napistu_torch.utils.base_utils

General utilities for the Python standard library.

Classes

CorruptionError

Raised when MPS memory corruption is detected.

Public Functions

ensure_path(path: Union[str, Path], expand_user: bool = True) -> Path

Convert a string or Path to a Path object, optionally expanding user home directory.

normalize_and_validate_indices(indices: Union[int, List[int], tuple, range], max_value: int, param_name: str = “indices”) -> List[int]

Normalize indices to a list and validate they are integers in valid range.

shortest_common_prefix(names: List[str], min_length: int = 3) -> str

Find shortest common prefix respecting word boundaries.

Functions

ensure_path(path[, expand_user])

Convert a string or Path to a Path object, optionally expanding user home directory.

normalize_and_validate_indices(indices, ...)

Normalize indices to a list and validate they are integers in valid range.

shortest_common_prefix(names[, min_length])

Find shortest common prefix respecting word boundaries.

Exceptions

CorruptionError

Raised when MPS memory corruption is detected.
exception napistu_torch.utils.base_utils.CorruptionError

Bases: ValueError

Raised when MPS memory corruption is detected.

napistu_torch.utils.base_utils.ensure_path(path: str | Path, expand_user: bool = True) → Path

Convert a string or Path to a Path object, optionally expanding user home directory.

Parameters:

  • path (Union[str, Path]) – Path to convert. Can be a string (e.g., “~/data/store”) or Path object.

  • expand_user (bool, default=True) – If True, expand tildes (~) to the user’s home directory.

Returns:

Path object, with user expanded if expand_user=True.

Return type:

Path

Raises:

TypeError – If path is not a str or Path object.

Examples

>>> ensure_path("~/data/store")
PosixPath('/home/user/data/store')
>>> ensure_path(Path("./relative/path"))
PosixPath('./relative/path')
>>> ensure_path("~/data", expand_user=False)
PosixPath('~/data')
napistu_torch.utils.base_utils.normalize_and_validate_indices(indices: int | List[int] | tuple | range, max_value: int, param_name: str = 'indices') → List[int]

Normalize indices to a list and validate they are integers in valid range.

Parameters:

  • indices (int, List[int], tuple, or range) – Indices to normalize and validate. Can be a single integer, list, tuple, or range.

  • max_value (int) – Maximum valid index value (exclusive). Valid range is [0, max_value).

  • param_name (str, optional) – Name of the parameter for error messages (default: “indices”)

Returns:

Normalized list of validated indices

Return type:

List[int]

Raises:

ValueError – If indices are invalid (wrong type, not integers, out of range, or duplicates)

napistu_torch.utils.base_utils.shortest_common_prefix(names: List[str], min_length: int = 3) → str

Find shortest common prefix respecting word boundaries.

Parameters:

  • names (List[str]) – Feature names to find common prefix for

  • min_length (int, default=3) – Minimum acceptable prefix length

Returns:

Shortest common prefix, or alphabetically first name if prefix too short

Return type:

str

Examples

>>> shortest_common_prefix(['is_string_x', 'is_string_y'])
'is_string'
>>> shortest_common_prefix(['is_omnipath_kinase', 'is_omnipath_phosphatase'])
'is_omnipath'
>>> shortest_common_prefix(['is_a', 'is_b'])  # Too short
'is_a'