preprocessing.extra

logger_config

logger

validate_extra_features

def validate_extra_features(
        config: dict,
        sessions: Optional[List[str]] = None,
        extra_features: Optional[List[str]] = None) -> None

Validate that every session's processed .nc satisfies the extra-feature contract.

For each name in the extra-feature list and each session, checks:

• The processed .nc file exists.
• The data variable is present.
• Its dims are (time,) or (time, individuals).
• Its time length matches the session's position_processed time length.
• Its dtype is numeric (float or int).
• The first-individual slice contains no NaN or Inf values.

All problems across all sessions are collected and reported in a single ValueError so users can fix them in one pass.

Parameters

• config (dict): Configuration dictionary.
• sessions (list of str, optional): Sessions to validate. Defaults to config["session_names"].
• extra_features (list of str, optional): Override the list of feature names to validate. Defaults to config["extra_features"]. Inference paths pass the list recovered from training metadata so they validate against what the model expects, not what the live config currently contains.

Returns

• None: Returns silently on success.

Raises

• ValueError: If any registered extra feature fails the contract for any session.