Qlik Sense Profile Insights API

openapi: 3.2.0
info:
  title: ml Profile Insights API
  version: ''
servers:
- url: https://{tenant}.{region}.qlikcloud.com
  variables:
    region:
      default: us
      description: The region the tenant is hosted in
    tenant:
      default: your-tenant
      description: Name of the tenant that will be called
tags:
- name: profile-insights
paths:
  /api/v1/ml/profile-insights:
    post:
      tags:
      - profile-insights
      summary: Start profile insights creation
      responses:
        '200':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ProfileInsightsGetResponse'
          description: '`OK`'
        '202':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ProfileInsightsGetResponse'
          headers:
            Content-Location:
              schema:
                type: string
                format: uri
              required: true
              description: 'URL to poll for the status of this resource. When it''s `ready`

                you have everything you need.

                '
          description: '`Accepted`'
        '400':
          $ref: '#/components/responses/400BadRequest'
        '401':
          $ref: '#/components/responses/401Unauthorized'
        '403':
          $ref: '#/components/responses/403Forbidden'
        '500':
          $ref: '#/components/responses/500InternalError'
        '502':
          $ref: '#/components/responses/502BadGateway'
        '503':
          $ref: '#/components/responses/503ServiceUnavailable'
        default:
          $ref: '#/components/responses/FailureDefault'
      description: 'Starts creating profile insights for an experiment dataset.

        This is an asynchronous operation. A `202 Accepted` response indicates

        that the process has started successfully. Use the link in the response

        to check the status.'
      operationId: postProfileInsights
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/ProfileInsightsInput'
      x-qlik-visibility: public
      x-qlik-stability: stable
      x-qlik-deprecated: false
      x-qlik-tier:
        tier: '2'
        limit: 100
  /api/v1/ml/profile-insights/{dataSetId}:
    get:
      tags:
      - profile-insights
      summary: Get profile insights
      responses:
        '200':
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/ProfileInsightsGetResponse'
          description: '`OK`'
        '400':
          $ref: '#/components/responses/400BadRequest'
        '401':
          $ref: '#/components/responses/401Unauthorized'
        '403':
          $ref: '#/components/responses/403Forbidden'
        '404':
          $ref: '#/components/responses/404NotFound'
        '500':
          $ref: '#/components/responses/500InternalError'
        '502':
          $ref: '#/components/responses/502BadGateway'
        '503':
          $ref: '#/components/responses/503ServiceUnavailable'
        default:
          $ref: '#/components/responses/FailureDefault'
      parameters:
      - in: query
        name: experimentVersionId
        schema:
          type: string
        required: false
        description: 'The optional experimentVersionId query parameter for profile-insights

          GET requests. When provided after a version has been trained, it gets

          the profile insights snapshot used in previous versions rather than

          new results.

          '
      - $ref: '#/components/parameters/dataSetId'
      - $ref: '#/components/parameters/target'
      - $ref: '#/components/parameters/experimentType'
      description: 'Retrieves profile insights for the specified dataset. If you received a

        `202 Accepted` response from `POST /ml/profile-insights`, poll this

        endpoint until a `200 OK` response with `ready` status is returned.'
      operationId: getProfileInsights
      x-qlik-visibility: public
      x-qlik-stability: stable
      x-qlik-deprecated: false
      x-qlik-tier:
        tier: '1'
        limit: 1000
components:
  schemas:
    TableProfile:
      type: object
      required:
      - name
      - numberOfRows
      - fieldProfiles
      properties:
        name:
          type: string
        sizeInBytes:
          type: number
        numberOfRows:
          type: integer
        fieldProfiles:
          type: array
          items:
            $ref: '#/components/schemas/FieldProfile'
      description: Profile metadata for a single table/dataset.
      additionalProperties: false
    FeaturesList:
      type: array
      items:
        $ref: '#/components/schemas/Feature'
      description: 'List of features from your dataset for creating Experiment

        Versions. This appears in from ProfileInsights response (in the

        defaultVersionConfig). You can adjust the default settings before

        using it as input to create or update Experiment Versions.

        '
    TableSample:
      type: object
      properties:
        name:
          type: string
        records:
          type: array
          items:
            $ref: '#/components/schemas/TableRecord'
        fieldNames:
          type: array
          items:
            type: string
      description: Sample values from a table/dataset.
      additionalProperties: false
    DatasetOrigin:
      enum:
      - new
      - changed
      - refreshed
      - same
      type: string
      default: new
      description: Whether this is a new or other dataset
    ChangeType:
      enum:
      - categorical
      - numeric
      - date
      - freetext
      type:
      - string
      - 'null'
      default: null
      description: 'Indicates if you want to change the featureType for this

        feature within the experiment version

        '
    TenantId:
      type: string
      format: uuid
      description: Tenant ID for this entity
      x-qlik-filterable: true
    DataSetProfile:
      type: object
      required:
      - meta
      properties:
        meta:
          $ref: '#/components/schemas/ProfileMetadata'
        samples:
          type: array
          items:
            $ref: '#/components/schemas/TableSample'
        profiles:
          type: array
          items:
            $ref: '#/components/schemas/TableProfile'
      description: Full dataset profile from the Profile Service.
      additionalProperties: false
    TableRecord:
      type: object
      properties:
        values:
          type: array
          items:
            type: string
      description: A single record of sample values.
      additionalProperties: false
    IncludeProfileInsightsField:
      enum:
      - dataSetProfile
      type: string
      example: dataSetProfile
      description: 'Fields to include in the response.


        Currently only supported value is `dataSetProfile`.


        When `dataSetProfile` is specified, the response includes the full dataset profile.

        '
    DataSetId:
      type: string
      example: 672e55cfcadfb8a18281523e
      description: The Qlik catalog dataset ID
    Feature:
      type: object
      properties:
        name:
          type: string
          example: ColumnA
          description: Name of the feature column
        include:
          type: boolean
          description: 'Include this feature in your experiment version? Default

            here is based on insights for this feature

            (e.g. willBeDropped).

            '
        dataType:
          $ref: '#/components/schemas/DataType'
        changeType:
          $ref: '#/components/schemas/ChangeType'
        featureType:
          $ref: '#/components/schemas/FeatureType'
        parentFeature:
          type:
          - string
          - 'null'
          default: null
          description: The parent feature name for engineered features. e.g. `OrderDate` may be the parent of its engineered features (features extracted from parent) like `OrderDate.YEAR`, `OrderDate.MONTH`, etc.
      description: A feature (column) from your dataset
      additionalProperties: false
    FeatureInsights:
      type: object
      required:
      - name
      - experimentTypes
      - willBeDropped
      - insights
      - cannotBeTarget
      properties:
        name:
          type: string
          description: Name of the feature insight
        insights:
          type: array
          items:
            $ref: '#/components/schemas/Insights'
          description: 'List of insights about this feature.

            '
        willBeDropped:
          type: boolean
          description: 'Whether this feature will be dropped. Traits like high cardinality

            make some features less predictive or too costly to merit use.

            '
        cannotBeTarget:
          type: boolean
          description: Whether a feature cannot be the target field
        experimentTypes:
          type: array
          items:
            $ref: '#/components/schemas/ExperimentType'
          description: Experiment types in this feature insight
        defaultFeatureType:
          $ref: '#/components/schemas/FeatureType'
        engineeredFeatures:
          type: array
          items:
            type: string
          example: "[\n  `${featureName}.YEAR`,\n  `${featureName}.MONTH`\n]\n"
          description: 'Preliminary list of engineered features as strings. If subsequent

            processing validates them, they''ll be converted to EngineeredFeature

            objects within a NestedColumn, each of which may contain its own

            FeatureInsights.

            '
        estimatedMaxForecastHorizon:
          type: integer
          description: 'Only applies for time series experiment types. This intial estimate

            of the combined max forecast window and gap (aka - horizon). It only

            applies to possible date index columns. After the experiment version

            is created, we get a more precise number for subsequent versions.

            When training data is grouped, this estimate may be less accurate.

            '
      description: 'Metadata about the features in your dataset, generated when you create

        ProfileInsights.

        '
      additionalProperties: false
    ModelAlgorithm:
      enum:
      - catboost_classifier
      - catboost_regression
      - elasticnet_regression
      - gaussian_nb
      - kneighbors_classifier
      - lasso_regression
      - lasso
      - lgbm_classifier
      - lgbm_regression
      - linear_regression
      - logistic_regression
      - random_forest_classifier
      - random_forest_regression
      - sgd_regression
      - xgb_classifier
      - xgb_regression
      type: string
      description: 'Enumeration of model algorithms.


        A subset of these may be provided based on your ExperimentType. This is

        based on your target''s featureType. When you select a target feature

        from your ProfileInsights response (defaultVersionConfig.featuresList),

        it tells you which algorithms will be available if you choose it as your

        target.

        '
    FieldProfile:
      type: object
      properties:
        name:
          type: string
        tags:
          type: array
          items:
            type: string
        index:
          type: integer
        median:
          type: number
        average:
          type: number
        dataType:
          type: string
        kurtosis:
          type: number
        skewness:
          type: number
        fractiles:
          type: array
          items:
            type: number
        sampleValues:
          type: array
          items:
            type: string
        technicalName:
          type: string
        classification:
          $ref: '#/components/schemas/ProfileClassification'
        nullValueCount:
          type: integer
        textValueCount:
          type: integer
        zeroValueCount:
          type: integer
        maxNumericValue:
          type: number
        maxStringLength:
          type: integer
        minNumericValue:
          type: number
        minStringLength:
          type: integer
        sumStringLength:
          type: integer
        emptyStringCount:
          type: integer
        sumNumericValues:
          type: number
        numericValueCount:
          type: integer
        standardDeviation:
          type: number
        distinctValueCount:
          type: integer
        mostFrequentValues:
          type: array
          items:
            $ref: '#/components/schemas/ProfileFrequency'
        negativeValueCount:
          type: integer
        positiveValueCount:
          type: integer
        averageStringLength:
          type: number
        frequencyDistribution:
          type: array
          items:
            $ref: '#/components/schemas/BinnedFrequency'
        lastSortedStringValue:
          type: string
        firstSortedStringValue:
          type: string
        sumSquaredNumericValues:
          type: number
      description: Profile metadata for a single field/column.
      additionalProperties: false
    APIError:
      type: object
      required:
      - code
      - title
      properties:
        code:
          type: string
          description: Qlik error code (not HTTP response status code)
        meta:
          type: object
          properties:
            issue:
              type: string
              description: The issue code
            details:
              type: string
              description: Extra details for what may have caused the error
            errorId:
              type: string
              description: The unique id of the error instance
            argument:
              type: string
              description: The argument
            resource:
              type: string
              description: The resource type that the error occurred on
            resourceId:
              type: string
              description: The resource id that the error occurred on
          description: 'Additional details about the error. These may vary by error.

            '
          additionalProperties: false
        title:
          type: string
          description: Short summary of error
        detail:
          type: string
          description: Description of the error
      description: An error object
      additionalProperties: false
    OwnerId:
      type: string
      description: ID of owner/user for this entity
      x-qlik-filterable: true
    ExperimentMode:
      enum:
      - intelligent
      - manual
      - manual_hpo
      type: string
      default: intelligent
      description: The model training mode for the experiment version
    ProfileClassificationTag:
      type: object
      properties:
        tag:
          type: string
        score:
          type: number
      description: A classification tag with a confidence score.
      additionalProperties: false
    Failure:
      type: object
      required:
      - errors
      properties:
        errors:
          type: array
          items:
            $ref: '#/components/schemas/APIError'
          uniqueItems: true
        traceId:
          type: string
      additionalProperties: false
    DataType:
      enum:
      - DATE
      - TIME
      - DATETIME
      - TIMESTAMP
      - STRING
      - DOUBLE
      - DECIMAL
      - INTEGER
      - BOOLEAN
      - BINARY
      - CUSTOM
      - FLOAT
      - OBJECT
      type: string
      example: STRING
      description: 'The data type of this feature in your dataset

        '
    ProfileMetadata:
      type: object
      required:
      - dataSetId
      - status
      properties:
        status:
          enum:
          - FINISHED
          - PARTIAL
          - QUEUED
          - RUNNING
          - STALE
          - CANCELLED
          - FAILED
          - INCOMPLETE
          - DEFERRED
          type: string
        messages:
          type: array
          items:
            type: string
        dataSetId:
          $ref: '#/components/schemas/DataSetId'
        resultType:
          enum:
          - NORMAL
          - BASIC
          - BASIC_TOO_LARGE
          type: string
        connectionId:
          type: string
        lastLoadTime:
          type: string
        maxSizeBytes:
          type: number
        computationEndTime:
          type: string
        computationStartTime:
          type: string
      description: Metadata about a dataset profile computation.
      additionalProperties: false
    ProfileFrequency:
      type: object
      properties:
        value:
          type: string
        frequency:
          type: number
      description: A value and its frequency count.
      additionalProperties: false
    EntityId:
      type: string
      format: uuid
      description: ID of this entity
      x-qlik-filterable: true
    ProfileClassification:
      type: object
      properties:
        pii:
          type: boolean
        tags:
          type: array
          items:
            $ref: '#/components/schemas/ProfileClassificationTag'
        sensitive:
          type: boolean
        obfuscation:
          type: string
      description: Classification metadata for a field.
      additionalProperties: false
    Insights:
      enum:
      - constant
      - high_cardinality
      - high_cardinality_integer
      - too_many_nulls
      - will_be_impact_encoded
      - will_be_one_hot_encoded
      - possible_free_text_encoded
      - valid_index
      - underrepresented_class
      - invalid_column_name
      - will_be_date_engineered
      type: string
      description: 'Insights about a feature, such as why it''s not included in a model

        '
    ProfileInsightsInput:
      type: object
      properties:
        data:
          type: object
          properties:
            type:
              enum:
              - profile-insights
              type: string
              default: profile-insights
            attributes:
              type: object
              properties:
                target:
                  type: string
                  description: 'Optional selected target provided on subsequent requests

                    '
                include:
                  $ref: '#/components/schemas/IncludeProfileInsightsField'
                dataSetId:
                  $ref: '#/components/schemas/DataSetId'
                shouldWait:
                  type: boolean
                  default: false
                  description: 'Whether the server should or client should manage polling/waiting

                    '
                experimentType:
                  $ref: '#/components/schemas/ExperimentType'
              description: The request body for this resource
              additionalProperties: false
          description: Data wrapper for request input
          additionalProperties: false
      description: 'Input to get dataset and feature metadata needed to create

        experiment versions

        '
      additionalProperties: false
    ProfileInsightsGetResponse:
      type: object
      required:
      - data
      properties:
        data:
          type: object
          required:
          - type
          - id
          - attributes
          properties:
            id:
              $ref: '#/components/schemas/EntityId'
            type:
              enum:
              - profile-insights
              type: string
            attributes:
              $ref: '#/components/schemas/ProfileInsights'
          additionalProperties: false
      additionalProperties: false
    ExperimentType:
      enum:
      - binary
      - multiclass
      - regression
      type: string
      description: Experiment type
    ProfileInsights:
      type: object
      required:
      - tenantId
      - ownerId
      - status
      - dataSetId
      properties:
        status:
          enum:
          - pending
          - error
          - ready
          type: string
          default: pending
          description: 'Status of profile insights. Not available until `ready`.

            '
        ownerId:
          $ref: '#/components/schemas/OwnerId'
        insights:
          type: array
          items:
            $ref: '#/components/schemas/FeatureInsights'
          description: 'List of feature insights object, one per feature in the dataset

            '
        tenantId:
          $ref: '#/components/schemas/TenantId'
        algorithms:
          type: array
          items:
            $ref: '#/components/schemas/ModelAlgorithm'
          description: 'List of algorithms available for the selected experiment type

            '
        isLargeCsv:
          type: boolean
          description: Is this a CSV dataset > 1GB?
        sizeInBytes:
          type: integer
          description: Size of the profiled dataset in bytes.
        numberOfRows:
          type: number
          description: 'Number of rows in the dataset.


            When isLargeCsv is true, this is an estimate since their metadata is

            based on a sample rather than the full dataset.


            For datasets over 1GB, multiply `rows` (this) by `columns` (features

            included in experiment version) to calculate total `cells` to

            ensure it stays under your license limit. Large CSVs have a hard

            100M cell limit.

            '
        dataSetProfile:
          $ref: '#/components/schemas/DataSetProfile'
        experimentVersionId:
          type: string
          description: 'Optional experiment version ID. When included, it indicates that

            this dataset profile is a snapshot from a previous version.

            '
        defaultVersionConfig:
          type: object
          required:
          - name
          - datasetOrigin
          - dataSetId
          - experimentMode
          properties:
            name:
              type: string
              default: Experiment version name. Defaults to current date/time.
              example: '1999-12-31T23:59:59.123Z'
            dataSetId:
              $ref: '#/components/schemas/DataSetId'
            featuresList:
              $ref: '#/components/schemas/FeaturesList'
            datasetOrigin:
              $ref: '#/components/schemas/DatasetOrigin'
            experimentMode:
              $ref: '#/components/schemas/ExperimentMode'
          description: 'Default configuration for creating an experiment version from this

            dataset. Not returned when `experimentVersionId` is provided.

            '
      description: Insights (metadata) about an experiment dataset
      additionalProperties: false
    BinnedFrequency:
      type: object
      properties:
        binEdge:
          type: number
        frequency:
          type: number
      description: A frequency bin in a field's frequency distribution.
      additionalProperties: false
    FeatureType:
      enum:
      - categorical
      - numeric
      - date
      - freetext
      type: string
      description: 'The default feature type based on the feature''s data type.

        If you want a value to be interpreted differently (e.g. 0/1

        as categorical/boolean instead of numeric), use `changeType`.

        '
  responses:
    FailureDefault:
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/Failure'
      description: '`Unexpected Error`'
    500InternalError:
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/Failure'
      description: '`Internal Error`'
    503ServiceUnavailable:
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/Failure'
      description: '`Service Unavailable`'
    403Forbidden:
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/Failure'
      description: '`Forbidden`'
    502BadGateway:
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/Failure'
      description: '`Bad Gateway`'
    400BadRequest:
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/Failure'
      description: '`Bad Request`'
    404NotFound:
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/Failure'
      description: '`Not Found`'
    401Unauthorized:
      content:
        application/json:
          schema:
            $ref: '#/components/schemas/Failure'
      description: '`Unauthorized`'
  parameters:
    experimentType:
      in: query
      name: experimentType
      schema:
        $ref: '#/components/schemas/ExperimentType'
      required: false
      description: 'The optional experiment type for profile-insights GET requests after

        this is known.

        '
    target:
      in: query
      name: target
      schema:
        type: string
      required: false
      description: 'The optional target feature for profile-insights GET requests after this

        is known.

        '
    dataSetId:
      in: path
      name: dataSetId
      schema:
        $ref: '#/components/schemas/DataSetId'
      required: true
      description: The Qlik catalog dataset ID