LowRank.CurveDecomposer

LowRank.CurveDecomposer#

LowRank.CurveDecomposer.py

This module contains the decompose functions used to decompose a given I-curve into a set of component curves.

safe_log10(x)#

Compute the logarithm base 10 of x, returning a large negative number for non-positive x.

Parameters:

x (float) – The input value.

Returns:

The logarithm base 10 of x, or -10 if x is non-positive.

Return type:

float

compute_areas(x, peak_list)#

Compute the areas under the peaks defined by peak_list over the x values.

Parameters:

  • x (array-like) – The x values.

  • peak_list (list of tuples) – The list of peak parameters, where each tuple contains (height, tR, sigma, tau).

Returns:

areas – The areas under each peak.

Return type:

array-like

class CurveDecomposer#

Bases: object

A class for decomposing curves into component curves.

decompose_icurve_impl(icurve, num_components, **kwargs)#

Decompose a curve into component curves. :param icurve: The input curve to be decomposed. :type icurve: Curve :param num_components: The number of components to decompose into. :type num_components: int :param curve_model: The model to use for curve decomposition.

Currently, only ‘EGH’ (Exponentially-Gaussian Hybrid) is supported. Default is ‘EGH’.

Parameters:

  • smoothing (bool, optional) – Whether to apply smoothing to the input curve before decomposition. Default is False.

  • debug (bool, optional) – If True, print debug information and plot intermediate results. Default is False.

  • kwargs (dict, optional) –

    Additional keyword arguments for decomposition. Possible keys include:

    • decompargs: dict or None

      If provided, use these arguments for decomposition instead of peak recognition.

    • tau_limit: float

      Limit for the tau parameter in the EGH model. Default is 0.5.

    • area_weight: float

      Weight for the area proportion penalty in the fitting objective. Default is 0.1.

    • proportions: list of float or None

      Target proportions for the areas of the components. If None, use areas from initial peak recognition.

    • sec_constraints: bool

      If True, apply SEC-related constraints during fitting. Default is False.

    • data_matrix: array-like or None

      The data matrix for SEC constraints. Required if sec_constraints is True.

    • qv: array-like or None

      The q-values corresponding to the data matrix. Required if sec_constraints is True.

    • num_plates: int or None

      The number of plates to use for decomposition. If None, use the default value.

    • randomize: float

      Standard deviation of Gaussian noise to add to initial parameters for randomization. Default is 0 (no randomization).

    • seed: int or None

      Random seed for parameter randomization. Default is None.

    • global_opt: bool

      If True, use global optimization (basinhopping) for fitting. Default is False.

Returns:

ret_curves – The list of decomposed component curves.

Return type:

list of ComponentCurve

Contents