Datalayer VS Code Extension - v0.0.17
    Preparing search index...

    All validated Datalayer extension settings.

    interface DatalayerSettings {
        agentChat: {
            agentSpecId: string;
            protocol: "ag-ui" | "acp" | "a2a" | "vercel-ai" | "vercel-ai-jupyter";
        };
        autoConnect: { strategies: ("Active Runtime" | "Ask" | "Pyodide")[] };
        inlineLlmCompletion: {
            contextBlocks: number;
            debounceMs: number;
            enabled: boolean;
            triggerMode: "auto" | "manual";
        };
        logging: {
            enableDatalayerLogging: boolean;
            enablePerformanceMonitoring: boolean;
            includeContext: boolean;
            includeTimestamps: boolean;
            level: "error"
            | "trace"
            | "debug"
            | "info"
            | "warn";
        };
        onboarding: { showWelcome: boolean };
        proseLlmCompletion: {
            contextBlocks: number;
            debounceMs: number;
            enabled: boolean;
            triggerKey: string;
            triggerMode: "auto" | "manual";
        };
        pyodide: {
            preloadBehavior: "auto"
            | "ask-once"
            | "ask-always"
            | "disabled";
            preloadPackages: string[];
            version: string;
        };
        runtime: { defaultMinutes: number; defaultType: "CPU"
        | "GPU" };
        services: {
            iamUrl: string;
            runtimesUrl: string;
            spacerUrl: string;
            spacerWsUrl: string;
        };
        tools: { responseFormat: "json"
        | "toon" };
    }
    Index

    Properties

    agentChat: {
        agentSpecId: string;
        protocol: "ag-ui" | "acp" | "a2a" | "vercel-ai" | "vercel-ai-jupyter";
    }

    Agent chat sidebar configuration.

    Type Declaration

    • agentSpecId: string

      Agent specification ID used when the sidebar provisions a fresh runtime.

    • protocol: "ag-ui" | "acp" | "a2a" | "vercel-ai" | "vercel-ai-jupyter"

      Chat transport protocol. Default kept in sync with the datalayer.agentChat.protocol contribution in package.json so the Zod fallback matches what VS Code reports when the user has not explicitly set the value.

    autoConnect: { strategies: ("Active Runtime" | "Ask" | "Pyodide")[] }

    Auto-connect strategies.

    Type Declaration

    • strategies: ("Active Runtime" | "Ask" | "Pyodide")[]

      Ordered list of auto-connect strategies.

    inlineLlmCompletion: {
        contextBlocks: number;
        debounceMs: number;
        enabled: boolean;
        triggerMode: "auto" | "manual";
    }

    Inline LLM completion configuration.

    Type Declaration

    • contextBlocks: number

      Number of context blocks (-1 for entire document).

    • debounceMs: number

      Debounce delay in milliseconds.

    • enabled: boolean

      Toggle LLM-powered inline code completions in Jupyter cells.

    • triggerMode: "auto" | "manual"

      How code completions are triggered (auto while typing or manual shortcut).

    logging: {
        enableDatalayerLogging: boolean;
        enablePerformanceMonitoring: boolean;
        includeContext: boolean;
        includeTimestamps: boolean;
        level: "error" | "trace" | "debug" | "info" | "warn";
    }

    Logging configuration.

    Type Declaration

    • enableDatalayerLogging: boolean

      Enable Datalayer operation logging.

    • enablePerformanceMonitoring: boolean

      Toggle performance monitoring with timing and memory tracking.

    • includeContext: boolean

      Include context information in log messages.

    • includeTimestamps: boolean

      Include timestamps in log messages.

    • level: "error" | "trace" | "debug" | "info" | "warn"

      Minimum log level.

    onboarding: { showWelcome: boolean }

    Onboarding preferences.

    Type Declaration

    • showWelcome: boolean

      Whether to show the welcome prompt.

    proseLlmCompletion: {
        contextBlocks: number;
        debounceMs: number;
        enabled: boolean;
        triggerKey: string;
        triggerMode: "auto" | "manual";
    }

    Prose LLM completion configuration.

    Type Declaration

    • contextBlocks: number

      Number of context blocks (-1 for entire document).

    • debounceMs: number

      Debounce delay in milliseconds.

    • enabled: boolean

      Toggle LLM-powered inline completions for prose and markdown content.

    • triggerKey: string

      Keyboard shortcut for manual trigger.

    • triggerMode: "auto" | "manual"

      How prose completions are triggered (auto while typing or manual shortcut).

    pyodide: {
        preloadBehavior: "auto" | "ask-once" | "ask-always" | "disabled";
        preloadPackages: string[];
        version: string;
    }

    Pyodide configuration.

    Type Declaration

    • preloadBehavior: "auto" | "ask-once" | "ask-always" | "disabled"

      Preload behavior for Pyodide packages.

    • preloadPackages: string[]

      Python packages to preload.

    • version: string

      Pyodide version string.

    runtime: { defaultMinutes: number; defaultType: "CPU" | "GPU" }

    Runtime configuration.

    Type Declaration

    • defaultMinutes: number

      Default runtime duration in minutes.

    • defaultType: "CPU" | "GPU"

      Default runtime type.

    services: {
        iamUrl: string;
        runtimesUrl: string;
        spacerUrl: string;
        spacerWsUrl: string;
    }

    Service endpoint URLs.

    Type Declaration

    • iamUrl: string

      IAM service URL.

    • runtimesUrl: string

      Runtimes service URL.

    • spacerUrl: string

      Spacer service URL.

    • spacerWsUrl: string

      WebSocket URL for Spacer real-time collaboration.

    tools: { responseFormat: "json" | "toon" }

    MCP tool response format.

    Type Declaration

    • responseFormat: "json" | "toon"

      Response format for MCP tool responses.