Dust Upsert a Document and Search the Data Source

arazzo: 1.0.1
info:
  title: Dust Upsert a Document and Search the Data Source
  summary: Upsert a document into a data source, wait for the upsert queue to drain, then search for it.
  description: >-
    Loads knowledge into a Dust data source and confirms it is retrievable. The
    workflow upserts a document into a folder data source, polls the upsert
    queue status until no upsert workflows are running, then runs a semantic
    search against the data source. Each step spells out its request inline so
    the flow can be read and executed without opening the underlying OpenAPI
    description.
  version: 1.0.0
sourceDescriptions:
- name: datasourcesApi
  url: ../openapi/dust-tt-datasources-api-openapi.yml
  type: openapi
workflows:
- workflowId: upsert-document-and-search
  summary: Upsert a document, wait for indexing, and search the data source.
  description: >-
    Upserts a single text document into a data source, polls the upsert queue
    until it is empty, and then issues a search query against the data source.
  inputs:
    type: object
    required:
    - apiToken
    - wId
    - spaceId
    - dsId
    - documentId
    - title
    - text
    - query
    properties:
      apiToken:
        type: string
        description: Dust API key used as the Bearer token.
      wId:
        type: string
        description: The workspace identifier.
      spaceId:
        type: string
        description: The space identifier containing the data source.
      dsId:
        type: string
        description: The data source identifier.
      documentId:
        type: string
        description: The document identifier to upsert.
      title:
        type: string
        description: Title of the document.
      text:
        type: string
        description: Plain text content of the document.
      query:
        type: string
        description: The search query to run after indexing.
      topK:
        type: number
        description: Number of search results to return (defaults to 5).
        default: 5
  steps:
  - stepId: upsertDocument
    description: >-
      Upsert the document into the data source synchronously so it is queued for
      indexing.
    operationPath: '{$sourceDescriptions.datasourcesApi.url}#/paths/~1api~1v1~1w~1{wId}~1spaces~1{spaceId}~1data_sources~1{dsId}~1documents~1{documentId}/post'
    parameters:
    - name: Authorization
      in: header
      value: Bearer $inputs.apiToken
    - name: wId
      in: path
      value: $inputs.wId
    - name: spaceId
      in: path
      value: $inputs.spaceId
    - name: dsId
      in: path
      value: $inputs.dsId
    - name: documentId
      in: path
      value: $inputs.documentId
    requestBody:
      contentType: application/json
      payload:
        title: $inputs.title
        mime_type: text/plain
        text: $inputs.text
        light_document_output: true
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      documentId: $response.body#/document/document_id
  - stepId: checkQueue
    description: >-
      Check how many upsert workflows are still running for this data source.
      Continue polling while any are running, otherwise proceed to search.
    operationPath: '{$sourceDescriptions.datasourcesApi.url}#/paths/~1api~1v1~1w~1{wId}~1spaces~1{spaceId}~1data_sources~1{dsId}~1check_upsert_queue/get'
    parameters:
    - name: Authorization
      in: header
      value: Bearer $inputs.apiToken
    - name: wId
      in: path
      value: $inputs.wId
    - name: spaceId
      in: path
      value: $inputs.spaceId
    - name: dsId
      in: path
      value: $inputs.dsId
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      runningCount: $response.body#/running_count
    onSuccess:
    - name: queueDrained
      type: goto
      stepId: searchDataSource
      criteria:
      - context: $response.body
        condition: $.running_count == 0
        type: jsonpath
    - name: stillIndexing
      type: goto
      stepId: checkQueue
      criteria:
      - context: $response.body
        condition: $.running_count > 0
        type: jsonpath
  - stepId: searchDataSource
    description: >-
      Run a semantic search against the data source to confirm the upserted
      document is retrievable.
    operationPath: '{$sourceDescriptions.datasourcesApi.url}#/paths/~1api~1v1~1w~1{wId}~1spaces~1{spaceId}~1data_sources~1{dsId}~1search/get'
    parameters:
    - name: Authorization
      in: header
      value: Bearer $inputs.apiToken
    - name: wId
      in: path
      value: $inputs.wId
    - name: spaceId
      in: path
      value: $inputs.spaceId
    - name: dsId
      in: path
      value: $inputs.dsId
    - name: query
      in: query
      value: $inputs.query
    - name: top_k
      in: query
      value: $inputs.topK
    - name: full_text
      in: query
      value: false
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      documents: $response.body#/documents
  outputs:
    documentId: $steps.upsertDocument.outputs.documentId
    documents: $steps.searchDataSource.outputs.documents