Hugging Face Hub API — Datasets

naftiko: 1.0.0-alpha2
info:
  label: Hugging Face Hub API — Datasets
  description: 'Hugging Face Hub API — Datasets. 4 operations. Lead operation: List Datasets on the Hub. Self-contained Naftiko
    capability covering one Hugging Face business surface.'
  tags:
  - Hugging Face
  - Datasets
  created: '2026-05-19'
  modified: '2026-05-19'
binds:
- namespace: env
  keys:
    HUGGING_FACE_API_KEY: HUGGING_FACE_API_KEY
capability:
  consumes:
  - type: http
    namespace: hub-datasets
    baseUri: https://huggingface.co/api
    description: Hugging Face Hub API — Datasets business capability. Self-contained, no shared references.
    resources:
    - name: datasets
      path: /datasets
      operations:
      - name: listdatasets
        method: GET
        description: List Datasets on the Hub
        outputRawFormat: json
        outputParameters:
        - name: result
          type: object
          value: $.
        inputParameters:
        - name: search
          in: query
          type: string
          description: Full-text search query to filter datasets
        - name: author
          in: query
          type: string
          description: Filter by dataset author or organization
        - name: filter
          in: query
          type: string
          description: Filter by tags (comma-separated)
        - name: sort
          in: query
          type: string
          description: Property to sort results by
        - name: direction
          in: query
          type: string
          description: Sort direction
        - name: limit
          in: query
          type: integer
          description: Number of results to return
        - name: full
          in: query
          type: boolean
          description: Whether to return full dataset information
    - name: datasets-tags-by-type
      path: /datasets-tags-by-type
      operations:
      - name: listdatasettags
        method: GET
        description: List Dataset Tags Grouped by Type
        outputRawFormat: json
        outputParameters:
        - name: result
          type: object
          value: $.
    - name: datasets-repo_id
      path: /datasets/{repo_id}
      operations:
      - name: getdataset
        method: GET
        description: Get Dataset Information
        outputRawFormat: json
        outputParameters:
        - name: result
          type: object
          value: $.
        inputParameters:
        - name: repo_id
          in: path
          type: string
          description: The dataset repository ID
          required: true
    - name: datasets-repo_id-parquet
      path: /datasets/{repo_id}/parquet
      operations:
      - name: getdatasetparquet
        method: GET
        description: Get Dataset Parquet Files
        outputRawFormat: json
        outputParameters:
        - name: result
          type: object
          value: $.
        inputParameters:
        - name: repo_id
          in: path
          type: string
          required: true
    authentication:
      type: bearer
      token: '{{env.HUGGING_FACE_API_KEY}}'
  exposes:
  - type: rest
    namespace: hub-datasets-rest
    port: 8080
    description: REST adapter for Hugging Face Hub API — Datasets. One Spectral-compliant resource per consumed operation,
      prefixed with /v1.
    resources:
    - path: /v1/datasets
      name: datasets
      description: REST surface for datasets.
      operations:
      - method: GET
        name: listdatasets
        description: List Datasets on the Hub
        call: hub-datasets.listdatasets
        with:
          search: rest.search
          author: rest.author
          filter: rest.filter
          sort: rest.sort
          direction: rest.direction
          limit: rest.limit
          full: rest.full
        outputParameters:
        - type: object
          mapping: $.
    - path: /v1/datasets-tags-by-type
      name: datasets-tags-by-type
      description: REST surface for datasets-tags-by-type.
      operations:
      - method: GET
        name: listdatasettags
        description: List Dataset Tags Grouped by Type
        call: hub-datasets.listdatasettags
        outputParameters:
        - type: object
          mapping: $.
    - path: /v1/datasets/{repo-id}
      name: datasets-repo-id
      description: REST surface for datasets-repo_id.
      operations:
      - method: GET
        name: getdataset
        description: Get Dataset Information
        call: hub-datasets.getdataset
        with:
          repo_id: rest.repo_id
        outputParameters:
        - type: object
          mapping: $.
    - path: /v1/datasets/{repo-id}/parquet
      name: datasets-repo-id-parquet
      description: REST surface for datasets-repo_id-parquet.
      operations:
      - method: GET
        name: getdatasetparquet
        description: Get Dataset Parquet Files
        call: hub-datasets.getdatasetparquet
        with:
          repo_id: rest.repo_id
        outputParameters:
        - type: object
          mapping: $.
  - type: mcp
    namespace: hub-datasets-mcp
    port: 9090
    transport: http
    description: MCP adapter for Hugging Face Hub API — Datasets. One tool per consumed operation, routed inline through this
      capability's consumes block.
    tools:
    - name: list-datasets-hub
      description: List Datasets on the Hub
      hints:
        readOnly: true
        destructive: false
        idempotent: true
      call: hub-datasets.listdatasets
      with:
        search: tools.search
        author: tools.author
        filter: tools.filter
        sort: tools.sort
        direction: tools.direction
        limit: tools.limit
        full: tools.full
      outputParameters:
      - type: object
        mapping: $.
    - name: list-dataset-tags-grouped-type
      description: List Dataset Tags Grouped by Type
      hints:
        readOnly: true
        destructive: false
        idempotent: true
      call: hub-datasets.listdatasettags
      outputParameters:
      - type: object
        mapping: $.
    - name: get-dataset-information
      description: Get Dataset Information
      hints:
        readOnly: true
        destructive: false
        idempotent: true
      call: hub-datasets.getdataset
      with:
        repo_id: tools.repo_id
      outputParameters:
      - type: object
        mapping: $.
    - name: get-dataset-parquet-files
      description: Get Dataset Parquet Files
      hints:
        readOnly: true
        destructive: false
        idempotent: true
      call: hub-datasets.getdatasetparquet
      with:
        repo_id: tools.repo_id
      outputParameters:
      - type: object
        mapping: $.