Skip to main content
POST
/
huggingface-datasets
curl --request POST \
  --url https://apigcp.trynia.ai/v2/huggingface-datasets \
  --header 'Authorization: Bearer <token>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "url": "dair-ai/emotion"
}
'
{
  "id": "550e8400-e29b-41d4-a716-446655440000",
  "dataset_id": "emotion",
  "url": "https://huggingface.co/datasets/dair-ai/emotion",
  "status": "processing",
  "created_at": "2024-01-15T10:30:00Z",
  "updated_at": "2024-01-15T10:30:00Z",
  "owner": "dair-ai",
  "description": "Emotion is a dataset of English Twitter messages with six basic emotions.",
  "splits": [
    "train",
    "test",
    "validation"
  ],
  "columns": [
    {
      "name": "text",
      "dtype": "string"
    },
    {
      "name": "label",
      "dtype": "int64"
    }
  ],
  "row_count": 20000,
  "indexed_row_count": 0,
  "chunk_count": 0
}

Authorizations

​
Authorization
string
header
required

API key must be provided in the Authorization header

Body

application/json
​
url
string
required

HuggingFace dataset URL or identifier. Supports multiple formats:

Example:

"dair-ai/emotion"

​
config
string | null

Dataset configuration name (for multi-config datasets)

​
add_as_global_source
boolean
default:true

Add to global shared pool (default true). Set false for private indexing.

Response

HuggingFace dataset indexing started or completed successfully

​
id
string

Unique identifier for the data source

​
dataset_id
string

Dataset identifier (e.g., "squad", "emotion")

​
url
string

Canonical HuggingFace dataset URL

​
status
enum<string>

Current indexing status

Available options:
pending,
processing,
completed,
failed,
error
​
created_at
string<date-time>
​
updated_at
string<date-time>
​
owner
string | null

Dataset owner/organization

​
description
string | null

Dataset description

​
splits
string[]

Available dataset splits (e.g., ["train", "test", "validation"])

​
columns
object[]

Dataset columns with names and data types

​
row_count
integer
default:0

Total number of rows in the dataset

​
indexed_row_count
integer
default:0

Number of rows actually indexed (may differ due to sampling)

​
chunk_count
integer
default:0

Number of text chunks created

​
sampling_strategy
string | null

Sampling strategy used (full or sampled)

​
license
string | null

Dataset license

​
error
string | null

Error message if status is 'failed'