def get_series(eventWise, fixed_params, vary_param, tag_index, name_start=""):
all_cols, var_cols, score_cols, table = CompareClusters.tabulate_scores(
eventWise)
name_mask = [
TypeTools.restring(name).startswith(name_start)
for name in table[:, all_cols.index("jet_name")]
]
table = table[name_mask]
mask = CompareClusters.filter_matching(all_cols,
table,
approx=fixed_params)
table = table[mask]
vary_values = table[:, all_cols.index(vary_param)]
if not len(set(vary_values)) == len(vary_values):
import jet_tools
jet_tools.st()
values_dict = {v: i for i, v in enumerate(vary_values)}
sorted_values = CompareClusters.generic_sort(vary_values)
order = [values_dict[v] for v in sorted_values]
jet_names = table[order, all_cols.index("jet_name")]
jet_indices = get_jet_index(eventWise, jet_names, tag_index)
labels = []
for value in sorted_values:
if isinstance(value, float):
labels.append(f"{vary_param}={value:.1}")
else:
labels.append(f"{vary_param}={value}")
return jet_names, jet_indices, labels
def test_tabulate_scores():
params1 = {}
params2 = {}
params1["DogJet_QualityWidth"] = 4
params1["DogJet_DeltaR"] = .5
params2["CatJet_QualityWidth"] = np.inf
with TempTestDir("tst") as dir_name:
# this will raise a value error if given an empty eventWise
file_name1 = "test1.parquet"
file_name2 = "test2.parquet"
ew1 = Components.EventWise(os.path.join(dir_name, file_name1))
ew1.append_hyperparameters(**params1)
ew1.append(DogJet_Label=ak.from_iter([[]]))
path1 = os.path.join(dir_name, file_name1)
ew2 = Components.EventWise(os.path.join(dir_name, file_name2))
ew2.append_hyperparameters(**params2)
ew2.append(CatJet_Label=ak.from_iter([[]]))
path2 = os.path.join(dir_name, file_name2)
all_cols, variable_cols, score_cols, table = CompareClusters.tabulate_scores(
[path1, path2])
assert len(all_cols) == len(table[0])
assert len(all_cols) == len(variable_cols) + len(score_cols) + 3
assert len(table) == 2
dog_row = next(
i for i, name in enumerate(table[:, all_cols.index("jet_name")])
if name == "DogJet")
cat_row = next(
i for i, name in enumerate(table[:, all_cols.index("jet_name")])
if name == "CatJet")
use_cols = [
all_cols.index("QualityWidth"),
all_cols.index("DeltaR"),
all_cols.index("AveBGMassRatio")
]
assert np.all(
TypeTools.soft_equality(table[dog_row, use_cols],
[4, .5, "Undefined"]))
assert np.all(
TypeTools.soft_equality(table[cat_row, use_cols],
[np.inf, "Undefined", "Undefined"]))