import numpy as np
from .Model import Model
from ...lib.readEPF import Parameters
from ...lib.split_channels import split
[docs]
class PolynomialModel(Model):
"""Polynomial Model"""
def __init__(self, **kwargs):
"""Initialize the polynomial model.
Parameters
----------
**kwargs : dict
Additional parameters to pass to
eureka.S5_lightcurve_fitting.models.Model.__init__().
Can pass in the parameters, longparamlist, nchan, and
paramtitles arguments here.
"""
# Inherit from Model class
super().__init__(**kwargs)
# Define model type (physical, systematic, other)
self.modeltype = 'systematic'
# Check for Parameters instance
self.parameters = kwargs.get('parameters')
# Generate parameters from kwargs if necessary
if self.parameters is None:
coeff_dict = kwargs.get('coeff_dict')
params = {cN: coeff for cN, coeff in coeff_dict.items()
if cN.startswith('c') and cN[1:].isdigit()}
self.parameters = Parameters(**params)
# Update coefficients
self._parse_coeffs()
@property
def time(self):
"""A getter for the time."""
return self._time
@time.setter
def time(self, time_array):
"""A setter for the time."""
self._time = time_array
if self.time is not None:
# Convert to local time
if self.multwhite:
self.time_local = []
for chan in self.fitted_channels:
# Split the arrays that have lengths
# of the original time axis
time = split([self.time, ], self.nints, chan)[0]
self.time_local.extend(time - time.mean())
self.time_local = np.array(self.time_local)
else:
self.time_local = self.time - self.time.mean()
def _parse_coeffs(self):
"""Convert dict of 'c#' coefficients into a list
of coefficients in decreasing order, i.e. ['c2','c1','c0'].
Returns
-------
np.ndarray
The sequence of coefficient values
"""
# Parse 'c#' keyword arguments as coefficients
self.coeffs = np.zeros((self.nchannel_fitted, 10))
for c in range(self.nchannel_fitted):
if self.nchannel_fitted > 1:
chan = self.fitted_channels[c]
else:
chan = 0
for i in range(9, -1, -1):
try:
if chan == 0:
self.coeffs[c, 9-i] = \
self.parameters.dict[f'c{i}'][0]
else:
self.coeffs[c, 9-i] = \
self.parameters.dict[f'c{i}_{chan}'][0]
except KeyError:
pass
# Trim zeros
self.coeffs = self.coeffs[:, ~np.all(self.coeffs == 0, axis=0)]
[docs]
def eval(self, channel=None, **kwargs):
"""Evaluate the function with the given values.
Parameters
----------
channel : int; optional
If not None, only consider one of the channels. Defaults to None.
**kwargs : dict
Must pass in the time array here if not already set.
Returns
-------
lcfinal : ndarray
The value of the model at the times self.time.
"""
if channel is None:
nchan = self.nchannel_fitted
channels = self.fitted_channels
else:
nchan = 1
channels = [channel, ]
# Get the time
if self.time is None:
self.time = kwargs.get('time')
# Create the polynomial from the coeffs
lcfinal = np.array([])
for c in range(nchan):
if self.nchannel_fitted > 1:
chan = channels[c]
else:
chan = 0
time = self.time_local
if self.multwhite:
# Split the arrays that have lengths of the original time axis
time = split([time, ], self.nints, chan)[0]
poly = np.poly1d(self.coeffs[chan])
lcpiece = np.polyval(poly, time)
lcfinal = np.append(lcfinal, lcpiece)
return lcfinal