Support Utilities

Configuration models (utils.config_models)

Configuration validation and normalization helpers.

utils.config_models.normalize_split_column(series)[source]

Normalize split labels to train/valid/test while preserving pandas semantics.

Return type:

Any

utils.config_models.validate_yaml_mapping(data, *, source=None)[source]

Validate that parsed YAML data is a non-empty mapping.

Parameters:

  • data (Any)

  • source (Path | None)

Return type:

dict[str, Any]

utils.config_models.normalize_loader_config(cfg)[source]

Validate and normalize a loader config without mutating the caller’s dict.

Parameters:

cfg (Any)

Return type:

dict[str, Any]

utils.config_models.normalize_runtime_config(cfg)[source]

Validate a runtime config passed into run_one_config/builders.

Parameters:

cfg (Any)

Return type:

dict[str, Any]

utils.config_models.normalize_echo_config(cfg)[source]

Normalize the config payload stored in analysis summaries.

Parameters:

cfg (Any)

Return type:

dict[str, Any]

Benchmark cleaning (utils.benchmark_cleaning)

Opt-in cleaning utilities for curated benchmark splits.

utils.benchmark_cleaning.clean_benchmark_splits(splits, task_type, *, reference_splits=DEFAULT_REFERENCE_SPLITS, remove_invalid=True, remove_conflicts=True, remove_contaminants=True)[source]

Return cleaned benchmark splits and a JSON-serializable cleaning report.

Cleaning is intentionally opt-in and operates only on in-memory split frames. Removal precedence is invalid rows, label-conflicting molecules, then exact contaminants in non-reference splits.

Parameters:

  • splits (Mapping[str, pandas.DataFrame])

  • task_type (str)

  • reference_splits (Sequence[str] | str)

  • remove_invalid (bool)

  • remove_conflicts (bool)

  • remove_contaminants (bool)

Return type:

Tuple[Dict[str, pandas.DataFrame], Dict[str, Any]]

Pydantic compatibility layer (utils.pydantic_compat)

Small compatibility layer for pydantic v1 and v2 APIs.

utils.pydantic_compat.pydantic_model_validate(model_cls, payload)[source]

Compatibility wrapper for pydantic v1/v2 model validation.

Parameters:

  • model_cls (Type[T])

  • payload (Any)

Return type:

T

utils.pydantic_compat.pydantic_model_dump(instance, **kwargs)[source]

Compatibility wrapper for pydantic v1/v2 model dumps.

Parameters:

  • instance (Any)

  • kwargs (Any)

Return type:

dict[str, Any]