Classes and Functions

In past versions of the python-cmethods package (v1.x) there was a “CMethods” class that implemented the bias correction methods. This class was removed in version v2.0.0. Since then, the cmethods.adjust function is used to apply the implemented techniques except for detrended quantile mapping.

cmethods.adjust(method: str, obs: XRData, simh: XRData, simp: XRData, **kwargs) → XRData

Function to apply a bias correction technique on single and multidimensional data sets. For more information please refer to the method specific requirements and execution examples.

See https://python-cmethods.readthedocs.io/en/latest/methods.html

The time dimension of obs, simh and simp must be named time.

If the sizes of time dimensions of the input data sets differ, you have to pass the hidden input_core_dims parameter, see https://python-cmethods.readthedocs.io/en/latest/getting_started.html#advanced-usage for more information.

Parameters:

  • method (str) – Technique to apply

  • obs (XRData) – The reference/observational data set

  • simh (XRData) – The modeled data of the control period

  • simp (XRData) – The modeled data of the period to adjust

  • kwargs (dict) – Any other method-specific parameter (like n_quantiles and kind)

Returns:

The bias corrected/adjusted data set

Return type:

XRData

distribution.detrended_quantile_mapping(simh: DataArray, simp: DataArray, n_quantiles: int, kind: str = '+', **kwargs: Any) → NPData

See https://python-cmethods.readthedocs.io/en/latest/methods.html#detrended_quantile_mapping

This function can only be applied to 1-dimensional data.

Some additional methods

utils.get_pdf(xbins: list | ndarray) → ndarray

Compuites and returns the the probability density function \(P(x)\) of x based on xbins.

Parameters:

  • x (list | np.ndarray) – The vector to get \(P(x)\) from

  • xbins (list | np.ndarray) – The boundaries/bins of \(P(x)\)

Returns:

The probability densitiy function of x

Return type:

np.ndarray

Compute the probability density function \(P(x)\)
1>>> from cmethods import CMethods as cm
2
3>>> x = [1, 2, 3, 4, 5, 5, 5, 6, 7, 8, 9, 10]
4>>> xbins = [0, 3, 6, 10]
5>>> print(cm.get_pdf(x=x, xbins=xbins))
6[2, 5, 5]
utils.get_cdf(xbins: list | ndarray) → ndarray

Computes and returns returns the cumulative distribution function \(F(x)\) of x based on xbins.

Parameters:

  • x (list | np.ndarray) – Vector to get \(F(x)\) from

  • xbins (list | np.ndarray) – The boundaries/bins of \(F(x)\)

Returns:

The cumulative distribution function of x

Return type:

np.ndarray

Compute the cmmulative distribution function \(F(x)\)
1>>> from cmethods import CMethods as cm
2
3>>> x = [1, 2, 3, 4, 5, 5, 5, 6, 7, 8, 9, 10]
4>>> xbins = [0, 3, 6, 10]
5>>> print(cm.get_cdf(x=x, xbins=xbins))
6[0, 2, 7, 12]
utils.get_inverse_of_cdf(insert_cdf: list | ndarray, xbins: list | ndarray) → ndarray

Returns the inverse cumulative distribution function as: \(F^{-1}_{x}\left[y\right]\) where \(x\) represents base_cdf and insert_cdf is represented by \(y\).

Parameters:

  • base_cdf (list | np.ndarray) – The basis

  • insert_cdf (list | np.ndarray) – The CDF that gets inserted

  • xbins (list | np.ndarray) – Probability boundaries

Returns:

The inverse CDF

Return type:

np.ndarray