# -*- coding: utf-8 -*-
"""
This file contains models of exponential decay fitting routines for qudi based on the lmfit package.
Copyright (c) 2021, the qudi developers. See the AUTHORS.md file at the top-level directory of this
distribution and on <https://github.com/Ulm-IQO/qudi-core/>
This file is part of qudi.
Qudi is free software: you can redistribute it and/or modify it under the terms of
the GNU Lesser General Public License as published by the Free Software Foundation,
either version 3 of the License, or (at your option) any later version.
Qudi is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY;
without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
See the GNU Lesser General Public License for more details.
You should have received a copy of the GNU Lesser General Public License along with qudi.
If not, see <https://www.gnu.org/licenses/>.
"""
__all__ = ('ExponentialDecay', 'multiple_exponential_decay')
import warnings
import numpy as np
from scipy.ndimage import filters
from qudi.util.fit_models.model import FitModelBase, estimator
[docs]
def multiple_exponential_decay(x, amplitudes, decays, stretches):
"""
Mathematical definition of the sum of multiple stretched exponential decays without any bias.
WARNING: Iterable parameters "amplitudes", "decays", and "stretches" must have the same length.
Parameters
----------
x : float
The independent variable to calculate f(x).
amplitudes : iterable
Iterable containing amplitudes for all exponentials.
decays : iterable
Iterable containing decay constants for all exponentials.
stretches : iterable
Iterable containing stretch constants for all exponentials.
Returns
-------
float or numpy.ndarray
The result given x for f(x).
"""
assert len(decays) == len(amplitudes) == len(stretches)
return sum(amp * np.exp(-(x / decay) ** stretch) for amp, decay, stretch in
zip(amplitudes, decays, stretches))
[docs]
class ExponentialDecay(FitModelBase):
"""
"""
[docs]
def __init__(self, **kwargs):
super().__init__(**kwargs)
self.set_param_hint('offset', value=0., min=-np.inf, max=np.inf)
self.set_param_hint('amplitude', value=1., min=-np.inf, max=np.inf)
self.set_param_hint('decay', value=1., min=0., max=np.inf)
self.set_param_hint('stretch', value=1., min=0., max=np.inf)
@staticmethod
def _model_function(x, offset, amplitude, decay, stretch):
return offset + multiple_exponential_decay(x, (amplitude,), (decay,), (stretch,))
@estimator('Decay')
def estimate_decay(self, data, x):
# Smooth very radically the provided data, so that noise fluctuations will not disturb the
# parameter estimation.
if len(data) <= 10:
data_smoothed = data.copy()
else:
sigma = max(1, int(round(len(data) / 10)))
data_smoothed = filters.gaussian_filter1d(data, sigma=sigma)
# Calculate mean value of first and last 10% of data array. Take the latter as offset.
mean_len = max(1, len(x) // 10)
start_mean = np.mean(data_smoothed[mean_len:])
offset = np.mean(data_smoothed[-mean_len:])
# subtraction of the offset and normalization of the decay direction
if start_mean < offset:
data_smoothed = offset - data_smoothed
else:
data_smoothed = data_smoothed - offset
# Make sure there are no negative values
smooth_min = min(data_smoothed)
if smooth_min <= 0:
data_smoothed -= smooth_min
# Take all values up to the standard deviation, the remaining values are
# more disturbing the estimation then helping:
tmp = np.argwhere(data_smoothed <= np.std(data_smoothed))
stop_index = tmp[0, 0] if tmp.size > 0 else len(data_smoothed)
# perform a linear fit on the logarithm of the remaining data
try:
poly_coef = np.polyfit(x[:stop_index], np.log(data_smoothed[:stop_index]), deg=1)
decay = 1 / abs(poly_coef[0])
amplitude = np.exp(poly_coef[1])
except:
warnings.warn('Estimation of decay constant and amplitude failed.')
decay = abs(data[1] - data[0]) / abs(x[-1] - x[0]) / abs(data[-1] - data[0])
amplitude = abs(start_mean - offset)
estimate = self.make_params()
if start_mean < offset:
estimate['amplitude'].set(value=-amplitude, max=0)
else:
estimate['amplitude'].set(value=amplitude, min=0)
estimate['offset'].set(value=offset)
estimate['decay'].set(value=decay, min=2 * min(abs(np.ediff1d(x))))
estimate['stretch'].set(value=1, vary=False)
return estimate
@estimator('Stretched Decay')
def estimate_stretched_decay(self, data, x):
estimate = self.estimate_decay(data, x)
# ToDo: Estimate stretch factor. Currently just a random starting point.
estimate['stretch'].set(value=2, min=0, max=np.inf, vary=True)
return estimate
@estimator('Decay (no offset)')
def estimate_decay_no_offset(self, data, x):
estimate = self.estimate_decay(data, x)
estimate['offset'].set(value=0, min=-np.inf, max=np.inf, vary=False)
return estimate
@estimator('Stretched Decay (no offset)')
def estimate_stretched_decay_no_offset(self, data, x):
estimate = self.estimate_stretched_decay(data, x)
estimate['offset'].set(value=0, min=-np.inf, max=np.inf, vary=False)
return estimate
[docs]
class DoubleExponentialDecay(FitModelBase):
"""
"""
[docs]
def __init__(self, **kwargs):
super().__init__(**kwargs)
self.set_param_hint('offset', value=0., min=-np.inf, max=np.inf)
self.set_param_hint('amplitude_1', value=1., min=-np.inf, max=np.inf)
self.set_param_hint('amplitude_2', value=1., min=-np.inf, max=np.inf)
self.set_param_hint('decay_1', value=1., min=0., max=np.inf)
self.set_param_hint('decay_2', value=1., min=0., max=np.inf)
self.set_param_hint('stretch_1', value=1., min=0., max=np.inf)
self.set_param_hint('stretch_2', value=1., min=0., max=np.inf)
@staticmethod
def _model_function(x, offset, amplitude_1, amplitude_2, decay_1, decay_2, stretch_1,
stretch_2):
return offset + multiple_exponential_decay(x,
(amplitude_1, amplitude_2),
(decay_1, decay_2),
(stretch_1, stretch_2))