def summarize(mgr, df_cfg, event_selection, scribblers, max_events=-1):
'''
Summarise the data in the tree into the data frames (DFs) given in
df_cfg.
:param mgr: HEPPY job manager
:param df_cfg(list): list of DF definitions
:param event_selection: pairs of event selections and collectors
:param scribblers: pairs of scribblers and empty collectors which
create new event content
:param max_events(int): Number of events to process.
Default is -1 -> all events
'''
reader_collector_pairs = scribblers + event_selection
# setting up defaults to complete the provided DF configs
tableConfigCompleter = TableConfigCompleter(
# using a composer to create a predictable output file name
# based on the names of the output columns
createOutFileName=TableFileNameComposer2(default_prefix='tbl_n'))
# combine configs and completers
reader_collector_pairs += complete(df_cfg, tableConfigCompleter)
# Hard-coded list of components is temporary, need to remove this in the
# near future
return mgr.run(reader_collector_pairs,
components=[
"TTWJetsToLNu_amcatnloFXFX",
"ZJetsToNuNu_HT100to200_madgraph_ext1"
])
def summarize(mgr, df_cfg, event_selection, scribblers, max_events=-1):
'''
Summarise the data in the tree into the data frames (DFs) given in
df_cfg.
:param mgr: HEPPY job manager
:param df_cfg(list): list of DF definitions
:param event_selection: pairs of event selections and collectors
:param scribblers: pairs of scribblers and empty collectors which
create new event content
:param max_events(int): Number of events to process.
Default is -1 -> all events
'''
reader_collector_pairs = scribblers + event_selection
# setting up defaults to complete the provided DF configs
tableConfigCompleter = TableConfigCompleter(
# using a composer to create a predictable output file name
# based on the names of the output columns
createOutFileName=TableFileNameComposer2(default_prefix='tbl_n'))
# combine configs and completers
reader_collector_pairs += complete(df_cfg, tableConfigCompleter)
# Unlike step6, now run over all components
return mgr.run(reader_collector_pairs)
def _create_rc_pairs(self):
keyAttrNames, keyOutColumnNames, binnings, keyIndices = zip(
*self._binning)
if not any(keyIndices):
keyIndices = None
base_cfg = dict(keyAttrNames=keyAttrNames,
keyOutColumnNames=keyOutColumnNames,
binnings=binnings,
keyIndices=keyIndices)
df_configs = {}
if not self._weights:
name_composer = TableFileNameComposer()
df_configs["unweighted"] = base_cfg
else:
for name, weights in self._weights.items():
config = copy.copy(base_cfg)
if weights != 1:
config["weight"] = weights
df_configs[name] = config
name_composer = WithInsertTableFileNameComposer(
TableFileNameComposer(), list(df_configs.keys()))
tableConfigCompleter = TableConfigCompleter(
createOutFileName=name_composer, defaultOutDir=self.output_dir)
return complete(df_configs.values(), tableConfigCompleter)
def summarize(tree, df_cfg, event_selection, scribblers, max_events = -1):
'''
Summarise the data in the tree into the data frames (DFs) given in
df_cfg.
:param tree(ROOT.TTree): the input tree
:param df_cfg(list): list of DF definitions
:param event_selection: pairs of event selections and collectors
:param scribblers: pairs of scribblers and empty collectors which
create new event content
:param max_events(int): Number of events to process.
Default is -1 -> all events
'''
reader_collector_pairs = scribblers + event_selection
# setting up defaults to complete the provided DF configs
tableConfigCompleter = TableConfigCompleter(
# using a composer to create a predictable output file name
# based on the names of the output columns
createOutFileName=TableFileNameComposer2(default_prefix='tbl_n')
)
# combine configs and completers
reader_collector_pairs += complete(df_cfg, tableConfigCompleter)
# wrap tree for the event loop
def event_builder():
return alphatwirl.roottree.BEvents(tree, maxEvents=max_events)
# create reader and collector collections
reader = alphatwirl.loop.ReaderComposite()
collector = alphatwirl.loop.CollectorComposite()
for r, c in reader_collector_pairs:
reader.add(r)
collector.add(c)
# loop over all events
eventLoop = alphatwirl.loop.EventLoop(event_builder, reader)
reader = eventLoop()
# collect all results and return them
return collector.collect(((None, (reader, )), ))
def summarize(out_dir, mgr, df_cfg, event_selection, scribblers,
components_df):
'''
Summarise the data in the tree into the data frames (DFs) given in
df_cfg.
:param mgr: NANOAOD job manager
:param df_cfg(list): list of DF definitions
:param event_selection: pairs of event selections and collectors
:param scribblers: pairs of scribblers and empty collectors which
create new event content
:param components_df: list of nanoaod component names to summarize
'''
reader_collector_pairs = [(s, NullCollector())
for s in scribblers] + event_selection
reader_collector_pairs += [(SaveTree(out_dir), NullCollector())]
name_composer = WithInsertTableFileNameComposer(TableFileNameComposer(),
df_cfg.keys())
# setting up defaults to complete the provided DF configs
tableConfigCompleter = TableConfigCompleter(
# using a composer to create a predictable output file name
# based on the names of the output columns
createOutFileName=name_composer,
defaultOutDir=out_dir)
# combine configs and completers
reader_collector_pairs += complete(df_cfg.values(), tableConfigCompleter)
print "Components:"
for d, e in zip(components_df["dataset"], components_df["era"]):
print "\t{}_{}".format(d, e)
print
return mgr.run(reader_collector_pairs, components=components_df)