export type MLRuntime = 'native' | 'web-worker' | 'wasm';
export interface ModelConfig {
    id: string;
    name: string;
    type: 'regression' | 'classification' | 'anomaly' | 'forecasting' | 'neural-network' | 'linear-regression' | 'signal-processor' | string;
    metadata?: Record<string, any>;
}
export interface PredictionInput {
    data: number[] | Float32Array;
    metadata?: Record<string, any>;
}
export interface PredictionResult {
    modelId: string;
    output: Float32Array;
    xValues?: Float32Array | Float64Array;
    outputShape: number[];
    timestamp: number;
    processingTime: number;
    confidence?: number[];
    metadata?: Record<string, any>;
}
export interface MLModelAPI {
    id: string;
    name: string;
    type: string;
    predict(input: PredictionInput): Promise<PredictionResult>;
    getInfo(): ModelConfig;
    isReady(): boolean;
    warmup(): Promise<void>;
    dispose(): void;
}
export interface VisualizationConfig {
    showConfidenceInterval?: boolean;
    intervalOpacity?: number;
    lineStyle?: {
        color?: string;
        width?: number;
        dash?: number[];
    };
}
export interface PluginMLIntegrationConfig {
    runtime?: MLRuntime;
    models?: ModelConfig[];
    defaultVisualization?: VisualizationConfig;
}
export interface TrainingResult {
    coefficients: number[];
    intercept: number;
    fitted: number[];
    residuals: number[];
    r2: number;
    rmse: number;
}
export interface MLIntegrationAPI {
    registerModel(model: MLModelAPI): void;
    runInference(modelId: string, seriesId: string): Promise<PredictionResult>;
    visualizeResults(result: PredictionResult, config?: VisualizationConfig): string;
    /** Intent-revealing alias for visualizeResults (prediction overlay). */
    visualizePredictions?(result: PredictionResult, config?: VisualizationConfig): string;
    clearResults(visualizationId?: string): void;
    /** Train a small regression model on the fly, returning fit diagnostics. */
    trainModel?(modelId: string, data: {
        x: number[][];
        y: number[];
    }): TrainingResult;
    /** Retrieve the last training diagnostics for a model. */
    getTrainingResult?(modelId: string): TrainingResult | null;
    /** List registered model descriptors. */
    listModels?(): ModelConfig[];
    stats: NativeStatsAPI;
}
export interface NativeStatsAPI {
    fft(data: number[]): {
        real: number[];
        imag: number[];
    };
    mean(data: number[]): number;
    standardDeviation(data: number[]): number;
    correlation(x: number[], y: number[]): number;
}
export interface ModelLoadedEvent {
    modelId: string;
    config: ModelConfig;
}
export interface PredictionEvent {
    modelId: string;
    input: PredictionInput;
    result: PredictionResult;
}
export interface ModelErrorEvent {
    modelId: string;
    error: string;
}