voiage.contracts.normalized_input.NormalizedInputBundle

Immutable tables plus the manifest that makes them meaningful.

The data frame representation is Apache Arrow because it has deterministic logical schemas and supports zero-copy handoff to Polars and other engines.

Methods

table

table([positional or keyword] self: None = None, [positional or keyword] table_id: str = None) -> pa.Table

Return the immutable Arrow table selected by its stable identifier.

Parameters:

• self
• table_id str

Returns: pa.Table

write_ipc

write_ipc([positional or keyword] self: None = None, [positional or keyword] path: Path = None) -> None

Write a deterministic single-table IPC file with its manifest metadata.

Parameters:

• self
• path Path

Returns: None

write_parquet

write_parquet([positional or keyword] self: None = None, [positional or keyword] path: Path = None) -> None

Write a deterministic single-table Parquet file with bundle metadata.

Parameters:

• self
• path Path

Returns: None

read_ipc

read_ipc([positional or keyword] cls: None = None, [positional or keyword] path: Path = None) -> NormalizedInputBundle

Restore a normalized bundle previously written by :meth:write_ipc.

Parameters:

• cls
• path Path

Returns: NormalizedInputBundle

read_parquet

read_parquet([positional or keyword] cls: None = None, [positional or keyword] path: Path = None) -> NormalizedInputBundle

Restore a normalized bundle previously written by :meth:write_parquet.

Parameters:

• cls
• path Path

Returns: NormalizedInputBundle