Member Junction
    Preparing search index...

    Interface AIRerankRunResult

    The result of AIRerankerRunner.RunRerank. The runner never throws: every failure is Success: false with an ErrorMessage.

    interface AIRerankRunResult {
        DriverClass?: string;
        ErrorMessage?: string;
        ExecutionTimeMS: number;
        ModelID?: string;
        ModelName?: string;
        PromptRun?: MJAIPromptRunEntityExtended;
        PromptRunID?: string;
        Response?: RerankResponse;
        Success: boolean;
    }
    Index

    Properties

    DriverClass? ErrorMessage? ExecutionTimeMS ModelID? ModelName? PromptRun? PromptRunID? Response? Success

    Properties

    DriverClass?: string

    The driver class of that model.

    ErrorMessage?: string

    Why the rerank failed, when it did.

    ExecutionTimeMS: number

    Wall-clock time for the whole call, in milliseconds.

    ModelID?: string

    The model that answered after any failover, or the last one tried when the call failed.

    ModelName?: string

    The name of that model.

    PromptRun?: MJAIPromptRunEntityExtended

    The run row itself, as the runner holds it. Its TotalCost and token rollups include the prompt runs of a prompt-backed driver once its queued finalize has run, so an agent step can carry it as its PromptRun and the agent run counts the rerank. Absent when PromptRunID is.

    PromptRunID?: string

    The MJ: AI Prompt Runs row for this call. Absent when the call never reached a model.

    Response?: RerankResponse

    The driver's response, ranked by relevance. Absent when no model was called.

    Success: boolean

    Whether the documents were reranked.

    Settings

    Member Visibility

    On This Page

    Properties