Jet clustering CMS data without CMSSW

cluster_cms.py

from pyjet import cluster
from root_numpy import root2array, stretch

branches = [
    'recoPFCandidates_particleFlow__RECO.obj.pt_',
    'recoPFCandidates_particleFlow__RECO.obj.eta_',
    'recoPFCandidates_particleFlow__RECO.obj.phi_',
    'recoPFCandidates_particleFlow__RECO.obj.mass_',
    ]

# http://opendata.cern.ch/research/CMS
filename = ("root://eospublic.cern.ch//eos/opendata/cms/Run2011A/DoubleMu/AOD/"
            "12Oct2013-v1/10000/000D143E-9535-E311-B88B-002618943934.root")

# Read the first event
events = root2array(filename, "Events", branches=branches, stop=1)

for event in events:
    flattened_event = stretch(event.reshape(-1)) # Stretch out the nestedness of the subarrays 
    sequence = cluster(flattened_event, R=0.5, p=-1)  # Anti-kt with R=0.5
    jets = sequence.inclusive_jets(ptmin=3)  # You get the same jets as CMS! Just need to calibrate them...