yohou.utils¶

Validation, panel data, weighting, tags, discovery, and other utility functions.

Name	Description
`all_estimators`	Get a list of all estimators from `yohou`.
`all_displays`	Get a list of all displays from `yohou`.
`all_functions`	Get a list of all functions from `yohou`.

Name	Description
`inspect_panel`	Inspect DataFrame columns to distinguish global and local (panel) data.
`get_group_df`	Extract and rename columns for a specific panel group.
`dict_to_panel`	Convert a dict of group DataFrames to a single DataFrame with prefixed columns.
`select_panel_columns`	Select panel group columns and optionally global columns of a DataFrame.
`panel_aware_rename`	Apply a rename function to a column name while preserving the panel group prefix.
`panel_aware_prefix`	Add a prefix to a column name while preserving the panel group prefix.
`panel_aware_suffix`	Add a suffix to a column name while preserving the panel group prefix.
`check_groups`	Validate and normalize panel group names for forecaster operations.
`check_groups_exist`	Validate all requested panel groups exist in fitted forecaster.
`check_panel_groups_match`	Validate that y and X have matching panel group structures.
`check_panel_internal_consistency`	Validate that all panel groups in a DataFrame have the same local column structure.

Name	Description
`validate_forecaster_data`	Validate data for forecasters.
`validate_transformer_data`	Validate data for transformers.
`validate_scorer_data`	Validate and prepare scorer input data.
`validate_splitter_data`	Validate data for splitters.
`validate_plotting_data`	Validate a DataFrame for plotting and resolve columns.
`validate_plotting_params`	Validate common plotting function parameters.
`validate_search_data`	Validate input data for hyperparameter search (GridSearchCV, RandomizedSearchCV).
`validate_time_weight`	Validate time_weight parameter for forecasters and scorers.
`validate_column_names`	Validate that __ separator is used only for panel data group names.

Name	Description
`check_time_column`	Validate that time column exists, has proper dtype, no nulls, and is sorted.
`check_interval_consistency`	Validate that a time series has uniform time spacing.
`check_continuity`	Validate temporal continuity between consecutive DataFrames.
`check_sufficient_rows`	Validate DataFrame has sufficient rows for operation.
`check_inputs`	Validate that target and feature DataFrames have consistent time intervals.
`check_schema`	Validate DataFrame schema and return with proper column ordering.
`check_X_actual_required`	Validate X_actual is provided when required for recursive prediction.
`check_forecasting_horizon_positive`	Validate forecasting horizon is positive.
`check_scorer_column_selection`	Subselect columns based on scorer configuration.

Name	Description
`exponential_decay_weight`	Generate exponential decay weights giving more weight to recent times.
`linear_decay_weight`	Generate linear decay weights giving more weight to recent times.
`seasonal_emphasis_weight`	Generate weights emphasizing specific seasonal positions.
`compose_weights`	Compose multiple weight functions by multiplication.
`validate_callable_signature`	Validate that callable has valid signature for time weighting.
`normalize_weights`	Normalize weights so they sum to the number of elements.
`validate_weight_array`	Validate a resolved weight array for NaN, negatives, infinities, and all-zero.
`resolve_dict_weights`	Map a `{key: weight}` dict to an aligned numpy array.
`combine_weight_vectors`	Combine weight vectors multiplicatively and normalize.
`resolve_weight_to_array`	Resolve a weight specification (callable, DataFrame, or dict) to a numpy array.

Name	Description
`add_interval`	Add n intervals to a datetime (handles variable-length intervals).
`interval_to_timedelta`	Convert fixed interval to timedelta, or None for variable intervals.
`parse_interval`	Parse interval string into (multiplier, unit).

Name	Description
`cast`	Cast columns according to schema with integer rounding.
`get_numeric_columns`	Get list of numeric column names from a DataFrame.
`tabularize`	Convert time series to tabular format using lags.