Implementation:Cohere ai Cohere python DatasetPart Model

Knowledge Sources	Cohere Python SDK
Domains	SDK, Datasets
Last Updated	2026-02-15 14:00 GMT

Overview

DatasetPart is a Pydantic model representing a single part (file segment) within a Cohere dataset, including metadata such as size, row count, and download URLs.

Description

The DatasetPart class models an individual part of a dataset stored in the Cohere platform. Datasets may be split into multiple parts for storage and retrieval purposes. Each part carries its own identifier, name, download URL, index position, byte size, row count, original URL, and a sample of parsed rows. All fields except id and name are optional, reflecting that metadata may not always be available depending on the dataset state.

The class extends UncheckedBaseModel and is auto-generated by the Fern API definition toolchain. It supports both Pydantic v1 and v2 through a compatibility layer, and allows extra fields to be passed through without validation errors.

Usage

Use DatasetPart when working with the Cohere Datasets API to inspect or iterate over the individual file parts that compose a dataset. This is useful when downloading dataset files, checking dataset upload progress, or examining row-level samples from each part.

Code Reference

Source Location

• Repository: Cohere Python SDK
• File: src/cohere/types/dataset_part.py

Signature

class DatasetPart(UncheckedBaseModel):
    id: str
    name: str
    url: typing.Optional[str] = None
    index: typing.Optional[int] = None
    size_bytes: typing.Optional[int] = None
    num_rows: typing.Optional[int] = None
    original_url: typing.Optional[str] = None
    samples: typing.Optional[typing.List[str]] = None

Import

from cohere.types import DatasetPart

I/O Contract

Fields

Field	Type	Required	Default	Description
id	str	Yes	--	The dataset part ID
name	str	Yes	--	The name of the dataset part
url	Optional[str]	No	None	The download URL of the file
index	Optional[int]	No	None	The index of the file within the dataset
size_bytes	Optional[int]	No	None	The size of the file in bytes
num_rows	Optional[int]	No	None	The number of rows in the file
original_url	Optional[str]	No	None	The download URL of the original file
samples	Optional[List[str]]	No	None	The first few rows of the parsed file

Usage Examples

Constructing a DatasetPart Instance

from cohere.types import DatasetPart

# Create a DatasetPart with all fields populated
part = DatasetPart(
    id="part-abc123",
    name="training_data_part_0",
    url="https://storage.cohere.ai/datasets/part-abc123/download",
    index=0,
    size_bytes=1048576,
    num_rows=5000,
    original_url="https://storage.cohere.ai/datasets/part-abc123/original",
    samples=["sample row 1", "sample row 2", "sample row 3"],
)

print(part.id)          # "part-abc123"
print(part.name)        # "training_data_part_0"
print(part.size_bytes)  # 1048576
print(part.num_rows)    # 5000

Accessing Parts from a Dataset Response

import cohere

co = cohere.Client()

dataset = co.datasets.get(id="my-dataset-id")
for part in dataset.dataset.parts:
    print(f"Part {part.index}: {part.name} ({part.size_bytes} bytes, {part.num_rows} rows)")
    if part.samples:
        print(f"  First sample: {part.samples[0]}")

Related Pages

• Environment:Cohere_ai_Cohere_python_Python_SDK_Runtime