Skip to main content

aip_sdk.DatasetVersion

aip_sdk.DatasetVersion(id: str, version: int, dataset_id: str | None = None, stage: str = 'raw', is_golden: bool = False, row_count: int | None = None, column_mapping: dict[str, str] = dict(), check_results: dict[str, Any] | None = None, dataset_metadata: dict[str, Any] = dict(), lineage: dict[str, Any] = dict(), created_at: datetime | None = None, span_table_id: str | None = None)

A single version of a dataset with its lifecycle stage.

Stages (in order): raw — just uploaded; column names may not match the GDI schema mapped — column mapping applied via Dataset.map_version() augmented — optional enrichment step (future) golden — promoted; used as the canonical evaluation set

span_table_id is an opaque token identifying the span table the version reads. Two agent-trace versions are row-comparable only when it matches — a ground-truth-derived version shares its base's token, a re-ingest does not. None outside the agent-trace schema. Compare tokens for equality; the value has no other meaning.

Typical flow:

dataset = project.upload_dataset(df, name="v1")  # stage: raw
v = dataset.latest_version()
dataset.map_version(v.id, {"prompt": "question"})  # stage: mapped
dataset.promote(v.id)  # stage: golden

aip_sdk.DatasetVersion.check_results​

aip_sdk.DatasetVersion.check_results: dict[str, Any] | None = None

No docstring is defined in the source.

aip_sdk.DatasetVersion.column_mapping​

aip_sdk.DatasetVersion.column_mapping: dict[str, str] = field(default_factory=dict)

No docstring is defined in the source.

aip_sdk.DatasetVersion.created_at​

aip_sdk.DatasetVersion.created_at: datetime | None = None

No docstring is defined in the source.

aip_sdk.DatasetVersion.dataset_id​

aip_sdk.DatasetVersion.dataset_id: str | None = None

No docstring is defined in the source.

aip_sdk.DatasetVersion.dataset_metadata​

aip_sdk.DatasetVersion.dataset_metadata: dict[str, Any] = field(default_factory=dict)

No docstring is defined in the source.

aip_sdk.DatasetVersion.id​

aip_sdk.DatasetVersion.id: str

No docstring is defined in the source.

aip_sdk.DatasetVersion.is_golden​

aip_sdk.DatasetVersion.is_golden: bool = False

No docstring is defined in the source.

aip_sdk.DatasetVersion.lineage​

aip_sdk.DatasetVersion.lineage: dict[str, Any] = field(default_factory=dict)

No docstring is defined in the source.

aip_sdk.DatasetVersion.row_count​

aip_sdk.DatasetVersion.row_count: int | None = None

No docstring is defined in the source.

aip_sdk.DatasetVersion.span_table_id​

aip_sdk.DatasetVersion.span_table_id: str | None = None

No docstring is defined in the source.

aip_sdk.DatasetVersion.stage​

aip_sdk.DatasetVersion.stage: str = 'raw'

No docstring is defined in the source.

aip_sdk.DatasetVersion.version​

aip_sdk.DatasetVersion.version: int

No docstring is defined in the source.