Circle V2 API Docs
    Preparing search index...

    Type Alias RunOptions

    type RunOptions = {
        auditQuestion: RunnableAuditQuestion;
        context?: QuestionRunContext;
        dateIntent?: DateIntent;
        expandedQuestion?: string;
        model?: ChatOpenAI;
        patient: PatientRow;
        relevantDocuments?: FindRelevantDocumentsResult;
        reviewCriteriaInstructions?: string | null;
        summaryInstructions?: string | null;
    }
    Index

    Properties

    auditQuestion context? dateIntent? expandedQuestion? model? patient relevantDocuments? reviewCriteriaInstructions? summaryInstructions?

    Properties

    auditQuestion: RunnableAuditQuestion
    context?: QuestionRunContext
    dateIntent?: DateIntent

    Pre-classified date intent, hoisted alongside expandedQuestion by SummaryTableShortcutRunner so classification runs once per question. Absent when a runner computes it itself, or when intent-aware planning is off.

    expandedQuestion?: string

    Pre-resolved question expansion + document selection, passed down by SummaryTableShortcutRunner so the expensive expand + find_documents LLM calls run once per question instead of once per runner. Absent when a runner is used standalone (eval harness, shortcut-disabled jobbables) — the runner then computes them itself.

    model?: ChatOpenAI
    patient: PatientRow
    relevantDocuments?: FindRelevantDocumentsResult
    reviewCriteriaInstructions?: string | null

    Template custom_review_criteria_prompt: rubric that guides retrieval and pass/fail grading.

    summaryInstructions?: string | null

    Template custom_summary_prompt: format for the summary deliverable.

    Settings

    Member Visibility

    On This Page

    Properties