def ROOTGetSize(filename):
root = import_root()
try:
msg.debug('Calling TFile.Open for {0}'.format(filename))
extraparam = '?filetype=raw'
if filename.startswith("https") or filename.startswith("davs"):
try:
pos = filename.find("?")
if pos >= 0:
extraparam = '&filetype=raw'
else:
extraparam = '?filetype=raw'
except Exception:
extraparam = '?filetype=raw'
fname = root.TFile.Open(filename + extraparam, 'READ')
fsize = fname.GetSize()
msg.debug('Got size {0} from TFile.GetSize'.format(fsize))
except ReferenceError:
msg.error('Failed to get size of {0}'.format(filename))
return None
fname.Close()
del root
return fsize
def _get_file_size(filename):
if _urlType(filename) == 'posix':
try:
fsize = os.stat(filename)[6]
except IOError:
fsize = None
else:
from PyUtils.RootUtils import import_root
root = import_root()
try:
msg.debug('Calling TFile.Open for {0}'.format(filename))
pos = filename.find("?")
if pos>=0:
extraparam = '&filetype=raw'
else:
extraparam = '?filetype=raw'
file = root.TFile.Open(filename + extraparam, 'READ')
fsize = file.GetSize()
msg.debug('Got size {0} from TFile.GetSize'.format(fsize))
except ReferenceError:
msg.error('Failed to get size of {0}'.format(filename))
fsize = None
file.Close()
del root
return fsize
def PRWEntries(fileName, integral=False):
root = import_root()
fname = root.TFile.Open(fileName, 'READ')
if not (isinstance(fname, root.TFile) and fname.IsOpen()):
return None
rundir = None
for key in fname.GetListOfKeys():
if key.GetName() == 'PileupReweighting':
rundir = fname.Get('PileupReweighting')
break
# Not PRW...
if rundir is None: return None
total = 0
for key in rundir.GetListOfKeys():
if 'pileup' in key.GetName():
msg.debug('Working on file ' + fileName + ' histo ' +
key.GetName())
if integral:
total += rundir.Get(key.GetName()).Integral()
else:
total += rundir.Get(key.GetName()).GetEntries()
# Was not one of our histograms
return total
def _process_tag_file(self, evtmax=1):
tag_guid = None
nentries = 0
runs = []
evts = []
try:
from PyUtils.RootUtils import import_root
root = import_root()
f = root.TFile.Open(self._filename, 'READ')
metadata = f.Get('CollectionMetadata') if f else None
if metadata:
metadata.GetEntry(0)
# note: we used to use .rstrip('\0') b/c of the change in
# semantics in PyROOT (char[] and const char* may not mean
# the same thing)
# see https://savannah.cern.ch/bugs/?100920 for the gory details
# but in the end, we use ctypes...
# see https://savannah.cern.ch/bugs/?101200 for the gory details
#
# make sure it is what we think it is
import ctypes
key_name = str(ctypes.c_char_p(metadata.Key).value)
assert key_name == 'POOLCollectionID'
tag_guid = str(ctypes.c_char_p(metadata.Value).value)
del metadata
coll_tree = f.Get('POOLCollectionTree') if f else None
if coll_tree:
nentries = coll_tree.GetEntries()
if evtmax in (-1, None):
evtmax = nentries
evtmax = int(evtmax)
for row in range(evtmax):
if coll_tree.GetEntry(row) < 0:
break
runnbr = coll_tree.RunNumber
runs.append(runnbr)
evtnbr = coll_tree.EventNumber
evts.append(evtnbr)
del coll_tree
f.Close()
del f
self._metadata['stream_names'] = ['TAG']
self._metadata['file_guid'] = tag_guid
self._metadata['nentries'] = nentries
self._metadata['run_number'] = runs
self._metadata['evt_number'] = evts
except Exception as e:
print("Exception raised when processing TAG file {0}: {1}".format(
self._filename, e),
file=sys.stderr)
raise
def _process_inp_file(self):
pool_guid = None
nentries = 0
try:
from PyUtils.RootUtils import import_root
root = import_root()
f = root.TFile.Open(self._filename, 'READ')
if f:
# FIXME EventStreamInfo is more authoritative source for nentries
tree = f.Get('POOLContainer')
if not tree: # support for old files
tree = f.Get("POOLContainer_DataHeader")
if tree:
nentries = tree.GetEntriesFast()
del tree
# _get_guid() code from FilePeeker class by Sebastian Binet
pool = f.Get('##Params')
if pool:
pool_token = re.compile(r'\[NAME=(?P<name>.*?)\]'\
r'\[VALUE=(?P<value>.*?)\]').match
params = []
for i in range(pool.GetEntries()):
if pool.GetEntry(i) > 0:
# Work around apparent pyroot issue:
# If we try to access pool.db_string directly,
# we see trailing garbage, which can confuse
# python's bytes->utf8 conversion
# and result in an error.
param = pool.GetLeaf('db_string').GetValueString()
match = pool_token(param)
if not match:
continue
d = match.groupdict()
params.append((d['name'], d['value']))
if d['name'].lower() == 'fid':
pool_guid = d['value']
del pool
f.Close()
del f
self._metadata['file_guid'] = pool_guid
self._metadata['nentries'] = nentries
except Exception as e:
print("Exception raised when processing POOL file {0}: {1}".format(
self._filename, e),
file=sys.stderr)
raise
def NTUPEntries(fileName, treeNames):
if not isinstance(treeNames, list):
treeNames = [treeNames]
root = import_root()
fname = root.TFile.Open(fileName, 'READ')
if not (isinstance(fname, root.TFile) and fname.IsOpen()):
return None
prevNum = None
prevTree = None
for treeName in treeNames:
tree = fname.Get(treeName)
if not isinstance(tree, root.TTree):
return None
num = tree.GetEntriesFast()
if not num >= 0:
msg.warning(
'GetEntriesFast returned non positive value for tree %s in NTUP file %s.',
treeName, fileName)
return None
if prevNum is not None and prevNum != num:
msg.warning(
"Found diffferent number of entries in tree %s and tree %s of file %s.",
treeName, prevTree, fileName)
return None
numberOfEntries = num
prevTree = treeName
del num
del tree
fname.Close()
return numberOfEntries
def _process_inp_file(self):
pool_guid = None
nentries = 0
try:
from PyUtils.RootUtils import import_root
root = import_root()
f = root.TFile.Open(self._filename, 'READ')
if f:
# FIXME EventStreamInfo is more autoritative source for nentries
tree = f.Get('POOLContainer')
if not tree: # support for old files
tree = f.Get("POOLContainer_DataHeader")
if tree:
nentries = tree.GetEntriesFast()
del tree
# _get_guid() code from FilePeeker class by Sebastian Binet
pool = f.Get('##Params')
if pool:
pool_token = re.compile(r'[[]NAME=(?P<name>.*?)[]]'\
r'[[]VALUE=(?P<value>.*?)[]]').match
params = []
for i in xrange(pool.GetEntries()):
if pool.GetEntry(i)>0:
match = pool_token(pool.db_string)
if not match:
continue
d = match.groupdict()
params.append((d['name'], d['value']))
if d['name'].lower() == 'fid':
pool_guid = d['value']
del pool
f.Close()
del f
self._metadata['file_guid'] = pool_guid
self._metadata['nentries'] = nentries
except Exception, e:
print >>sys.stderr, "Exception raised when processing POOL file {0}: {1}".format(self._filename, e)
raise
def HISTEntries(fileName):
root = import_root()
fname = root.TFile.Open(fileName, 'READ')
if not (isinstance(fname, root.TFile) and fname.IsOpen()):
return None
rundir = None
keys = fname.GetListOfKeys()
for key in keys:
name = key.GetName()
if name.startswith('run_') and name != 'run_multiple':
if rundir is not None:
msg.warning(
'Found two run_ directories in HIST file %s: %s and %s',
fileName, rundir, name)
return None
else:
rundir = name
del name
if rundir is None:
msg.warning('Unable to find run directory in HIST file %s', fileName)
fname.Close()
return None
msg.info('Using run directory %s for event counting of HIST file %s. ',
rundir, fileName)
hpath = '%s/GLOBAL/DQTDataFlow/events_lb' % rundir
possibleLBs = []
if 'tmp.HIST_' in fileName:
msg.info('Special case for temporary HIST file {0}. '.format(fileName))
h = fname.Get('{0}'.format(rundir))
for directories in h.GetListOfKeys():
if 'lb' in directories.GetName():
msg.info('Using {0} in tmp HIST file {1}. '.format(
directories.GetName(), fileName))
hpath = rundir + '/' + str(
directories.GetName()) + '/GLOBAL/DQTDataFlow/events_lb'
possibleLBs.append(hpath)
else:
msg.info('Classical case for HIST file {0}. '.format(fileName))
possibleLBs.append(hpath)
nev = 0
if len(possibleLBs) == 0:
msg.warning('Unable to find events_lb histogram in HIST file %s',
fileName)
fname.Close()
return None
for hpath in possibleLBs:
h = fname.Get(hpath)
if not isinstance(h, root.TH1):
msg.warning('Unable to retrieve %s in HIST file %s.', hpath,
fileName)
fname.Close()
return None
nBinsX = h.GetNbinsX()
nevLoc = 0
for i in range(1, nBinsX):
if h[i] < 0:
msg.warning(
'Negative number of events for step %s in HIST file %s.',
h.GetXaxis().GetBinLabel(i), fileName)
fname.Close()
return None
elif h[i] == 0:
continue
if nevLoc == 0:
nevLoc = h[i]
else:
if nevLoc != h[i]:
msg.warning(
'Mismatch in events per step in HIST file %s; most recent step seen is %s.',
fileName,
h.GetXaxis().GetBinLabel(i))
fname.Close()
return None
nev += nevLoc
fname.Close()
return nev