bluemath_tk.predictor package

Submodules

bluemath_tk.predictor.xwt module

class bluemath_tk.predictor.xwt.XWT(steps: Dict[str, BlueMathModel])[source]

Bases: BlueMathModel, BlueMathPipeline

Xly Weather Types (XWT) class.

This class implements the XWT method to identify and classify weather patterns in a dataset. The XWT method is a combination of Principal Component Analysis (PCA) and K-means clustering (KMA).

steps

The steps of the XWT method.

Type:

Dict[str, BlueMathModel]

num_clusters

The number of clusters.

Type:

int

kma_bmus

The KMA best matching units (BMUs).

Type:

pd.DataFrame

property clusters_annual_probs_df
property clusters_monthly_probs_df
property clusters_perpetual_year_probs_df
property clusters_probs_df
property clusters_seasonal_probs_df
property data: Dataset
fit(data: Dataset, fit_params: Dict[str, Dict[str, Any]] = {}, variable_to_sort_bmus: str = None) → None[source]

Fit the XWT model to the data.

Parameters:

  • data (xr.Dataset) – The data to fit the model to. Must be PCA formatted.

  • fit_params (Dict[str, Dict[str, Any]], optional) – The fitting parameters for the PCA and KMA models. Default is {}.

  • variable_to_sort_bmus (str, optional) – The variable to sort the BMUs. Default is None.

Raises:

  • XWTError – If the data is not PCA formatted.

  • TODO – Standarize PCs by first PC variance.: pca.pcs_df / pca.pcs.stds.isel(n_component=0).values ??

property get_conditioned_probabilities
plot_dwts_probs(vmax: float = 0.15, vmax_seasonality: float = 0.15, plot_text: bool = False) → None[source]

Plot Daily Weather Types bmus probabilities.

Parameters:

  • vmax (float, optional) – The maximum value of the colorbar. Default is 0.15.

  • vmax_seasonality (float, optional) – The maximum value of the colorbar for seasonality. Default is 0.15.

  • plot_text (bool, optional) – Whether to plot the text in each cell. Default is False.

Raises:

ValueError – If the kma_bmus time sampling is not daily.

plot_map_features(ax: Axes, land_color: str = array([0.9375, 0.9375, 0.859375])) → None[source]

Plot map features on an axis.

Parameters:

  • ax (Axes) – The axis to plot the map features on.

  • land_color (str, optional) – The color of the land. Default is cfeature.COLORS[“land”].

plot_perpetual_year() → Axes[source]

Plot perpetual year bmus probabilities.

Returns:

The plot with the perpetual year bmus probabilities.

Return type:

Axes

plot_xwts(var_to_plot: str, anomaly: bool = False, map_center: tuple = None) → Collection[source]

Plot the XWTs for a variable.

Parameters:

  • var_to_plot (str) – The variable to plot.

  • anomaly (bool, optional) – Whether to plot the anomaly of the variable. Default is False.

  • map_center (tuple, optional) – The center of the map. Default is None.

Returns:

The grid specification with the XWTs plot.

Return type:

GridSpec

exception bluemath_tk.predictor.xwt.XWTError(message='XWT error occurred.')[source]

Bases: Exception

Custom exception for XWT class.

bluemath_tk.predictor.xwt.check_model_is_fitted(func)[source]

Decorator to check if the model is fitted.

bluemath_tk.predictor.xwt.get_dynamic_estela_predictor(data: Dataset, estela: Dataset, check_interpolation: bool = True, verbose: bool = False) → Dataset[source]

Transform an xarray dataset of longitude, latitude, and time into one where each longitude, latitude value at each time is replaced by the corresponding time - t, where t is specified in the estela dataset.

Parameters ———-ltimes = estela.where(estela.F >= 0, np.nan).traveltime.astype(int) estela_max_traveltime = estela_traveltimes.max().values for traveltime in range(estela_max_traveltime):

data = data.w

dataxr.Dataset

The input dataset with dimensions longitude, latitude, and time.

estelaxr.Dataset

The dataset containing the F values with dimensions longitude and latitude.

check_interpolationbool, optional

Whether to check if the data is interpolated. Default is True.

verbosebool, optional

Whether to print verbose output. Default is False. If False, Dask logs are suppressed. If True, Dask logs are shown.

Returns:

The transformed dataset.

Return type:

xr.Dataset

Module contents

Project: BlueMath_tk Sub-Module: predictor Author: GeoOcean Research Group, Universidad de Cantabria Repository: https://github.com/GeoOcean/BlueMath_tk.git Status: Under development (Working)

class bluemath_tk.predictor.XWT(steps: Dict[str, BlueMathModel])[source]

Bases: BlueMathModel, BlueMathPipeline

Xly Weather Types (XWT) class.

This class implements the XWT method to identify and classify weather patterns in a dataset. The XWT method is a combination of Principal Component Analysis (PCA) and K-means clustering (KMA).

steps

The steps of the XWT method.

Type:

Dict[str, BlueMathModel]

num_clusters

The number of clusters.

Type:

int

kma_bmus

The KMA best matching units (BMUs).

Type:

pd.DataFrame

property clusters_annual_probs_df
property clusters_monthly_probs_df
property clusters_perpetual_year_probs_df
property clusters_probs_df
property clusters_seasonal_probs_df
property data: Dataset
fit(data: Dataset, fit_params: Dict[str, Dict[str, Any]] = {}, variable_to_sort_bmus: str = None) → None[source]

Fit the XWT model to the data.

Parameters:

  • data (xr.Dataset) – The data to fit the model to. Must be PCA formatted.

  • fit_params (Dict[str, Dict[str, Any]], optional) – The fitting parameters for the PCA and KMA models. Default is {}.

  • variable_to_sort_bmus (str, optional) – The variable to sort the BMUs. Default is None.

Raises:

  • XWTError – If the data is not PCA formatted.

  • TODO – Standarize PCs by first PC variance.: pca.pcs_df / pca.pcs.stds.isel(n_component=0).values ??

property get_conditioned_probabilities
plot_dwts_probs(vmax: float = 0.15, vmax_seasonality: float = 0.15, plot_text: bool = False) → None[source]

Plot Daily Weather Types bmus probabilities.

Parameters:

  • vmax (float, optional) – The maximum value of the colorbar. Default is 0.15.

  • vmax_seasonality (float, optional) – The maximum value of the colorbar for seasonality. Default is 0.15.

  • plot_text (bool, optional) – Whether to plot the text in each cell. Default is False.

Raises:

ValueError – If the kma_bmus time sampling is not daily.

plot_map_features(ax: Axes, land_color: str = array([0.9375, 0.9375, 0.859375])) → None[source]

Plot map features on an axis.

Parameters:

  • ax (Axes) – The axis to plot the map features on.

  • land_color (str, optional) – The color of the land. Default is cfeature.COLORS[“land”].

plot_perpetual_year() → Axes[source]

Plot perpetual year bmus probabilities.

Returns:

The plot with the perpetual year bmus probabilities.

Return type:

Axes

plot_xwts(var_to_plot: str, anomaly: bool = False, map_center: tuple = None) → Collection[source]

Plot the XWTs for a variable.

Parameters:

  • var_to_plot (str) – The variable to plot.

  • anomaly (bool, optional) – Whether to plot the anomaly of the variable. Default is False.

  • map_center (tuple, optional) – The center of the map. Default is None.

Returns:

The grid specification with the XWTs plot.

Return type:

GridSpec