import Rivet_i
> rivet = Rivet_i()
> 
> rivet.AnalysisPath =
> '/afs/cern.ch/work/h/hao/public/AZh_sys_accept/AZh_lltautau_analysis'
> <http://cern.ch/work/h/hao/public/AZh_sys_accept/AZh_lltautau_analysis'>
> rivet.Analyses += ['azh_lltautau']
> 
> rivet.File = lpath+'/'+file2+'.aida'
> job += rivet
> *
> 
> Best,
> Hao
> On Apr 28, 2014, at 12:34 PM, Andy Buckley <andy.buckley at cern.ch
> <mailto:andy.buckley at cern.ch>> wrote:
> 
>> On 28/04/14 18:10, Hao Liu wrote:
>>> Hi Andy,
>>>
>>> I tried this one this time.
>>>
>>> */afs/cern.ch/sw/lcg/external/MCGenerators_lcgcmt65/rivet/2.1.1/x86_64-slc5-gcc47-opt/rivetenv.sh
>>> <http://cern.ch/sw/lcg/external/MCGenerators_lcgcmt65/rivet/2.1.1/x86_64-slc5-gcc47-opt/rivetenv.sh>
>>> <http://cern.ch/sw/lcg/external/MCGenerators_lcgcmt65/rivet/2.1.1/x86_64-slc5-gcc47-opt/rivetenv.sh>*
>>> *
>>> *
>>> while I got the same error. Is this caused by my environment?
>>> The python I am using right now is */usr/bin/python*
>>
>> The SLC5 (and 6) system Python and GCC aren't a supported LCG platform,
>> so yes it's an environment thing.
>>
>>> what do I need to do before I source this rivetenv.sh file?
>>
>> First of all, do you really need to use SLC5? SLC6 is much more up to
>> date and is what's installed on grid nodes now. If you can use SLC6 then
>> the system Python will be ok, I think, and maybe you can even get away
>> with the system GCC but I always set up a newer CC like this:
>>
>> source
>> /afs/cern.ch/sw/lcg/contrib/gcc/4.6/x86_64-slc6-gcc46-opt/setup.sh
>> <http://cern.ch/sw/lcg/contrib/gcc/4.6/x86_64-slc6-gcc46-opt/setup.sh>
>>
>> (or use an even newer one like 4.7 or 4.8 if that's what the Genser
>> builds require)
>>
>> To use a newer Python, do something like this:
>>
>> PATH=/afs/cern.ch/sw/lcg/external/Python/2.6.5/x86_64-slc5-gcc46-opt/bin:$PATH
>> <http://cern.ch/sw/lcg/external/Python/2.6.5/x86_64-slc5-gcc46-opt/bin:$PATH>
>>
>> where the exact path depends on the Python version you need, whether
>> you're using SLC5 or 6, and the GCC version.
>>
>>> When I was setting up the old version rivet on slc5. I typed in the
>>> command below.
>>>
>>> *export
>>> ATLAS_LOCAL_ROOT_BASE=/cvmfs/atlas.cern.ch/repo/ATLASLocalRootBase
>>> <http://atlas.cern.ch/repo/ATLASLocalRootBase>
>>> <http://atlas.cern.ch/repo/ATLASLocalRootBase>
>>> source
>>> /cvmfs/atlas.cern.ch/repo/ATLASLocalRootBase/user/atlasLocalSetup.sh
>>> <http://atlas.cern.ch/repo/ATLASLocalRootBase/user/atlasLocalSetup.sh>
>>> <http://atlas.cern.ch/repo/ATLASLocalRootBase/user/atlasLocalSetup.sh>
>>> asetup 17.2.3.3,slc5
>>> source
>>> $SITEROOT/sw/lcg/external/MCGenerators_hepmc2.06.05/rivet/1.8.0.a/i686-slc5-gcc43-opt/rivetenv.sh*
>>
>> That's probably not the best idea, since you're then setting up an ATLAS
>> environment (which contains a Rivet version of its own) and then hoping
>> that the rivetenv.sh won't spoil that environment. It might work for
>> Rivet but I don't think running ATLAS commands would be safe anymore!
>>
>> In fact, if you want this for ATLAS purposes then a better thing to do
>> would be to call "asetup 19.0.1" which should automatically provide all
>> the Rivet2 and YODA scripts, plus the ability to run it in Athena.
>>
>>> the reason I want to use a new version of rivet is that I want to use
>>> the efficiency function in YODA which I believe does’t exist in older
>>> version.
>>
>> Great, a real use-case ;-)  Please let us know if you need any help with
>> this, or want to suggest improvements!
>>
>> Andy
>>
>>
>>>> Hi Hao,
>>>>
>>>> That build, as the path indicates, is just for our experimental
>>>> pre-release testing. It was built on SLC6 and hence can only be used
>>>> from SLC6. And anyway it is old and will be removed/replaced shortly:
>>>> the current released version is 2.1.1.
>>>>
>>>> The best way to use a pre-built up-to-date Rivet from AFS is to source
>>>> it from a Genser repository at
>>>> /afs/cern.ch/sw/lcg/external/MCGenerators_lcgcmt*
>>>> <http://cern.ch/sw/lcg/external/MCGenerators_lcgcmt*>
>>>> <http://cern.ch/sw/lcg/external/MCGenerators_lcgcmt*>
>>>>
>>>> In particular, for SLC5 and Rivet 2.1.1 lcgcmt65 looks "good":
>>>>
>>>> /afs/cern.ch/sw/lcg/external/MCGenerators_lcgcmt65/rivet/2.1.1/x86_64-slc5-gcc47-opt/rivetenv.sh
>>>> <http://cern.ch/sw/lcg/external/MCGenerators_lcgcmt65/rivet/2.1.1/x86_64-slc5-gcc47-opt/rivetenv.sh>
>>>> <http://cern.ch/sw/lcg/external/MCGenerators_lcgcmt65/rivet/2.1.1/x86_64-slc5-gcc47-opt/rivetenv.sh>
>>>>
>>>> Andy
>>>>
>>>>
>>>> On 28/04/14 17:42, Hao Liu wrote:
>>>>> Dear Rivet experts,
>>>>>
>>>>> I encountered some problem when I am trying to setup Rivet on
>>>>> lxplus5 at
>>>>> CERN.
>>>>>
>>>>> I found that there is a 2.0 version rivet in our machine which is here:
>>>>>
>>>>> */afs/cern.ch/sw/lcg/experimental/rivet/setup.sh
>>>>> <http://cern.ch/sw/lcg/experimental/rivet/setup.sh>
>>>>> <http://cern.ch/sw/lcg/experimental/rivet/setup.sh>
>>>>> <http://cern.ch/sw/lcg/experimental/rivet/setup.sh>*
>>>>>
>>>>> but after I sourced this setup file. When I typed into rivet —help, I
>>>>> got the message below:
>>>>>
>>>>> [hao at lxplus438 ~]$  rivet --help
>>>>> *The rivet Python module could not be loaded: is your PYTHONPATH set
>>>>> correctly?*
>>>>> *
>>>>> *
>>>>> This version works fine on lxplus6 machine at CERN. I believe this has
>>>>> something to do with the gcc version but I am not so sure.
>>>>>
>>>>> Best,
>>>>> Hao
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> _______________________________________________
>>>>> Rivet mailing list
>>>>> Rivet at projects.hepforge.org <mailto:Rivet at projects.hepforge.org>
>>>>> <mailto:Rivet at projects.hepforge.org>
>>>>> https://www.hepforge.org/lists/listinfo/rivet
>>>>>
>>>>
>>>>
>>>> -- 
>>>> Dr Andy Buckley, Royal Society University Research Fellow
>>>> Particle Physics Expt Group, University of Glasgow / PH Dept, CERN
>>>
>>
>>
>> -- 
>> Dr Andy Buckley, Royal Society University Research Fellow
>> Particle Physics Expt Group, University of Glasgow / PH Dept, CERN
> 


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