# -*- coding: utf-8 -*-
from our_lib import tables
import synthetic_data
synthetic_data.set_seed(42)
n = 100
mu_key = 'exp_y'
sigma_key = 'err'
table_keys = ('Dpp', 'Dpm', 'Ppp', 'Ppm')
synth_tables = {}
for key in table_keys:
table = tables[key]
synth_tables[key] = synthetic_data.synth_data_from_table(
table, mu_key, sigma_key, n
)
synth_tables[key].to_excel('data/synthetic/{}.xlsx'.format(key),
index=False)
p_indices = ('a1', 'a2', 'a3', 'a4', 'b')
parameters = pd.DataFrame(columns=p_indices)
start_p = (1.0, 1.0, 1.0, 1.0, 1.0)
# Generate and fit the synthetic data.
for i in range(N):
# Generate the synthetic data. (4 tables are generated.)
synth_tables = {}
for k in ('Dpp', 'Dpm', 'Ppp', 'Ppm'):
table = our_lib.tables[k]
synth_tables[k] = (
synthetic_data.synth_data_from_table(table, 'exp_y', 'err')
)
tables_funcs = [(synth_tables[k],
our_lib.theor_funcs[k]) for k in ('Dpp', 'Dpm', 'Ppp', 'Ppm')]
# Fit the synthetic data.
res_of_p = partial(residuals.res_array_from_tables, tables_funcs)
opt_p = our_lib.leastsq(res_of_p, start_p)[0]
# Save the parameters.
parameters = parameters.append(pd.Series(dict(zip(p_indices, opt_p))),
ignore_index=True)
# Watch the time go by.
