Input Validation

This section documents validation utilities provided by eb-features to enforce structural assumptions on panel time-series data.

These helpers ensure column presence, monotonic timestamps, and feature integrity.

All content below is generated automatically from NumPy-style docstrings in the source code.

Validation API

eb_features.panel.validation

Validation utilities for panel feature engineering.

This module centralizes lightweight input validation helpers used throughout the panel feature engineering subpackage.

Key invariants

• Required columns must exist.
• Within each entity, timestamps must be strictly increasing in the given row order. (No sorting is performed inside validation; callers may sort afterward for deterministic computation, but validation should catch out-of-order input.)

validate_required_columns(df, *, required_cols)

Validate that required columns exist on a DataFrame.

Parameters:

Name	Type	Description	Default
df	DataFrame	Input DataFrame.	required
required_cols	Sequence[str]	Columns that must be present.	required

Raises:

Type	Description
KeyError	If any required columns are missing.

ensure_columns_present(df, *, columns, label)

Ensure that a set of configured columns exists on a DataFrame.

Parameters:

Name	Type	Description	Default
df	DataFrame	Input DataFrame.	required
columns	Iterable[str]	Columns that must be present.	required
label	str	Human-readable label used in error messages (e.g., "Static", "Regressor").	required

Raises:

Type	Description
KeyError	If any specified columns are missing.

validate_monotonic_timestamps(df, *, entity_col, timestamp_col)

Validate that timestamps are strictly increasing within each entity in row order.

This function does not sort. It checks the DataFrame exactly as provided.

Parameters:

Name	Type	Description	Default
df	DataFrame	Input panel DataFrame.	required
entity_col	str	Entity identifier column.	required
timestamp_col	str	Timestamp column.	required

Raises:

Type	Description
KeyError	If required columns are missing.
ValueError	If any entity has non-strictly increasing timestamps (ties or out-of-order), or if timestamps cannot be parsed.