arazzo: 1.0.1
info:
  title: Viam Curate a Training Dataset
  summary: Find binary data by filter, tag the matches, and add them to a dataset.
  description: >-
    Prepares captured images for ML training. The flow queries binary data items
    with a filter, applies tags to the supplied binary ids, and adds those items
    to a dataset that can later feed a training job. Each request body is inlined
    so the curation can be executed directly against the Viam Data Client API.
  version: 1.0.0
sourceDescriptions:
- name: dataClientApi
  url: ../openapi/viam-data-client-api-openapi.yml
  type: openapi
workflows:
- workflowId: curate-dataset
  summary: Query binary data, tag the items, and add them to a dataset.
  description: >-
    Runs a binary-data filter query, adds tags to the supplied binary ids, and
    adds the same items to the target dataset.
  inputs:
    type: object
    required:
    - apiKey
    - filter
    - binaryIds
    - tags
    - datasetId
    properties:
      apiKey:
        type: string
        description: Viam API key value sent in the key header.
      filter:
        type: object
        description: Data request filter object used to scope the binary-data query.
      binaryIds:
        type: array
        description: Binary id objects to tag and add to the dataset.
        items:
          type: object
      tags:
        type: array
        description: String tags to apply to the matched binary data items.
        items:
          type: string
      datasetId:
        type: string
        description: The dataset the binary items are added to.
  steps:
  - stepId: findBinary
    description: Query binary data items matching the supplied filter (metadata only).
    operationId: binaryDataByFilter
    parameters:
    - name: key
      in: header
      value: $inputs.apiKey
    requestBody:
      contentType: application/json
      payload:
        data_request: $inputs.filter
        include_binary: false
        count_only: false
    successCriteria:
    - condition: $statusCode == 200
  - stepId: tagBinary
    description: Apply the supplied tags to the supplied binary data items.
    operationId: addTagsToBinaryDataByIDs
    parameters:
    - name: key
      in: header
      value: $inputs.apiKey
    requestBody:
      contentType: application/json
      payload:
        binary_ids: $inputs.binaryIds
        tags: $inputs.tags
    successCriteria:
    - condition: $statusCode == 200
  - stepId: addToDataset
    description: Add the tagged binary data items to the target dataset.
    operationId: addBinaryDataToDatasetByIDs
    parameters:
    - name: key
      in: header
      value: $inputs.apiKey
    requestBody:
      contentType: application/json
      payload:
        binary_ids: $inputs.binaryIds
        dataset_id: $inputs.datasetId
    successCriteria:
    - condition: $statusCode == 200
  outputs:
    datasetId: $inputs.datasetId