Structify structure API

openapi: 3.2.0
info:
  contact:
    email: team@structify.ai
    name: Structify Team
  description: Every enterprise's data team.
  license:
    name: Discuss directly with founders for license.
    url: https://structify.ai
  title: Structify account Structure API
  version: 0.1.0
servers:
- description: Production server
  url: https://api.structify.ai
- description: Local server
  url: http://localhost:8080
security:
- api_key: []
- session_token: []
tags:
- name: structure
paths:
  /jobs/status_aggregated:
    post:
      description: 'Returns counts of jobs by status (completed, running, failed, queued).

        Exactly one of job_ids or dataset_name must be provided.

        This endpoint can handle large numbers of job IDs since it returns aggregated data instead of individual job details.'
      operationId: jobs_status_aggregated
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/AggregatedJobRequest'
        required: true
      responses:
        '200':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/JobCount'
          description: Aggregated status counts for specified jobs.
        '400':
          description: Bad request - exactly one of job_ids or dataset_name must be provided.
      security:
      - api_key: []
      - session_token: []
      summary: Get aggregated status counts for multiple jobs.
      tags:
      - structure
  /structure/enhance_property:
    post:
      description: Returns a job id that can be waited on until the request is finished.
      operationId: structure_enhance_property
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/EnhancePropertyParams'
        required: true
      responses:
        '200':
          content:
            text/plain:
              schema:
                type: string
          description: The job id.
      security:
      - session_token: []
      - api_key: []
      summary: enhance_property a given entity by searching specifically for a certain missing property.
      tags:
      - structure
  /structure/enhance_relationship:
    post:
      description: Returns a job id that can be waited on until the request is finished.
      operationId: structure_enhance_relationship
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/EnhanceRelationshipParams'
        required: true
      responses:
        '200':
          content:
            text/plain:
              schema:
                type: string
          description: The job id.
      security:
      - session_token: []
      - api_key: []
      summary: Enhance relationships between entities by searching for missing connections.
      tags:
      - structure
  /structure/find_relationship:
    post:
      operationId: structure_find_relationship
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/FindRelationshipParams'
        required: true
      responses:
        '200':
          content:
            text/plain:
              schema:
                type: string
          description: The job id.
      security:
      - session_token: []
      - api_key: []
      summary: Attempt to find the given relation between two entities.
      tags:
      - structure
  /structure/is_complete:
    post:
      operationId: structure_is_complete
      requestBody:
        content:
          application/json:
            schema:
              items:
                type: string
              type: array
        required: true
      responses:
        '200':
          content:
            text/plain:
              schema:
                type: string
          description: All tasks finished.
      security:
      - api_key: []
      - session_token: []
      summary: Wait for all specified async tasks to be completed.
      tags:
      - structure
  /structure/job_status:
    post:
      description: If only dataset_name is provided, up to 1000 of the most recent jobs for that dataset will be returned.
      operationId: structure_job_status
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/JobRequest'
        required: true
      responses:
        '200':
          content:
            application/json:
              schema:
                items:
                  $ref: '#/components/schemas/JobNotification'
                type: array
          description: Status of all specified jobs.
      security:
      - api_key: []
      - session_token: []
      summary: Get status for specified jobs with detailed information similar to websocket endpoint.
      tags:
      - structure
  /structure/pdf:
    post:
      description: 'For single mode, creates one job for the entire PDF.

        For batch mode, creates one job per PDF page.'
      operationId: structure_pdf
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/StructurePdfRequest'
        required: true
      responses:
        '200':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/StructurePdfResponse'
          description: List of created job IDs
      security:
      - api_key: []
      - session_token: []
      summary: Structure a PDF document into the given dataset.
      tags:
      - structure
  /structure/run_async:
    post:
      description: Returns a token that can be waited on until the request is finished.
      operationId: structure_run_async
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/RunRequestParams'
        required: true
      responses:
        '200':
          content:
            text/plain:
              schema:
                type: string
          description: The job id.
      security:
      - api_key: []
      - session_token: []
      summary: Structure an unstructured data source into the given dataset in an async fashion.
      tags:
      - structure
components:
  schemas:
    EnhancePropertyParams:
      allOf:
      - $ref: '#/components/schemas/EnhancePropertyArgs'
      - properties:
          node_id:
            allOf:
            - $ref: '#/components/schemas/WorkflowSessionNodeId'
        type: object
    WebSearch:
      properties:
        banned_domains:
          items:
            type: string
          type: array
        starting_searches:
          items:
            type: string
          type: array
        starting_urls:
          items:
            type: string
          type: array
      type: object
    EnhanceRelationshipArgs:
      allOf:
      - $ref: '#/components/schemas/WebSearch'
      - properties:
          allow_extra_entities:
            type: boolean
          entity_id:
            $ref: '#/components/schemas/EntityId'
          relationship_name:
            type: string
        required:
        - entity_id
        - relationship_name
        type: object
    JobStatus:
      enum:
      - Queued
      - Running
      - Completed
      - Failed
      type: string
    LLMRelationship:
      additionalProperties: false
      properties:
        properties:
          additionalProperties:
            $ref: '#/components/schemas/BasicPropValue'
          type: object
        source:
          minimum: 0
          type: integer
        target:
          minimum: 0
          type: integer
        type:
          type: string
      required:
      - type
      - source
      - target
      type: object
    PDFIngestor:
      additionalProperties: false
      description: Ingest all pages of a PDF and process them independently.
      properties:
        page:
          format: int32
          minimum: 0
          type:
          - integer
          - 'null'
        path:
          type: string
      required:
      - path
      type: object
    FindRelationshipArgs:
      allOf:
      - $ref: '#/components/schemas/WebSearch'
      - properties:
          allow_extra_entities:
            type: boolean
          from_id:
            $ref: '#/components/schemas/EntityId'
          relationship_name:
            type: string
          to_id:
            $ref: '#/components/schemas/EntityId'
        required:
        - from_id
        - to_id
        - relationship_name
        type: object
    JobTarget:
      properties:
        entity_id:
          $ref: '#/components/schemas/EntityId'
        property_names:
          items:
            type: string
          type:
          - array
          - 'null'
        relationship_name:
          type:
          - string
          - 'null'
      required:
      - entity_id
      type: object
    LLMEntity:
      additionalProperties: false
      properties:
        id:
          minimum: 0
          type: integer
        properties:
          additionalProperties:
            $ref: '#/components/schemas/BasicPropValue'
          type: object
        type:
          type: string
      required:
      - id
      - type
      - properties
      type: object
    StructurePdfRequest:
      additionalProperties: false
      properties:
        dataset:
          type: string
        instructions:
          type:
          - string
          - 'null'
        model:
          type:
          - string
          - 'null'
        node_id:
          allOf:
          - $ref: '#/components/schemas/WorkflowSessionNodeId'
        pages:
          items:
            format: int32
            minimum: 0
            type: integer
          type:
          - array
          - 'null'
        path:
          type: string
      required:
      - path
      - dataset
      type: object
    EnhancePropertyArgs:
      allOf:
      - $ref: '#/components/schemas/WebSearch'
      - properties:
          allow_extra_entities:
            type: boolean
          entity_id:
            $ref: '#/components/schemas/EntityId'
          property_name:
            type: string
        required:
        - entity_id
        - property_name
        type: object
    RunRequestParams:
      additionalProperties: false
      properties:
        dataset:
          type: string
        instructions:
          type:
          - string
          - 'null'
        model:
          type:
          - string
          - 'null'
        node_id:
          allOf:
          - $ref: '#/components/schemas/WorkflowSessionNodeId'
        save_requirement:
          items:
            $ref: '#/components/schemas/SaveRequirement'
          type: array
        seeded_entity:
          $ref: '#/components/schemas/LLMKnowledgeGraph'
        source:
          $ref: '#/components/schemas/Input'
      required:
      - source
      - dataset
      type: object
    RequiredEntity:
      additionalProperties: false
      properties:
        entity_id:
          allOf:
          - $ref: '#/components/schemas/EntityId'
        seeded_entity_id:
          description: The integer id corresponding to an entity in the seeded entity graph (different from the global dataset entity id)
          minimum: 0
          type: integer
      required:
      - seeded_entity_id
      type: object
    JobNotification:
      properties:
        dataset_name:
          type: string
        job_id:
          $ref: '#/components/schemas/JobId'
        status:
          allOf:
          - $ref: '#/components/schemas/JobStatus'
        target:
          allOf:
          - $ref: '#/components/schemas/JobTarget'
      required:
      - dataset_name
      - job_id
      type: object
    JobCount:
      properties:
        completed:
          format: int64
          type: integer
        failed:
          format: int64
          type: integer
        queued:
          format: int64
          type: integer
        running:
          format: int64
          type: integer
        total:
          format: int64
          type: integer
      required:
      - total
      - completed
      - failed
      - running
      - queued
      type: object
    AggregatedJobRequest:
      properties:
        dataset_name:
          type:
          - string
          - 'null'
        job_ids:
          items:
            $ref: '#/components/schemas/JobId'
          type:
          - array
          - 'null'
        node_id:
          allOf:
          - $ref: '#/components/schemas/WorkflowSessionNodeId'
      type: object
    RequiredProperty:
      additionalProperties: false
      properties:
        property_names:
          description: If there are multiple properties, it can match just one of them
          items:
            type: string
          type: array
        table_name:
          description: The table name of the entity to update
          type: string
      required:
      - table_name
      - property_names
      type: object
    SaveRequirement:
      description: It's an OR statement across these.
      oneOf:
      - $ref: '#/components/schemas/RequiredRelationship'
      - $ref: '#/components/schemas/RequiredEntity'
      - $ref: '#/components/schemas/RequiredProperty'
    StructurePdfResponse:
      properties:
        job_ids:
          items:
            $ref: '#/components/schemas/JobId'
          type: array
      required:
      - job_ids
      type: object
    Input:
      oneOf:
      - properties:
          PDF:
            $ref: '#/components/schemas/PDFIngestor'
        required:
        - PDF
        type: object
      - properties:
          Web:
            $ref: '#/components/schemas/WebSearch'
        required:
        - Web
        type: object
    EntityId:
      format: uuid
      type: string
    BasicPropValue:
      oneOf:
      - type: string
      - format: int64
        type: integer
      - type: boolean
      - format: double
        type: number
    FindRelationshipParams:
      allOf:
      - $ref: '#/components/schemas/FindRelationshipArgs'
      - type: object
    RequiredRelationship:
      additionalProperties: false
      properties:
        relationship_name:
          type: string
      required:
      - relationship_name
      type: object
    JobRequest:
      properties:
        dataset_name:
          type:
          - string
          - 'null'
        job_ids:
          items:
            $ref: '#/components/schemas/JobId'
          type:
          - array
          - 'null'
      type: object
    WorkflowSessionNodeId:
      format: uuid
      type: string
    LLMKnowledgeGraph:
      additionalProperties: false
      description: 'Knowledge graph info structured to deserialize and display

        in the same format that the LLM outputs.

        Also the first representation of an LLM output in the pipeline from raw tool output to being merged into a DB'
      properties:
        entities:
          items:
            $ref: '#/components/schemas/LLMEntity'
          type: array
        relationships:
          items:
            $ref: '#/components/schemas/LLMRelationship'
          type: array
      required:
      - entities
      type: object
    EnhanceRelationshipParams:
      allOf:
      - $ref: '#/components/schemas/EnhanceRelationshipArgs'
      - properties:
          node_id:
            allOf:
            - $ref: '#/components/schemas/WorkflowSessionNodeId'
        type: object
    JobId:
      format: uuid
      type: string
  securitySchemes:
    api_key:
      in: header
      name: api_key
      type: apiKey
    session_token:
      scheme: bearer
      type: http