# This code is part of Qiskit.
#
# (C) Copyright IBM 2021.
#
# This code is licensed under the Apache License, Version 2.0. You may
# obtain a copy of this license in the LICENSE.txt file in the root directory
# of this source tree or at http://www.apache.org/licenses/LICENSE-2.0.
#
# Any modifications or derivative works of this code must retain this
# copyright notice, and modified files need to carry a notice indicating
# that they have been altered from the originals.
"""
Curve data classes.
"""
import dataclasses
import itertools
from typing import Any, Dict, Union, List, Tuple, Optional, Iterable
import numpy as np
import uncertainties
from uncertainties.unumpy import uarray
from qiskit_experiments.exceptions import AnalysisError
[docs]
class CurveFitResult:
"""Result of Qiskit Experiment curve analysis."""
def __init__(
self,
method: Optional[str] = None,
model_repr: Optional[Dict[str, str]] = None,
success: Optional[bool] = True,
nfev: Optional[int] = None,
message: Optional[str] = "",
dof: Optional[float] = None,
init_params: Optional[Dict[str, float]] = None,
chisq: Optional[float] = None,
reduced_chisq: Optional[float] = None,
aic: Optional[float] = None,
bic: Optional[float] = None,
params: Optional[Dict[str, float]] = None,
var_names: Optional[List[str]] = None,
x_data: Optional[np.ndarray] = None,
y_data: Optional[np.ndarray] = None,
weighted_residuals: Optional[np.ndarray] = None,
residuals: Optional[np.ndarray] = None,
covar: Optional[np.ndarray] = None,
):
"""Create new Qiskit curve analysis result object.
Args:
method: A name of fitting algorithm used for the curve fitting.
model_repr: String representation of fit functions of each curve.
success: True when the fitting is successfully performed.
nfev: Number of fit function evaluation until the solution is obtained.
message: Any message from the fitting software.
dof: Degree of freedom in this fitting, i.e. number of free parameters.
init_params: Initial parameters provided to the fitter.
chisq: Chi-squared value.
reduced_chisq: Reduced Chi-squared value.
aic: Akaike's information criterion.
bic: Bayesian information criterion.
params: Estimated fitting parameters keyed on the parameter names in the fit function.
var_names: Name of variables, i.e. fixed parameters are excluded from the list.
x_data: X values used for the fitting.
y_data: Y values used for the fitting.
weighted_residuals: The residuals from the fitting after assigning weights for each ydata.
residuals: residuals of the fitted model.
covar: Covariance matrix of fitting variables.
"""
self.method = method
self.model_repr = model_repr
self.success = success
self.nfev = nfev
self.message = message
self.dof = dof
self.init_params = init_params
self.chisq = chisq
self.reduced_chisq = reduced_chisq
self.aic = aic
self.bic = bic
self.params = params
self.var_names = var_names
self.x_data = x_data
self.y_data = y_data
self.weighted_residuals = weighted_residuals
self.residuals = residuals
self.covar = covar
@property
def x_range(self) -> Tuple[float, float]:
"""Range of x_data values."""
return min(self.x_data), max(self.x_data)
@property
def y_range(self) -> Tuple[float, float]:
"""Range of y_data values."""
return min(self.y_data), max(self.y_data)
@property
def ufloat_params(self) -> Dict[str, uncertainties.UFloat]:
"""UFloat representation of fit parameters."""
if hasattr(self, "_ufloat_params"):
# Return cache
return getattr(self, "_ufloat_params")
if self.params is None:
ufloat_params = None
else:
if self.covar is not None:
ufloat_fitvals = uncertainties.correlated_values(
nom_values=[self.params[name] for name in self.var_names],
covariance_mat=self.covar,
tags=self.var_names,
)
else:
# Invalid covariance matrix. Std dev is set to nan, i.e. not computed.
with np.errstate(invalid="ignore"):
# Setting std_devs to NaN will trigger floating point exceptions
# which we can ignore. See https://stackoverflow.com/q/75656026
ufloat_fitvals = uarray(
nominal_values=[self.params[name] for name in self.var_names],
std_devs=np.full(len(self.var_names), np.nan),
)
# Combine fixed params and fitting variables into a single dictionary
# Fixed parameter has zero std_dev
ufloat_params = {}
for name in self.params.keys():
try:
uind = self.var_names.index(name)
ufloat_params[name] = ufloat_fitvals[uind]
except ValueError:
ufloat_params[name] = uncertainties.ufloat(self.params[name], std_dev=0.0)
setattr(self, "_ufloat_params", ufloat_params)
return ufloat_params
@property
def correl(self):
"""Correlation matrix of fit parameters."""
if hasattr(self, "_correl"):
# Return cache
return getattr(self, "_correl")
if self.covar is not None:
# This is how uncertainties computes correlation matrix
stdevs = np.sqrt(np.diag(self.covar))
correl = self.covar / stdevs / stdevs[:, np.newaxis]
else:
correl = None
setattr(self, "_correl", correl)
return correl
def __str__(self):
ret = "CurveFitResult:"
ret += f"\n - fitting method: {self.method}"
ret += f"\n - number of sub-models: {len(self.model_repr)}"
for model_name, model_expr in self.model_repr.items():
if len(model_expr) > 60:
model_expr = f"{model_expr[:60]}..."
ret += f"\n * F_{model_name}(x) = {model_expr}"
ret += f"\n - success: {self.success}"
ret += f"\n - number of function evals: {self.nfev}"
ret += f"\n - degree of freedom: {self.dof}"
ret += f"\n - chi-square: {self.chisq}"
ret += f"\n - reduced chi-square: {self.reduced_chisq}"
ret += f"\n - Akaike info crit.: {self.aic}"
ret += f"\n - Bayesian info crit.: {self.bic}"
if self.init_params is not None:
ret += "\n - init params:"
for name, value in self.init_params.items():
ret += f"\n * {name} = {value}"
if self.ufloat_params is not None:
ret += "\n - fit params:"
for name, param in self.ufloat_params.items():
if np.isfinite(param.std_dev):
ret += f"\n * {name} = {param.nominal_value} ± {param.std_dev}"
else:
ret += f"\n * {name} = {param.nominal_value}"
if self.correl is not None:
ret += "\n - correlations:"
correlated = {}
for pi, pj in itertools.combinations(range(len(self.var_names)), 2):
correlated[(pi, pj)] = self.correl[pi, pj]
for (pi, pj), corr in sorted(correlated.items(), key=lambda item: item[1]):
ret += f"\n * ({self.var_names[pi]}, {self.var_names[pj]}) = {corr}"
return ret
def __copy__(self):
instance = CurveFitResult(**self.__json_encode__())
# Copying ufloat invalidate parameter correlation.
# Note that ufloat object has `self._linear_part.linear_combo` dictionary
# to store parameter correlation keyed on the ufloat objects.
# Copying the ufloat object may change object id, which is the identifier
# of ufloat value, thus it invalidates the `linear_combo` dictionary.
# To avoid missing correlation, the copy invalidate ufloat parameter object cache.
return instance
def __deepcopy__(self, memo):
return self.__copy__()
def __json_encode__(self):
return {
"method": self.method,
"model_repr": self.model_repr,
"success": self.success,
"nfev": self.nfev,
"message": self.message,
"dof": self.dof,
"init_params": self.init_params,
"chisq": self.chisq,
"reduced_chisq": self.reduced_chisq,
"aic": self.aic,
"bic": self.bic,
"params": self.params,
"var_names": self.var_names,
"x_data": self.x_data,
"y_data": self.y_data,
"covar": self.covar,
}
@classmethod
def __json_decode__(cls, value):
return cls(**value)
[docs]
@dataclasses.dataclass
class ParameterRepr:
"""Detailed description of fitting parameter.
Attributes:
name: Original name of the fit parameter being defined in the fit model.
repr: Optional. Human-readable parameter name shown in the analysis result and in the figure.
unit: Optional. Physical unit of this parameter if applicable.
"""
# Fitter argument name
name: str
# Unicode representation
repr: Optional[str] = None
# Unit
unit: Optional[str] = None
class OptionsDict(dict):
"""General extended dictionary for fit options.
This dictionary provides several extra features.
- A value setting method which validates the dict key and value.
- Dictionary keys are limited to those specified in the constructor as ``parameters``.
"""
def __init__(
self,
parameters: List[str],
defaults: Optional[Union[Iterable[Any], Dict[str, Any]]] = None,
):
"""Create new dictionary.
Args:
parameters: List of parameter names used in the fit model.
defaults: Default values.
Raises:
AnalysisError: When defaults is provided as array-like but the number of
element doesn't match with the number of fit parameters.
"""
if defaults is not None:
if not isinstance(defaults, dict):
if len(defaults) != len(parameters):
raise AnalysisError(
f"Default parameter {defaults} is provided with array-like "
"but the number of element doesn't match. "
f"This fit requires {len(parameters)} parameters."
)
defaults = dict(zip(parameters, defaults))
full_options = {p: self.format(defaults.get(p, None)) for p in parameters}
else:
full_options = {p: None for p in parameters}
super().__init__(**full_options)
def __setitem__(self, key, value):
"""Set value with validations.
Raises:
AnalysisError: When key is not previously defined.
"""
if key not in self:
raise AnalysisError(f"Parameter {key} is not defined in this fit model.")
super().__setitem__(key, self.format(value))
def __hash__(self):
return hash(tuple(sorted(self.items())))
def set_if_empty(self, **kwargs):
"""Set value to the dictionary if not assigned.
Args:
kwargs: Key and new value to assign.
"""
for key, value in kwargs.items():
if self.get(key) is None:
self[key] = value
@staticmethod
def format(value: Any) -> Any:
"""Format dictionary value.
Subclasses may override this method to provide their own validation.
Args:
value: New value to assign.
Returns:
Formatted value.
"""
return value
class InitialGuesses(OptionsDict):
"""Dictionary providing a float validation for initial guesses."""
@staticmethod
def format(value: Any) -> Optional[float]:
"""Validate that value is float a float or None.
Args:
value: New value to assign.
Returns:
Formatted value.
Raises:
AnalysisError: When value is not a float or None.
"""
if value is None:
return None
try:
return float(value)
except (TypeError, ValueError) as ex:
raise AnalysisError(f"Input value {value} is not valid initial guess. ") from ex
class Boundaries(OptionsDict):
"""Dictionary providing a validation for boundaries."""
@staticmethod
def format(value: Any) -> Optional[Tuple[float, float]]:
"""Validate if value is a min-max value tuple.
Args:
value: New value to assign.
Returns:
Formatted value.
Raises:
AnalysisError: When value is invalid format.
"""
if value is None:
return None
try:
minv, maxv = value
if minv >= maxv:
raise AnalysisError(
f"The first value is greater than the second value {minv} >= {maxv}."
)
return float(minv), float(maxv)
except (TypeError, ValueError) as ex:
raise AnalysisError(f"Input boundary {value} is not a min-max value tuple.") from ex
# pylint: disable=invalid-name
[docs]
class FitOptions:
"""Collection of fitting options.
This class is initialized with a list of parameter names used in the fit model
and corresponding default values provided by users.
This class is hashable, and generates fitter keyword arguments.
"""
def __init__(
self,
parameters: List[str],
default_p0: Optional[Union[Iterable[float], Dict[str, float]]] = None,
default_bounds: Optional[Union[Iterable[Tuple], Dict[str, Tuple]]] = None,
**extra,
):
# These are private members so that user cannot directly override values
# without implicitly implemented validation logic. No setter will be provided.
self.__p0 = InitialGuesses(parameters, default_p0)
self.__bounds = Boundaries(parameters, default_bounds)
self.__extra = extra
def __hash__(self):
return hash((self.__p0, self.__bounds, tuple(sorted(self.__extra.items()))))
def __eq__(self, other):
if isinstance(other, FitOptions):
checks = [
self.__p0 == other.__p0,
self.__bounds == other.__bounds,
self.__extra == other.__extra,
]
return all(checks)
return False
[docs]
def copy(self):
"""Create copy of this option."""
return FitOptions(
parameters=list(self.__p0.keys()),
default_p0=dict(self.__p0),
default_bounds=dict(self.__bounds),
**self.__extra,
)
@property
def p0(self) -> InitialGuesses:
"""Return initial guess dictionary."""
return self.__p0
@property
def bounds(self) -> Boundaries:
"""Return bounds dictionary."""
return self.__bounds
@property
def fitter_opts(self) -> Boundaries:
"""Return fitter options dictionary."""
return self.__extra
@property
def options(self):
"""Generate keyword arguments of the curve fitter."""
bounds = {k: v if v is not None else (-np.inf, np.inf) for k, v in self.__bounds.items()}
return {"p0": dict(self.__p0), "bounds": bounds, **self.__extra}