Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Reimplement DY #2248

Merged
merged 13 commits into from
Jan 27, 2025
98 changes: 98 additions & 0 deletions nnpdf_data/nnpdf_data/commondata/DYE605_Z0_38P8GEV_DW/filter.py
Original file line number Diff line number Diff line change
@@ -0,0 +1,98 @@
from dataclasses import dataclass
from pathlib import Path

import numpy as np
import pandas as pd
import yaml

from nnpdf_data.filter_utils.hera_utils import commondata
from nnpdf_data.filter_utils.utils import prettify_float

yaml.add_representer(float, prettify_float)


def mergetables() -> pd.DataFrame:

table_paths = []
for i in range(1, 8):
table_paths.append(Path(f"./rawdata/Table{i}.csv"))

# List with the rapidity bins for tables 1 to 7.
yrap = [-0.2, -0.1, 0.0, 0.1, 0.2, 0.3, 0.4]

col_names = ["M2", "dsig", "statp", "statm", "normp", "normm", "sysp", "sysm"]
col_names_all = col_names + ["y", "sqrts"]

combined_df = pd.DataFrame(columns=col_names_all)
for i, path in enumerate(table_paths):
df = pd.read_csv(path, header=11, names=col_names)
df["y"] = yrap[i]
df["sqrts"] = 38.8
df = df[pd.to_numeric(df['dsig'], errors='coerce').notnull()]
combined_df = pd.concat([combined_df, df], ignore_index=True)

# In the table we have sqrt(tau) not M2; compute M2=tau*s
combined_df["M2"] = (combined_df["M2"] * 38.8) ** 2

return combined_df


def nuclear_uncert_dw(tableN: Path, tablep: Path):
dfN = pd.read_table(tableN)
dfp = pd.read_table(tablep)
return dfN, dfp


@dataclass
class E605_commondata(commondata):
def __init__(self, data: pd.DataFrame, dataset_name: str, process: str):

# Kinematic quantities.
self.central_values = data["dsig"].astype(float).to_numpy()
self.kinematics = data[["y", "M2", "sqrts"]].astype(float).to_numpy()
self.kinematic_quantities = ["y", "M2", "sqrts"]

# Statistical uncertainties.
self.statistical_uncertainties = data["statp"]

# the overall 10% statistical uncertainty is treated as
# additive, while normalisation uncertainty is always treated
# multiplicatively
syst = pd.DataFrame(0.1 * self.central_values)

# Systematic uncertainties.
syst["norm"] = self.central_values * data["normp"].str.strip("%").astype(float) / 100

# self.systematic_uncertainties = np.dstack((stat,norm))[0]
self.systypes = [("ADD", "UNCORR"), ("MULT", "CORR")]

# Compute the point-to-point uncertainties
nrep = 999
norm = np.sqrt(nrep)
dfN, dfp = nuclear_uncert_dw(
Path("rawdata/nuclear/output/tables/group_result_table.csv"),
Path("rawdata/proton_ite/output/tables/group_result_table.csv"),
)

for rep in range(1, nrep + 1):
Delta = (dfN[f"rep_{rep:05d}"] - dfp["theory_central"]) / norm
syst[f"NUCLEAR{rep:05d}"] = Delta
self.systypes.append(("ADD", f"NUCLEAR{rep:05d}"))

self.systematic_uncertainties = syst.to_numpy()

self.process = process
self.dataset_name = dataset_name


def main():
data = mergetables()
# First create the commondata variant without the nuclear uncertainties.
DYE605 = E605_commondata(data, "DYE605_Z0_38P8GEV", "Z0")
DYE605.write_new_commondata(
Path("data.yaml"), Path("kinematics.yaml"), Path("uncertainties.yaml")
)


if __name__ == "__main__":
main()
Loading
Loading