Cognite Document AI API
The Document AI API from Cognite — 2 operation(s) for document ai.
OpenAPI Specification
openapi: 3.1.0
info:
title: Cognite 3D Asset Mapping Document AI API
description: "# Introduction\nThis is the reference documentation for the Cognite API with\nan overview of all the available methods.\n\n# Postman\nSelect the **Download** button to download our OpenAPI specification to get started.\n\nTo import your data into Postman, select **Import**, and the Import modal opens.\nYou can import items by dragging or dropping files or folders. You can choose how to import your API and manage the import settings in **View Import Settings**.\n\nIn the Import Settings, set the **Folder organization** to **Tags**, select\n**Enable optional parameters** to turn off the settings, and select **Always inherit authentication** to turn on the settings. Select **Import**.\n\nSet the Authorization to **Oauth2.0**. By default, the settings are for Open Industrial Data. Navigate to [Cognite Hub](https://hub.cognite.com/open-industrial-data-211) to understand how to get the credentials for use in Postman.\n\nFor more information, see [Getting Started with Postman](https://developer.cognite.com/dev/guides/postman/).\n\n# Pagination\nMost resource types can be paginated, indicated by the field `nextCursor` in the response.\nBy passing the value of `nextCursor` as the cursor you will get the next page of `limit` results.\nNote that all parameters except `cursor` has to stay the same.\n\n# Parallel retrieval\nAs general guidance, Parallel Retrieval is a technique that should be used when due to query complexity, retrieval of data in a single request is significantly slower than it would otherwise be for a simple request. Parallel retrieval does not act as a speed multiplier on optimally running queries. By parallelizing such requests, data retrieval performance can be tuned to meet the client application needs. \n\nCDF supports parallel retrieval through the `partition` parameter, which has the format `m/n` where `n` is the amount of partitions you would like to split the entire data set into.\nIf you want to download the entire data set by splitting it into 10 partitions, do the following in parallel with `m` running from 1 to 10:\n - Make a request to `/events` with `partition=m/10`.\n - Paginate through the response by following the cursor as explained above. Note that the `partition` parameter needs to be passed to all subqueries.\n\nProcessing of parallel retrieval requests is subject to concurrency quota availability. The request returns the `429` response upon exceeding concurrency limits. See the Request throttling chapter below.\n\nTo prevent unexpected problems and to maximize read throughput, you should at most use 10 partitions. \nSome CDF resources will automatically enforce a maximum of 10 partitions.\nFor more specific and detailed information, please read the ```partition``` attribute documentation for the CDF resource you're using. \n\n# Requests throttling\nCognite Data Fusion (CDF) returns the HTTP `429` (too many requests) response status code when project capacity exceeds the limit.\n\nThe throttling can happen:\n - If a user or a project sends too many (more than allocated) concurrent requests.\n - If a user or a project sends a too high (more than allocated) rate of requests in a given amount of time.\n\nCognite recommends using a retry strategy based on truncated exponential backoff to handle sessions with HTTP response codes 429.\n\nCognite recommends using a reasonable number (up to 10) of `Parallel retrieval` partitions.\n\nFollowing these strategies lets you slow down the request frequency to maximize productivity without having to re-submit/retry failing requests.\n\nSee more [here](https://docs.cognite.com/dev/concepts/resource_throttling).\n\n# API versions\n## Version headers\nThis API uses calendar versioning, and version names follow the `YYYYMMDD` format.\nYou can find the versions currently available by using the version selector at the top of this page.\n\nTo use a specific API version, you can pass the `cdf-version: $version` header along with your requests to the API.\n\n## Beta versions\nThe beta versions provide a preview of what the stable version will look like in the future.\nBeta versions contain functionality that is reasonably mature, and highly likely to become a part of the stable API.\n\nBeta versions are indicated by a `-beta` suffix after the version name. For example, the beta version header for the\n2023-01-01 version is then `cdf-version: 20230101-beta`.\n\n## Alpha versions\nAlpha versions contain functionality that is new and experimental, and not guaranteed to ever become a part of the stable API.\nThis functionality presents no guarantee of service, so its use is subject to caution.\n\nAlpha versions are indicated by an `-alpha` suffix after the version name. For example, the alpha version header for\nthe 2023-01-01 version is then `cdf-version: 20230101-alpha`."
version: v1
contact:
name: Cognite Support
url: https://support.cognite.com
email: support@cognite.com
servers:
- url: https://{cluster}.cognitedata.com/api/v1/projects/{project}
description: The URL for the CDF cluster to connect to
variables:
cluster:
enum:
- api
- az-tyo-gp-001
- az-eastus-1
- az-power-no-northeurope
- westeurope-1
- asia-northeast1-1
- gc-dsm-gp-001
default: api
description: The CDF cluster to connect to
project:
default: publicdata
description: The CDF project name.
security:
- oidc-token:
- https://{cluster}.cognitedata.com/.default
- oauth2-client-credentials:
- https://{cluster}.cognitedata.com/.default
- oauth2-open-industrial-data:
- https://api.cognitedata.com/.default
- oauth2-auth-code:
- https://{cluster}.cognitedata.com/.default
tags:
- name: Document AI
paths:
/ai/tools/documents/ask:
post:
description: 'This API endpoint uses a language model to answer questions about documents. Provided with a natural language question and a list of files, the API returns a multi-part answer with references to the locations in the documents that were used to build the answer.
### Limitations
- Currently, we only support PDF files.
- The PDF files can be a maximum of 400 pages long.
- You can only pass 100 files in a single request.'
operationId: document_questioning_ai_tools_documents_ask_post
parameters:
- in: header
name: content-type
required: true
schema:
title: Content-Type
type: string
requestBody:
content:
application/json:
schema:
$ref: '#/components/schemas/DocumentQARequest'
required: true
responses:
'200':
content:
application/json:
schema:
$ref: '#/components/schemas/Answer'
description: Successful Response
'400':
$ref: '#/components/responses/ErrorResponse'
summary: Ask questions about one or more documents.
tags:
- Document AI
/ai/tools/documents/summarize:
post:
description: Generates concise summaries for a list of input documents using a language model.
operationId: documents_summary_ai_tools_documents_summarize_post
parameters:
- in: header
name: content-type
required: true
schema:
title: Content-Type
type: string
requestBody:
content:
application/json:
schema:
$ref: '#/components/schemas/DocumentSummarizationRequest'
required: true
responses:
'200':
content:
application/json:
schema:
$ref: '#/components/schemas/DocumentSummarizationResponse'
description: Successful Response
'400':
$ref: '#/components/responses/ErrorResponse'
summary: Summarize documents.
tags:
- Document AI
components:
schemas:
DocumentInternalId:
additionalProperties: false
properties:
id:
description: The file ID for the document.
title: Id
type: integer
required:
- id
title: DocumentInternalId
type: object
Content:
additionalProperties: false
properties:
references:
description: The document locations that this part of the answer is sourced from.
items:
$ref: '#/components/schemas/ReferenceAI'
title: References
type: array
text:
description: A part of the answer from the LLM.
title: Text
type: string
required:
- text
title: Content
type: object
DocumentSummarizationResponseItemInternalId:
additionalProperties: false
properties:
id:
description: The file ID for the document.
title: Id
type: integer
summary:
description: The summary of the document.
title: Summary
type: string
required:
- id
- summary
title: DocumentSummarizationResponseItemInternalId
type: object
DocumentSummarizationRequest:
additionalProperties: false
properties:
ignoreUnknownIds:
default: false
description: If `true`, the API will not fail if any documents are missing, but summaries for the missing documents will be excluded from the response.
title: Ignoreunknownids
type: boolean
items:
description: A list of documents to summarize.
items:
oneOf:
- $ref: '#/components/schemas/DocumentInternalId'
- $ref: '#/components/schemas/DocumentExternalId'
- $ref: '#/components/schemas/DocumentInstanceId'
maxItems: 1
minItems: 1
title: Items
type: array
language:
$ref: '#/components/schemas/DocumentsLanguage'
default: English
description: The language in which the answer should be provided.
required:
- items
title: DocumentSummarizationRequest
type: object
DocumentSummarizationResponseItemExternalId:
additionalProperties: false
properties:
externalId:
description: The external ID for the document. The value matches the externalId set in the Files API.
title: Externalid
type: string
summary:
description: The summary of the document.
title: Summary
type: string
required:
- externalId
- summary
title: DocumentSummarizationResponseItemExternalId
type: object
DocumentSummarizationResponseItemInstanceId:
additionalProperties: false
properties:
instanceId:
$ref: '#/components/schemas/InstanceIdAI'
description: The data modeling instance ID for the document.
summary:
description: The summary of the document.
title: Summary
type: string
required:
- instanceId
- summary
title: DocumentSummarizationResponseItemInstanceId
type: object
InstanceIdAI:
additionalProperties: false
properties:
externalId:
description: The data modeling external ID of the document.
pattern: "^[^\0]{1,256}$"
title: Externalid
type: string
space:
description: The data modeling space of the document.
pattern: ^[a-zA-Z][a-zA-Z0-9_-]{0,41}[a-zA-Z0-9]?$
title: Space
type: string
required:
- space
- externalId
title: InstanceId
type: object
DocumentSummarizationResponse:
additionalProperties: false
properties:
items:
description: A list of documents with summaries.
items:
oneOf:
- $ref: '#/components/schemas/DocumentSummarizationResponseItemInternalId'
- $ref: '#/components/schemas/DocumentSummarizationResponseItemExternalId'
- $ref: '#/components/schemas/DocumentSummarizationResponseItemInstanceId'
title: Items
type: array
required:
- items
title: DocumentSummarizationResponse
type: object
Answer:
additionalProperties: false
properties:
content:
description: The content of an answer consists of one or more parts. Each part can have a different set of document locations connected to it.
items:
$ref: '#/components/schemas/Content'
title: Content
type: array
required:
- content
title: Answer
type: object
DocumentsLanguage:
enum:
- Chinese
- Dutch
- English
- French
- German
- Italian
- Japanese
- Korean
- Latvian
- Norwegian
- Portuguese
- Spanish
- Swedish
title: DocumentsLanguage
type: string
DocumentQARequest:
additionalProperties: false
properties:
additionalContext:
description: Optional additional context that the model can use to improve its answer
maxLength: 2048
minLength: 10
title: Additionalcontext
type: string
fileIds:
description: A list of file IDs, external IDs, or instance IDs pointing to PDF documents
items:
oneOf:
- $ref: '#/components/schemas/DocumentInternalId'
- $ref: '#/components/schemas/DocumentExternalId'
- $ref: '#/components/schemas/DocumentInstanceId'
maxItems: 100
minItems: 1
title: Fileids
type: array
ignoreUnknownIds:
default: false
description: If `true`, the API will not fail if any documents are missing or not fully processed, but generate an answer based on available documents.
title: Ignoreunknownids
type: boolean
language:
$ref: '#/components/schemas/DocumentsLanguage'
default: English
description: The language in which the answer should be provided.
question:
description: The question to ask about the documents.
maxLength: 2048
minLength: 1
title: Question
type: string
required:
- question
- fileIds
title: DocumentQARequest
type: object
DocumentExternalId:
additionalProperties: false
properties:
externalId:
description: The external ID for the document. The value matches the externalId set in the Files API.
title: Externalid
type: string
required:
- externalId
title: DocumentExternalId
type: object
LocationAI:
additionalProperties: false
properties:
bottom:
title: Bottom
type: number
left:
title: Left
type: number
pageNumber:
description: The page number within the file. Page numbers start at 1.
title: Pagenumber
type: integer
right:
title: Right
type: number
top:
title: Top
type: number
required:
- pageNumber
- left
- right
- top
- bottom
title: Location
type: object
Error:
type: object
required:
- code
- message
description: Cognite API error.
properties:
code:
type: integer
description: HTTP status code.
format: int32
example: 401
message:
type: string
description: Error message.
example: Could not authenticate.
missing:
type: array
description: List of lookup objects that do not match any results.
items:
type: object
additionalProperties: true
duplicated:
type: array
description: List of objects that are not unique.
items:
type: object
additionalProperties: true
ReferenceAI:
additionalProperties: false
properties:
externalId:
description: The external ID of the file.
title: Externalid
type: string
fileId:
description: The ID of the file.
title: Fileid
type: integer
fileName:
description: The name of the file.
title: Filename
type: string
instanceId:
$ref: '#/components/schemas/InstanceIdAI'
description: The instance ID of the file.
locations:
description: The locations in the document that contain the relevant text.
items:
$ref: '#/components/schemas/LocationAI'
title: Locations
type: array
required:
- fileId
- fileName
- locations
title: Reference
type: object
DocumentInstanceId:
additionalProperties: false
properties:
instanceId:
$ref: '#/components/schemas/InstanceIdAI'
description: The data modeling instance ID for the document.
required:
- instanceId
title: DocumentInstanceId
type: object
responses:
ErrorResponse:
description: The response for a failed request.
content:
application/json:
schema:
type: object
required:
- error
properties:
error:
$ref: '#/components/schemas/Error'
securitySchemes:
oidc-token:
type: http
scheme: bearer
bearerFormat: OpenID Connect or OAuth2 token
description: Access token issued by the CDF project's configured identity provider. Access token must be an OpenID Connect token, and the project must be configured to accept OpenID Connect tokens. Use a header key of 'Authorization' with a value of 'Bearer $accesstoken'. The token can be obtained through any flow supported by the identity provider.
oauth2-client-credentials:
type: oauth2
description: Access token issued by the CDF project's configured identity provider. Access token must be an OpenID Connect token, and the project must be configured to accept OpenID Connect tokens. Use a header key of 'Authorization' with a value of 'Bearer $accesstoken'. The token can be obtained through any flow supported by the identity provider.
flows:
clientCredentials:
tokenUrl: https://your-idps.token.url/
scopes:
default: https://{cluster}.cognitedata.com/.default
oauth2-auth-code:
type: oauth2
description: Access token issued by the CDF project's configured identity provider. Access token must be an OpenID Connect token, and the project must be configured to accept OpenID Connect tokens. Use a header key of 'Authorization' with a value of 'Bearer $accesstoken'. The token can be obtained through any flow supported by the identity provider.
flows:
authorizationCode:
authorizationUrl: https://your-idps.authorization.url/
tokenUrl: https://your-idps.token.url/
scopes:
default: https://{cluster}.cognitedata.com/.default
oauth2-open-industrial-data:
type: oauth2
description: Auth flow for Open Industrial Data. Get your client secret from https://hub.cognite.com/open-industrial-data-211.
flows:
clientCredentials:
tokenUrl: https://login.microsoftonline.com/48d5043c-cf70-4c49-881c-c638f5796997/oauth2/v2.0/token
scopes:
default: https://api.cognitedata.com/.default
org-oidc-token:
type: openIdConnect
openIdConnectUrl: https://auth.cognite.com/.well-known/openid-configuration
description: 'Access token issued by the Cognite authorization server, and valid for the target organization. The token must
be an OpenID Connect token, and it can be obtained by performing an OIDC login flow toward `auth.cognite.com`.
This is a single URL for all CDF organizations.'
x-tagGroups:
- name: Changelog
tags:
- Changelog
- name: Organizations and projects
tags:
- Organizations
- Projects
- name: Identity and access management
tags:
- Principals
- Groups
- Security categories
- Sessions
- Token
- User profiles
- Project Deletion Reporting
- name: Data modeling
tags:
- Data Modeling
- Data models
- Spaces
- Views
- Containers
- Nodes
- Instances
- Statistics
- Streams
- Records
- name: Asset-centric data model
tags:
- Assets
- Time series
- Synthetic Time Series
- Data point subscriptions
- Events
- Files
- Sequences
- Geospatial
- Seismic
- name: 3D
tags:
- 3D Models
- 3D Model Revisions
- 3D Files
- 3D Asset Mapping
- 3D Contextualization
- 3D Jobs
- 3D Migration
- 3D Scenes
- name: Contextualization
tags:
- Entity matching
- Entity matching pipelines
- Engineering diagrams
- Vision
- Advanced joins
- name: Cognite AI
tags:
- Agents
- Skills
- Chat Completions
- Document AI
- Models
- name: Documents
tags:
- Documents
- Document preview
- name: Data ingestion
tags:
- Raw
- Extraction Pipelines
- Extraction Pipelines Runs
- Extraction Pipelines Config
- Extractors
- name: Data organization
tags:
- Data sets
- Data domains
- Data products
- Rule sets
- Labels
- Relationships
- Annotations
- name: Transformations
tags:
- Transformations
- Transformation Jobs
- Transformation Schedules
- Transformation Notifications
- Query
- Schema
- name: Functions
tags:
- Functions
- Function calls
- Function schedules
- name: Hosted Extractors
tags:
- Sources
- Jobs
- Destinations
- Mappings
- name: PostgreSQL Gateway
tags:
- Postgres Gateway Users
- Postgres Gateway Tables
- name: SAP Writeback
tags:
- SAP Instances
- SAP Endpoints
- Schema Mappings
- Writeback Requests
- name: Data workflows
tags:
- Workflows
- Workflow versions
- Workflow executions
- Workflow triggers
- Tasks
- Workers
- name: Simulators
tags:
- Simulators
- Simulator Integrations
- Simulator Models
- Simulator Routines
- Simulation Runs
- Simulator Logs
- name: Units
tags:
- Units
- Unit Systems
- name: ''
tags:
- ''