# Generated by API Evangelist (build-phrasing.py). Our phrasing, not observed demand.
overlay: 1.0.0
info:
  title: API Evangelist conversational phrasing for Microsoft Azure Anomaly Detector Multivariate API
  version: 1.0.0
extends: openapi/microsoft-azure-multivariate-api-openapi.yml
actions:
- target: $.info
  update:
    x-apievangelist-phrasing:
      method: generated
      generated: '2026-09-26'
      generator: build-phrasing.py
      label: Generated by API Evangelist
      operations: 7
- target: $.paths['/multivariate/detect-batch/{resultId}'].get
  update:
    x-apievangelist-phrasing:
      intent: Get a batch anomaly detection result
      effect: read
      questions:
      - How do I get the results of an asynchronous multivariate anomaly detection?
      - Is my batch detection result ready yet?
      instructions:
      - text: Get batch detection result {result}.
        slots:
          result: path.resultId
      - text: Show the anomalies found in detection result {result}.
        slots:
          result: path.resultId
      method: generated
      generated: '2026-09-26'
- target: $.paths['/multivariate/models'].get
  update:
    x-apievangelist-phrasing:
      intent: List multivariate anomaly models
      effect: read
      questions:
      - Which multivariate anomaly detection models have I trained?
      - Can I page through my anomaly detector models?
      instructions:
      - text: List my multivariate anomaly models.
      - text: Show {top} anomaly models, skipping the first {skip}.
        slots:
          top: query.top
          skip: query.skip
      method: generated
      generated: '2026-09-26'
- target: $.paths['/multivariate/models'].post
  update:
    x-apievangelist-phrasing:
      intent: Train a multivariate anomaly model
      effect: write
      questions:
      - How do I train a multivariate anomaly detection model from data in Blob Storage?
      - Can I set the sliding window when training an anomaly model?
      instructions:
      - text: Train an anomaly model on {source} from {start} to {end}.
        slots:
          source: requestBody.dataSource
          start: requestBody.startTime
          end: requestBody.endTime
      - text: Train model {name} on {source} between {start} and {end} with sliding window {window}.
        slots:
          name: requestBody.displayName
          source: requestBody.dataSource
          start: requestBody.startTime
          end: requestBody.endTime
          window: requestBody.slidingWindow
      method: generated
      generated: '2026-09-26'
- target: $.paths['/multivariate/models/{modelId}'].get
  update:
    x-apievangelist-phrasing:
      intent: Get a multivariate anomaly model
      effect: read
      questions:
      - Has my multivariate anomaly model finished training?
      - Which variables does a trained anomaly model use?
      instructions:
      - text: Get multivariate model {model}.
        slots:
          model: path.modelId
      - text: Show the training status and variables of model {model}.
        slots:
          model: path.modelId
      method: generated
      generated: '2026-09-26'
- target: $.paths['/multivariate/models/{modelId}'].delete
  update:
    x-apievangelist-phrasing:
      intent: Delete a multivariate anomaly model
      effect: destructive
      questions:
      - Can I delete an anomaly detection model I no longer need?
      - What happens to a model's detection results after I delete it?
      instructions:
      - text: Delete multivariate model {model}.
        slots:
          model: path.modelId
      - text: Remove anomaly model {model} permanently.
        slots:
          model: path.modelId
      method: generated
      generated: '2026-09-26'
- target: $.paths['/multivariate/models/{modelId}:detect-batch'].post
  update:
    x-apievangelist-phrasing:
      intent: Run batch multivariate anomaly detection
      effect: write
      questions:
      - How do I detect anomalies across a whole time range of stored data?
      - Can I ask for the top contributing variables per anomaly in a batch run?
      instructions:
      - text: Run batch detection with model {model} on {source} from {start} to {end}.
        slots:
          model: path.modelId
          source: requestBody.dataSource
          start: requestBody.startTime
          end: requestBody.endTime
      - text: Detect batch anomalies using model {model} on {source}, {start} to {end}, top {count} contributors.
        slots:
          model: path.modelId
          source: requestBody.dataSource
          start: requestBody.startTime
          end: requestBody.endTime
          count: requestBody.topContributorCount
      method: generated
      generated: '2026-09-26'
- target: $.paths['/multivariate/models/{modelId}:detect-last'].post
  update:
    x-apievangelist-phrasing:
      intent: Detect anomalies in the latest data point
      effect: read
      questions:
      - Can I check whether my newest data point is anomalous right away?
      - How do I get synchronous anomaly detection with inline variables?
      instructions:
      - text: Check the last point of {variables} for anomalies with model {model}.
        slots:
          variables: requestBody.variables
          model: path.modelId
      - text: Score the latest timestamp with model {model} and return the top {count} contributors.
        slots:
          model: path.modelId
          count: requestBody.topContributorCount
      method: generated
      generated: '2026-09-26'