API Reference

Parameter Configuration

Default Configuration Parameters

The following parameters define PyRth’s default behavior and can be overridden by passing values in the parameter dictionary to evaluation methods.

Standard Evaluation Parameters

Standard Evaluation Parameters

precision (default: 250)

int: Number of points used for internal calculations involving the impedance curve.

log_time_size (default: 250)

int: Number of points in the logarithmically spaced time array used for spectrum calculations.

filter_name (default: “hann”)

str: Name of the filter for FFT deconvolution. Options: “fermi”, “gauss”, “nuttall”, “blackman_nuttall”, “hann”, “blackman_harris”, “rectangular”.

filter_range (default: 0.6)

float: Range parameter for the FFT deconvolution filter (if applicable).

filter_parameter (default: 0.0)

float: Additional parameter for the FFT deconvolution filter (if applicable).

deconv_mode (default: “bayesian”)

str: Deconvolution method. Options: ‘bayesian’, ‘fft’, ‘lasso’.

bay_steps (default: 1000)

int: Number of steps for Bayesian deconvolution.

pad_factor_pre (default: 0.01)

float: Padding factor to prepend zeros before deconvolution.

pad_factor_after (default: 0.01)

float: Padding factor to append zeros after deconvolution.

lasso_alpha (default: 0.0001)

array-like or float: Regularization parameter(s) for Lasso deconvolution.

lasso_max_iter (default: 10000)

int: Maximum number of iterations for Lasso deconvolution.

lasso_tol (default: 0.0001)

float: Tolerance for convergence in Lasso deconvolution.

lasso_cv_folds (default: 1)

int: Number of cross-validation folds for Lasso deconvolution.

lasso_selection (default: “cyclic”)

str: Selection method for Lasso deconvolution. Options: “cyclic”, “random”.

lasso_precompute (default: True)

bool: Whether to precompute the Gram matrix for Lasso deconvolution.

struc_method (default: “sobhy”)

str: Method for structure function calculation. Options: “sobhy”, “lanczos”, “boor_golub”, “khatwani”, “polylong”.

timespec_interpolate_factor (default: 1.0)

float: Interpolation factor for the time constant spectrum (used by Lanczos).

blockwise_sum_width (default: 20)

int: Number of RC rungs to combine for smoothing (used by Lanczos).

theo_inverse_specs (default: None)

dict | None: Dictionary of theoretical inverse specifications.

theo_resistances (default: None)

list | None: List of resistances for the theoretical Foster network model.

theo_capacitances (default: None)

list | None: List of capacitances for the theoretical Foster network model.

theo_time (default: [4e-08, 1000.0])

list[float, float]: Time range [start, end] in seconds for theoretical model calculation.

theo_time_size (default: 30000)

int: Number of time points for the theoretical model calculation.

signal_to_noise_ratio (default: 100.0)

float: Signal-to-noise ratio for adding noise to theoretical impedance data.

theo_delta (default: 0.008726646259971648)

float: Angle (radians) to rotate Z(s) into the complex plane for theoretical calculations to avoid singularities.

calib (default: None)

np.ndarray | None: 2D array of calibration data [temps, voltages]. Required if input_mode is ‘voltage’.

kfac_fit_deg (default: 2)

int: Degree of the polynomial fit for K-factor calculation from calibration data.

extrapolate (default: True)

bool: Whether to extrapolate the thermal response using a sqrt(time) fit at early times.

lower_fit_limit (default: None)

float | None: Lower time limit (seconds) for the extrapolation fit range.

upper_fit_limit (default: None)

float | None: Upper time limit (seconds) for the extrapolation fit range.

data_cut_lower (default: 0)

int: Index to cut data; points below this index are excluded from the transient analysis.

data_cut_upper (default: float(‘inf’))

int | float: Index to cut data; points above this index are excluded from the transient analysis.

temp_0_avg_range (default: (0, 1))

tuple[int, int]: Index range (start, end) to average the initial temperature/voltage to determine the baseline.

power_step (default: 1.0)

float: Power step [W] applied during the measurement. Used for impedance calculation.

power_scale_factor (default: 1.0)

float: Scaling factor applied to power, useful for analyzing multiple DUTs in series to get per-component properties.

optical_power (default: 0.0)

float: Optical power [W] to subtract, relevant for LED testing.

is_heating (default: False)

bool: True if the transient corresponds to a heating step (positive power), False for cooling.

power_data (default: None)

np.ndarray | None: Excitation power curve for temperature prediction. 2D array: [time, power].

lin_sampling_period (default: 1e-06)

float: Sampling period [s] for linear interpolation of the impulse response in temperature prediction.

minimum_window_length (default: 0.35)

float: Minimum window length (in log10(time) units) for the adaptive derivative calculation.

maximum_window_length (default: 3.0)

float: Maximum window length (in log10(time) units) for the adaptive derivative calculation.

minimum_window_size (default: 70)

int: Minimum number of data points within the derivative calculation window.

window_increment (default: 0.1)

float: Increment (+/-) applied to the window length during the adaptive derivative calculation update step.

expected_var (default: 0.09)

float: Expected variance of the noise in the thermal transient data, used in derivative calculation.

min_index (default: 3)

int: Minimum index from which to start the derivative calculation.

opt_recalc_forward (default: False)

bool: Whether to recalculate the forward solution during optimization.

opt_use_extrapolate (default: True)

bool: Whether to use the extrapolated impedance curve during optimization.

opt_method (default: “Powell”)

str: Optimization method to use (passed to scipy.optimize.minimize).

struc_init_method (default: “optimal_fit”)

str: Method to determine the initial structure function approximation for optimization.

opt_model_layers (default: 10)

int: Number of RC layers (Foster elements) for the optimization model.

input_mode (default: “impedance”)

str: Input data type. Options: ‘impedance’, ‘temperature’, ‘voltage’.

calc_struc (default: True)

bool: Whether to calculate the structure function after impedance calculation.

only_make_z (default: False)

bool: If True, only calculate the impedance curve and skip spectrum/structure function steps.

repetitions (default: 1000)

int: Number of repetitions for bootstrapping analysis.

random_seed (default: None)

int | None: Random seed for bootstrapping to ensure reproducibility.

bootstrap_mode (default: “from_data”)

str: Method for generating bootstrap samples. Options: “from_theo”, “from_data”, “given”, “given_with_opt”.

normalize_impedance_to_previous (default: False)

bool: In batch processing, normalize subsequent impedance curves to the first one.

evaluation_type (default: “standard”)

str: Type of evaluation module to run within standard_module_set.

iterable_keywords (default: [])

list[str]: Keyword argument names that should be iterated over in standard_module_set.

data (default: None)

np.ndarray | None: Input data. 2D array: [time, measurement].

output_dir (default: “output”)

str: Base directory for saving output files.

label (default: “no_label”)

str: Label used for naming output files and figures.

infile (default: None)

str | None: Input file path for T3ster data files.

infile_pwr (default: None)

str | None: Input file path for T3ster power files.

infile_tco (default: None)

str | None: Input file path for T3ster calibration files.

total_calls (default: 1)

int: Counter, possibly related to the number of analysis calls.

fig_total_calls (default: 1)

int: Counter, possibly related to the number of figures generated.

Evaluation Methods

class PyRth.Evaluation

User-facing class for running thermal transient evaluations.

The class owns the shared parameter dictionary, tracks generated modules, and offers convenience methods (standard_module, bootstrap_module …) that apply the appropriate defaults before delegating to StructureFunction. It also centralizes exporter wiring so each module can materialize CSV/figure outputs consistently.

Core Methods

Evaluation.standard_module(parameters)

Standard module for evaluation of the impedance function. This method compiles the necessary parameters and initializes the structure function based on validation and requirements.

Evaluation.standard_module_set(parameters)

Standard module set for evaluation of the impedance function. This method is used to evaluate multiple impedance approximations using different parameters.

Evaluation.bootstrap_module(parameters: Dict)

Bootstrap module for evaluation with specified parameters. The module is used to evaluate the impedance approximation using a structure function. The structure function is calculated from a given set of resistances and capacitances.

Evaluation.optimization_module(parameters: dict)

Optimizes the impedance approximation using a structure function. The structure function is calculated from a given set of resistances and capacitances.

Evaluation.theoretical_module(parameters: dict)

Calculates a thermal impedance from a structure function. The structure function is calculated from a given set of resistances and capacitances. Each resistance and capacitance is associated with constant RC-transmission line section, which is concatenated to the previous.

Parameters:

• theo_resistances (list of size N) – List of total resistances for each constant sections for the piecewise-uniform structure function.

• theo_capacitances (list of size N) – List of total capacitances for each constant sections for the piecewise-uniform structure function.

• output_dir (string, default="outputcsv") – Output directory for plots saved as PNG-files. Additionally a subfolder named “csv” is created where the CSV files are saved.

• label (string, default="no_label") – Label of the given parameter set. The label is used for the output files.

• theo_time (list of size 2, default=[-4e-8, 1e3]) – The time range for the thermal impedance in seconds.

• theo_time_size (integer, default=30000) – Number of points evenly distributed over the logarithmic time domain.

• theo_delta (float, default=0.5) – Angle to rotate the integration path into the complex plane.

• theo_added_noise (float) – The added gaussian noise to the impedance.

Examples

The following example calculates a thermal impedance for structure function with the five sections. The thermal impedance, impulse response, time constant spectrum, as well as other results are saved as CSV-files in the folder “theoretical”.

>>> import PyRth
>>> parameters = {
...     "output_dir": "theoretical/",
...     "label": "theoretical",
...     "theo_time": [3e-7, 200],
...     "theo_time_size": 10000,
...     "theo_delta": 1.5 * (2 * np.pi / 360),
...     "theo_resistances": [10, 10, 10, 10, 10],
...     "theo_capacitances": [1e-4, 1e-1, 1e-4, 1e-3, 1e0],
... }
>>> eval_instance = PyRth.Evaluation()
>>> eval_instance.theoretical_module(parameters)
>>> eval_instance.save_as_csv()
>>> eval_instance.save_figures()

Evaluation.comparison_module(parameters: dict)

Compares the impedance of a system to a theoretical impedance. The theoretical impedance is calculated from a given set of resistances and capacitances. Each resistance and capacitance is associated with constant RC-transmission line section, which is concatenated to the previous.

Evaluation.temperature_prediction_module(parameters)

Predicts the temperature of a system given a power function and an impulse response.

Utility Methods

Evaluation.save_as_csv()

Export all evaluation results to CSV files.

For each module stored in this Evaluation instance, saves all relevant data arrays (impedance, structure functions, time spectra, etc.) to separate CSV files in the configured output directories.

Evaluation.save_figures()

Generate and save figures for all evaluation results.

Creates visualizations of the data for each module stored in this Evaluation instance. Generates plots for impedance curves, structure functions, time constant spectra, etc., based on the output configuration.

Evaluation.save_all()

Save both CSV data files and figures for all evaluation results.

Convenience method that calls both save_as_csv() and save_figures().