pyprobe.analysis.differentiation module#
A module for differentiating experimental data.
- gradient(input_data, x, y)#
Differentiate smooth data with a finite difference method.
A wrapper of the numpy.gradient function. This method calculates the gradient of the data in the y column with respect to the data in the x column.
- Parameters:
- Returns:
A result object containing the columns, x, y and the calculated gradient.
- Return type:
- differentiate_LEAN(input_data, x, y, k=1, gradient='dydx', smoothing_filter=[0.0668, 0.2417, 0.383, 0.2417, 0.0668], section='longest')#
A method for differentiating noisy data.
Uses ‘Level Evaluation ANalysis’ (LEAN) method described in the paper of Feng et al.[1].
This method assumes \(x\) datapoints to be evenly spaced, it can return either \(\frac{dy}{dx}\) or \(\frac{dx}{dy}\) depending on the argument provided to the gradient parameter.
- Parameters:
input_data (RawData | Procedure | Experiment | Cycle | Step | Result) – The input data PyProBE object for the differentiation.
x (str) – The name of the x variable.
y (str) – The name of the y variable.
k (int) – The integer multiple to apply to the sampling interval for the bin size (\(\delta R\) in paper). Default is 1.
gradient (str) – The gradient to calculate, either ‘dydx’ or ‘dxdy’. Default is ‘dydx’.
smoothing_filter (List[float]) –
The coefficients of the smoothing matrix.
Examples provided by Feng et al.[1] include:
[0.25, 0.5, 0.25] for a 3-point smoothing filter.
- [0.0668, 0.2417, 0.3830, 0.2417, 0.0668] (default) for a 5-point
smoothing filter.
- [0.1059, 0.121, 0.1745, 0.1972, 0.1745, 0.121, 0.1059] for a
7-point smoothing filter.
section (str) – The section of the data with constant sample rate in x to be considered. Default is ‘longest’, which just returns the longest unifomly sampled section. Alternative is ‘all’, which returns all sections.
- Returns:
A result object containing the columns, x, y and the calculated gradient.
- Return type:
- pydantic model Differentiation(*, input_data)#
Bases:
BaseModelA class for differentiating experimental data.
- field input_data: RawData | Procedure | Experiment | Cycle | Step | Result [Required]#
The input data for the differentiation.
- differentiate_FD(x, y, gradient='dydx')#
Differentiate smooth data with the finite difference method.
A light wrapper of the numpy.gradient function.
- Parameters:
x (str) – The name of the x variable.
y (str) – The name of the y variable.
gradient (str, optional) – The gradient to calculate, either ‘dydx’ or ‘dxdy’. Defaults to “dydx”.
- Returns:
A result object containing the columns, x, y and the calculated gradient.
- Return type:
- differentiate_LEAN(x, y, k=1, gradient='dydx', smoothing_filter=[0.0668, 0.2417, 0.383, 0.2417, 0.0668], section='longest')#
A method for differentiating noisy data.
Uses ‘Level Evaluation ANalysis’ (LEAN) method described in the paper of Feng et al.[1].
This method assumes \(x\) datapoints to be evenly spaced, it can return either \(\frac{dy}{dx}\) or \(\frac{dx}{dy}\) depending on the argument provided to the gradient parameter.
- Parameters:
x (str) – The name of the x variable.
y (str) – The name of the y variable.
k (int, optional) – The integer multiple to apply to the sampling interval for the bin size (\(\delta R\) in paper). Default is 1.
gradient (str, optional) – The gradient to calculate, either ‘dydx’ or ‘dxdy’. Default is ‘dydx’.
smoothing_filter (List[float], optional) –
The coefficients of the smoothing matrix.
Examples provided by Feng et al.[1] include:
[0.25, 0.5, 0.25] for a 3-point smoothing filter.
[0.0668, 0.2417, 0.3830, 0.2417, 0.0668] (default) for a 5-point smoothing filter.
[0.1059, 0.121, 0.1745, 0.1972, 0.1745, 0.121, 0.1059] for a 7-point smoothing filter.
section (str, optional) – The section of the data with constant sample rate in x to be considered. Default is ‘longest’, which just returns the longest unifomly sampled section. Alternative is ‘all’, which returns all sections.
- Returns:
A result object containing the columns, x, y and the calculated gradient.
- Return type:
References