def test_readme():
momentum = from_root('TMath::Sqrt(X_PX**2 + X_PY**2 + X_PZ**2)')
assert momentum.to_numexpr(
) == 'sqrt(((X_PX ** 2) + (X_PY ** 2) + (X_PZ ** 2)))'
assert momentum.to_root(
) == 'TMath::Sqrt(((X_PX ** 2) + (X_PY ** 2) + (X_PZ ** 2)))'
my_selection = from_numexpr('X_PT > 5 & (Mu_NHits > 3 | Mu_PT > 10)')
assert my_selection.to_root(
) == '(X_PT > 5) && ((Mu_NHits > 3) || (Mu_PT > 10))'
assert my_selection.to_numexpr(
) == '(X_PT > 5) & ((Mu_NHits > 3) | (Mu_PT > 10))'
my_sum = from_auto('True + False')
assert my_sum.to_root() == 'true + false'
assert my_sum.to_numexpr() == 'True + False'
my_check = from_auto('(X_THETA*TMath::DegToRad() > pi/4) && D_PE > 9.2')
assert my_check.variables == {'D_PE', 'X_THETA'}
assert my_check.named_constants == {'DEG2RAD', 'PI'}
assert my_check.unnamed_constants == {'4', '9.2'}
new_selection = (momentum > 100) and (my_check or (np.sqrt(my_sum) < 1))
def numexpr_eval(string):
return numexpr.evaluate(string,
local_dict=dict(X_THETA=1234, D_PE=678))
assert pytest.approx(
numexpr_eval(new_selection.to_numexpr()),
numexpr_eval(
'((X_THETA * 0.017453292519943295) > (3.141592653589793 / 4)) & (D_PE > 9.2)'
))
def extract_branch_names(string):
try:
string = string.decode()
except (UnicodeDecodeError, AttributeError):
pass
return formulate.from_auto(string).variables
def selection_branches(selection: str) -> set[str]:
"""Construct the minimal set of branches required for a selection.
Parameters
-----------
selection : str
Selection string in ROOT or numexpr
Returns
-------
set(str)
Necessary branches/variables
Examples
--------
>>> from tdub.data import minimal_selection_branches
>>> selection = "(reg1j1b == True) & (OS == True) & (mass_lep1lep2 > 100)"
>>> minimal_branches(selection)
{'OS', 'mass_lep1lep2', 'reg1j1b'}
>>> selection = "reg2j1b == true && OS == true && (mass_lep1jet1 < 155)"
>>> minimal_branches(selection)
{'OS', 'mass_lep1jet1', 'reg2j1b'}
"""
return formulate.from_auto(selection).variables
def selection_as_root(selection: str) -> str:
"""Get the ROOT selection string from an arbitrary selection.
Parameters
-----------
selection : str
The selection string in ROOT or numexpr
Returns
-------
str
The same selection in ROOT format.
Examples
--------
>>> selection = "(reg1j1b == True) & (OS == True) & (mass_lep1jet1 < 155)"
>>> from tdub.data import selection_as_root
>>> selection_as_root(selection)
'(reg1j1b == true) && (OS == true) && (mass_lep1jet1 < 155)'
"""
return formulate.from_auto(selection).to_root()
def selection_as_numexpr(selection: str) -> str:
"""Get the numexpr selection string from an arbitrary selection.
Parameters
-----------
selection : str
Selection string in ROOT or numexpr
Returns
-------
str
Selection in numexpr format.
Examples
--------
>>> selection = "reg1j1b == true && OS == true && mass_lep1jet1 < 155"
>>> from tdub.data import selection_as_numexpr
>>> selection_as_numexpr(selection)
'(reg1j1b == True) & (OS == True) & (mass_lep1jet1 < 155)'
"""
return formulate.from_auto(selection).to_numexpr()
def do_cut(
tree_name,
files,
supercuts,
proposedBranches,
output_directory,
eventWeightBranch,
pids,
):
position = -1
if pids is not None:
# handle pid registration
if os.getpid() not in pids:
pids[np.argmax(pids == 0)] = os.getpid()
# this gives us the position of this particular process in our list of processes
position = np.where(pids == os.getpid())[0][0]
start = clock()
try:
branches = []
aliases = {}
missingBranches = False
for fname in files:
with uproot.open(fname) as f:
tree = f[tree_name]
for branch in proposedBranches:
if branch in tree:
branches.append(branch)
else:
if branch in tree.aliases:
aliases[branch.decode()] = formulate.from_auto(
tree.aliases[branch].decode())
branches.extend(
extract_branch_names(tree.aliases[branch]))
else:
logger.error(
'branch {} not found in {} for {}'.format(
branch, tree_name, fname))
missingBranches |= True
if missingBranches:
sys.exit(1)
for alias, alias_expr in aliases.items():
alias_expr = expand_definition(alias_expr, aliases)
branches.extend(extract_branch_names(alias_expr.to_numexpr()))
aliases[alias] = alias_expr
branches = set(branches)
eventWeightBranch = expand_selection(eventWeightBranch, aliases)
supercuts = expand_supercuts(supercuts, aliases)
# iterate over the cuts available
cuts = defaultdict(lambda: {'raw': 0, 'weighted': 0})
events_tqdm = tqdm(
total=uproot.numentries(files, tree_name),
disable=(position == -1),
position=2 * position + 1,
leave=False,
mininterval=5,
maxinterval=10,
unit="events",
dynamic_ncols=True,
)
for file, start, stop, events in uproot.iterate(
files,
tree_name,
branches=branches,
namedecode='utf-8',
reportfile=True,
reportentries=True,
):
events_tqdm.set_description("({1:d}) Working on {0:s}".format(
tree_name.decode('utf-8'), 2 * position + 1))
for cut in tqdm(
get_cut(copy.deepcopy(supercuts)),
desc="({1:d}) Applying cuts to {0:s}".format(
file.name.decode('utf-8'), 2 * position + 2),
total=get_n_cuts(supercuts),
disable=(position == -1),
position=2 * position + 2,
leave=False,
unit="cuts",
miniters=10,
dynamic_ncols=True,
):
cut_hash = get_cut_hash(cut)
rawEvents, weightedEvents = apply_cuts(events, cut,
eventWeightBranch)
cuts[cut_hash]['raw'] += rawEvents
cuts[cut_hash]['weighted'] += weightedEvents
events_tqdm.update(stop - start)
with open(
"{0:s}/{1:s}.json".format(output_directory,
tree_name.decode('utf-8')),
"w+") as f:
f.write(json.dumps(cuts, sort_keys=True, indent=4))
result = True
except:
logger.exception("Caught an error - skipping {0:s}".format(
tree_name.decode('utf-8')))
result = False
end = clock()
return (result, end - start)
def expand_selection(selection, aliases):
return expand_definition(formulate.from_auto(selection),
aliases).to_numexpr()
