Member Junction
    Preparing search index...

    Class PromptEngine

    Index

    Constructors

    constructor

    Methods

    executePrompt executePromptsParallel initialize setGuardrailCheck

    Constructors

    Methods

    • Execute a prompt with AI/Core Main entry point for DBAutoDoc analysis

      Type Parameters

      • T

      Parameters

      • promptName: string
      • context: any
      • Optionaloptions: {
            effortLevelOverride?: number;
            maxTokens?: number;
            modelOverride?: string;
            responseFormat?: "JSON" | "Text";
            systemPrompt?: string;
            temperature?: number;
        }
        • OptionaleffortLevelOverride?: number

          Override effort level for this specific call

        • OptionalmaxTokens?: number
        • OptionalmodelOverride?: string

          Override the model for this specific call (e.g., use a stronger model for FK pruning)

        • OptionalresponseFormat?: "JSON" | "Text"
        • OptionalsystemPrompt?: string
        • Optionaltemperature?: number

      Returns Promise<PromptExecutionResult<T>>

    • Execute multiple prompts in parallel Uses AI/Core's ChatCompletions for efficiency

      Type Parameters

      • T

      Parameters

      • requests: {
            context: any;
            options?: {
                maxTokens?: number;
                responseFormat?: "JSON" | "Text";
                systemPrompt?: string;
                temperature?: number;
            };
            promptName: string;
        }[]

      Returns Promise<PromptExecutionResult<T>[]>

    • Set guardrail checking callback This will be called before every LLM execution to check if we should continue

      Parameters

      • checkFn: GuardrailCheckFn

      Returns void

    Settings

    Member Visibility

    On This Page

    Constructors
    Methods