nemo_curator.tasks.lance

View as Markdown

Module Contents

Classes

NameDescription
LanceReadTaskTask containing Lance fragment ids assigned to one read partition.

Data

FragmentIds

API

class nemo_curator.tasks.lance.LanceReadTask(
dataset_name: str,
data: nemo_curator.tasks.lance.FragmentIds = list(),
_stage_perf: list[nemo_curator.utils.performance_utils.StagePerfStats] = list(),
_metadata: dict[str, typing.Any] = dict(),
path: str,
version: int
)
Dataclass

Bases: Task[FragmentIds]

Task containing Lance fragment ids assigned to one read partition.

Parameters:

path
str

Path or URI of the Lance dataset to read.

version
int

Lance dataset version to read.

data
FragmentIdsDefaults to list()

Lance fragment ids to read.

data
FragmentIds = field(default_factory=list)
num_items
int
path
str = field(kw_only=True)
version
int = field(kw_only=True)
nemo_curator.tasks.lance.LanceReadTask.get_deterministic_id() -> str
nemo_curator.tasks.lance.LanceReadTask.validate() -> bool
nemo_curator.tasks.lance.FragmentIds: TypeAlias = list[int]