def baserate_uncertainty(database, label, general_label):
with common.get_db(database) as conn:
cursor = conn.cursor()
cursor.execute(
'''
SELECT
Hall,
DetNo,
RunNo,
acc.Label,
singles.Rate_Hz,
singles.IsolatedEventCount,
singles.IsolatedEventRate_Hz,
acc.BaseRate_Hz
FROM
singles_rates AS singles
INNER JOIN
accidental_subtraction AS acc
USING (
RunNo,
DetNo
)
NATURAL JOIN
runs
WHERE
acc.Label = ?
AND singles.Label = ?
''', (label, general_label))
inputs = np.array(cursor.fetchall())
# Annoying casts because of numpy's homogeneous array types
halls = inputs[:, 0].astype(int)
dets = inputs[:, 1].astype(int)
runs = inputs[:, 2].astype(int)
labels = inputs[:, 3].astype(str)
singles_rate = inputs[:, 4].astype(float)
n_1fold = inputs[:, 5].astype(float)
r_1fold = inputs[:, 6].astype(float)
r_ss = inputs[:, 7].astype(float)
base_rate_error = (r_ss *
(ERROR_COEFFICIENT * r_1fold / singles_rate + 1) /
np.sqrt(n_1fold))
result_rows = np.vstack((base_rate_error, runs, dets, labels)).T
with common.get_db(database) as conn:
cursor = conn.cursor()
cursor.executemany(
'''
UPDATE
accidental_subtraction
SET
BaseRate_Hz_error = ?
WHERE
RunNo = ?
AND DetNo = ?
AND Label = ?
''', result_rows)
def daq_livetime_s(database, label):
"""Return an array of DAQ livetimes ordered from EH1-AD1 to EH3-AD4."""
with common.get_db(database) as conn:
cursor = conn.cursor()
# Total DAQ Livetime
cursor.execute('''
SELECT
SUM(Livetime_ns/Efficiency/1e9)
FROM
muon_rates
NATURAL JOIN
runs
WHERE
Label = ?
GROUP BY
Hall,
DetNo
ORDER BY
Hall,
DetNo
''',
(label,)
)
livetimes_s = np.array(cursor.fetchall()).reshape(-1)
return livetimes_s
def singles_rate_Hz(database, label):
"""Return an array of singles rates from EH1-AD1 to EH3-AD4."""
with common.get_db(database) as conn:
cursor = conn.cursor()
cursor.execute('''
SELECT
SUM(singles.Rate_Hz * Livetime_ns/Efficiency)/
SUM(Livetime_ns/Efficiency)
FROM
singles_rates AS singles
NATURAL JOIN
runs
INNER JOIN
muon_rates
USING (
RunNo,
DetNo,
Label
)
WHERE
Label = ?
GROUP BY
Hall,
DetNo
ORDER BY
Hall,
DetNo
''',
(label,)
)
singles_rates = np.array(cursor.fetchall()).reshape(-1)
return singles_rates
def counts(database, label, general_label):
"""Return an array of predicted accidentals counts."""
with common.get_db(database) as conn:
cursor = conn.cursor()
cursor.execute(
'''
SELECT
SUM(acc.BaseRate_Hz * Total_Acc_Eff * Livetime_ns/1e9)
FROM
accidental_subtraction AS acc
NATURAL JOIN
runs
INNER JOIN
muon_rates as mu
USING (
RunNo,
DetNo
)
WHERE
acc.Label = ?
AND mu.Label = ?
GROUP BY
Hall,
DetNo
ORDER BY
Hall,
DetNo
''', (label, general_label))
counts = np.array(cursor.fetchall()).reshape(-1)
return counts
def target_protons(database, label):
"""Return a 2D array of target protons (x1e25) and uncertainties from EH1-AD1 to EH3-AD4."""
with common.get_db(database) as conn:
cursor = conn.cursor()
cursor.execute('''
SELECT
GdLS_kg,
GdLS_err_kg,
LS_kg,
LS_err_kg,
Acrylic_kg,
Acrylic_err_kg
FROM
target_mass
ORDER BY
Hall,
DetNo
'''
)
masses = np.array(cursor.fetchall())
cursor.execute('''
SELECT
GdLS_density,
GdLS_err,
LS_density,
LS_err,
Acrylic_density,
Acrylic_err
FROM
proton_densities
WHERE
Source = ?
''',
(label,)
)
densities = np.array(cursor.fetchall()).reshape(-1)
num_protons_GdLS = masses[:, 0] * densities[0]
num_protons_LS = masses[:, 2] * densities[2]
num_protons_acrylic = masses[:, 4] * densities[4]
num_protons_total = num_protons_GdLS + num_protons_LS + num_protons_acrylic
err_GdLS = num_protons_GdLS * np.sqrt(
(masses[:, 1]/masses[:, 0])**2 + (densities[1]/densities[0])**2
)
err_LS = num_protons_LS * np.sqrt(
(masses[:, 3]/masses[:, 2])**2 + (densities[3]/densities[2])**2
)
err_acrylic = num_protons_acrylic * np.sqrt(
(masses[:, 5]/masses[:, 4])**2 + (densities[5]/densities[4])**2
)
err_total = np.sqrt(err_GdLS**2 + err_LS**2 + err_acrylic**2)
return np.stack((num_protons_total, err_total), axis=-1)
def count_errors(database, label, general_label):
"""Return the statistical error on the number of accidentals in each AD.
Return value is an array sorted from EH1-AD1 to EH3-AD4.
"""
with common.get_db(database) as conn:
cursor = conn.cursor()
cursor.execute(
'''
SELECT -- easier to compute variance in SQL then take sqrt later
SUM( -- add in quadrature
BaseRate_Hz * Total_Acc_Eff * Livetime_ns/1e9 * -- num acc squared
BaseRate_Hz * Total_Acc_Eff * Livetime_ns/1e9 *
( -- Add relative errors of R_ss and eps_total in quadrature
-- (R_ss = BaseRate_Hz, eps_total = Total_Acc_Eff)
BaseRate_Hz_error * BaseRate_Hz_error / (
BaseRate_Hz * BaseRate_Hz
)
+ Total_Acc_Eff_err * Total_Acc_Eff_err / (
Total_Acc_Eff * Total_Acc_Eff
)
)
)
FROM
accidental_subtraction AS acc
INNER JOIN
muon_rates as mu
USING (
RunNo,
DetNo
)
NATURAL JOIN
runs
WHERE
acc.Label = ?
AND mu.Label = ?
GROUP BY
Hall,
DetNo
ORDER BY
Hall,
DetNo
''', (label, general_label))
result = np.sqrt(np.array(cursor.fetchall()).reshape(-1))
return result
def load_acc_counts_to_db(read_database, label, general_label, bg_database,
bg_label):
"""Compute the counts and errors and load them into the specifified db."""
acc_counts = counts(read_database, label, general_label)
acc_errors = count_errors(read_database, label, general_label)
with common.get_db(bg_database) as conn:
cursor = conn.cursor()
for halldet, count, error in zip(common.all_ads, acc_counts,
acc_errors):
row = (bg_label, halldet[0], halldet[1], 'accidental', count,
error)
cursor.execute(
'''
INSERT INTO
bg_counts
VALUES
(?, ?, ?, ?, ?, ?)
''', row)
return
def muon_total_counts(database, label):
"""Return an array of muon counts from EH1-AD1 to EH3-AD4."""
with common.get_db(database) as conn:
cursor = conn.cursor()
cursor.execute('''
SELECT
SUM(`Count`)
FROM
muon_rates
NATURAL JOIN
runs
WHERE
Label = ?
GROUP BY
Hall,
DetNo
ORDER BY
Hall,
DetNo
''',
(label,)
)
muon_counts = np.array(cursor.fetchall()).reshape(-1)
return muon_counts
def muon_efficiency(database, label):
"""Return an array of muon efficiencies from EH1-AD1 to EH3-AD4."""
with common.get_db(database) as conn:
cursor = conn.cursor()
cursor.execute('''
SELECT
SUM(Livetime_ns)/SUM(Livetime_ns/Efficiency)
FROM
muon_rates
NATURAL JOIN
runs
WHERE
Label = ?
GROUP BY
Hall,
DetNo
ORDER BY
Hall,
DetNo
''',
(label,)
)
muon_effs = np.array(cursor.fetchall()).reshape(-1)
return muon_effs
def coincidences_counts(database, label):
"""Return an array of coincidence counts from EH1-AD1 to EH3-AD4."""
with common.get_db(database) as conn:
cursor = conn.cursor()
cursor.execute('''
SELECT
SUM(NumCoincidences)
FROM
num_coincidences_by_run
NATURAL JOIN
runs
WHERE
Label = ?
GROUP BY
Hall,
DetNo
ORDER BY
Hall,
DetNo
''',
(label,)
)
counts = np.array(cursor.fetchall()).reshape(-1)
return counts
def load_result(database, id_or_description):
"""Load saved results into (FitParams, dict of other attributes)."""
if isinstance(id_or_description, int):
use_id = True
else:
use_id = False
with common.get_db(database) as conn:
conn.row_factory = sqlite3.Row
cursor = conn.cursor()
cursor.execute(
f'''
SELECT
*
FROM
fits
WHERE
{"Id" if use_id else "Description"} = ?
LIMIT
1
''',
(id_or_description, ),
)
fit_info = dict(cursor.fetchone())
fit_info['IsRateOnly'] = bool(fit_info['IsRateOnly'])
fit_info['IsAvgNear'] = bool(fit_info['IsAvgNear'])
fit_info['Hierarchy'] = bool(fit_info['Hierarchy'])
fit_id = fit_info['Id']
fit_theta13 = 0.5 * np.arcsin(np.sqrt(fit_info['FitSinSqT13']))
cursor.execute(
'''
SELECT
ParamName,
ParamIndex,
ParamValue
FROM
pulls
WHERE
FitId = ?
ORDER BY
ParamName,
ParamIndex
''',
(fit_id, ),
)
rows = cursor.fetchall()
param_list = [None] * (2 + pred.FitParams.num_pulls
) # 2 for theta13, dm2
index_map = pred.FitParams.index_map()
param_list[index_map['theta13']] = fit_theta13
param_list[index_map['m2_ee']] = fit_info['FitDM2ee']
for row in rows:
index = index_map[row['ParamName']]
if isinstance(index, int):
param_list[index] = row['ParamValue']
else:
offset = row['ParamIndex']
real_index = index.start + offset
param_list[real_index] = row['ParamValue']
fit_params = pred.FitParams.from_list(param_list)
cursor.execute(
'''
SELECT
Contents
FROM
fit_configs
WHERE
Id = ?
''', (fit_info['FitConfigId'], ))
fit_config_str = cursor.fetchone()[0]
constants = pred.load_constants(fit_config_str, json_str=True)
return fit_params, constants, fit_info
def save_result(
database,
description,
constants,
fit_params,
rate_only,
avg_near,
fit_config_filename,
sin2_error_plus=None,
sin2_error_minus=None,
m2_ee_error_plus=None,
m2_ee_error_minus=None,
):
"""Save the specified results to the database.
The FitConstants object is not saved to the db.
It is only used to compute the chi-squareds.
The fit configuration used to generate the constants
is saved via the fit config file.
"""
chi2_poisson = chi_square(constants,
fit_params,
rate_only=rate_only,
avg_near=avg_near)
chi2_pearson = chi_square(constants,
fit_params,
rate_only=rate_only,
avg_near=avg_near,
variant='pearson')
fits_db_row = (
description,
int(rate_only),
int(avg_near),
fit_params.sin2_2theta13,
fit_params.m2_ee,
chi2_poisson,
chi2_pearson,
sin2_error_plus,
sin2_error_minus,
m2_ee_error_plus,
m2_ee_error_minus,
constants.input_osc_params.theta12,
constants.input_osc_params.m2_21,
int(constants.input_osc_params.hierarchy),
constants.input_osc_params.m2_ee_conversion,
)
params_list = fit_params.to_list()
params_rows = []
for name, index in fit_params.index_map().items():
if name in ('theta13', 'm2_ee'):
# The best-fit values get saved separately
continue
if isinstance(index, int):
# Only 1 value, not an array
param_value = params_list[index]
params_rows.append((name, None, param_value))
else:
# Multiple params in a list, index is a slice
param_values = params_list[index]
for param_index, param_value in enumerate(param_values):
params_rows.append((name, param_index, param_value))
with open(fit_config_filename, 'r') as fit_config_file:
fit_config_contents = fit_config_file.read()
with common.get_db(database) as conn:
cursor = conn.cursor()
cursor.execute(
f'''
INSERT OR IGNORE INTO
fit_configs (Contents)
VALUES
(?)
''', (fit_config_contents, ))
cursor.execute(
f'''
SELECT
Id
FROM
fit_configs
WHERE
Contents = ?
''', (fit_config_contents, ))
fit_config_id, = cursor.fetchone()
cursor.execute(
f'''
INSERT INTO
fits (
Description,
IsRateOnly,
IsAvgNear,
FitSinSqT13,
FitDM2ee,
ChiSquareFit,
ChiSquareGof,
SinSqT13_ErrorPlus,
SinSqT13_ErrorMinus,
DM2ee_ErrorPlus,
DM2ee_ErrorMinus,
Theta12,
DM2_21,
Hierarchy,
DM2ee_conversion,
FitConfigId
)
VALUES
({", ".join("?"*16)}); -- 15 parameters
''',
fits_db_row + (fit_config_id, ),
)
fit_id = cursor.lastrowid
# Update params_rows to include the fit_id
real_params_rows = []
for row in params_rows:
real_params_rows.append((fit_id, *row))
cursor.executemany(
'''
INSERT INTO
pulls
VALUES
(?, ?, ?, ?)
''',
real_params_rows,
)
return
