/**
 * EmbeddingService handles communication with Ollama API for generating embeddings.
 * Includes caching, error handling, and retry logic.
 */
export declare class EmbeddingService {
    private client;
    private model;
    private cache;
    private maxRetries;
    private retryDelay;
    constructor(ollamaUrl?: string, model?: string, maxRetries?: number, retryDelay?: number);
    /**
     * Generate an embedding vector for the given text.
     * Uses cache if available, otherwise calls Ollama API.
     *
     * @param text - Text to generate embedding for
     * @param useCache - Whether to use cached embeddings (default true)
     * @returns Embedding vector as array of numbers
     */
    generateEmbedding(text: string, useCache?: boolean): Promise<number[]>;
    /**
     * Generate embeddings for multiple texts in batch.
     *
     * @param texts - Array of texts to generate embeddings for
     * @param useCache - Whether to use cached embeddings (default true)
     * @returns Array of embedding vectors
     */
    generateEmbeddings(texts: string[], useCache?: boolean): Promise<number[][]>;
    /**
     * Call Ollama API to generate embedding.
     *
     * @param text - Text to generate embedding for
     * @returns Embedding vector
     */
    private callOllamaAPI;
    /**
     * Generate a cache key from text.
     * Uses a simple hash to avoid storing full text as keys.
     */
    private getCacheKey;
    /**
     * Sleep for specified milliseconds.
     */
    private sleep;
    /**
     * Clear the embedding cache.
     */
    clearCache(): void;
    /**
     * Get cache statistics.
     */
    getCacheStats(): {
        size: number;
        maxSize: number;
    };
    /**
     * Check if Ollama is available and the model is loaded.
     */
    healthCheck(): Promise<{
        available: boolean;
        model: string;
        error?: string;
    }>;
    /**
     * Calculate cosine similarity between two embedding vectors.
     *
     * @param vecA - First embedding vector
     * @param vecB - Second embedding vector
     * @returns Similarity score between 0 and 1
     */
    static cosineSimilarity(vecA: number[], vecB: number[]): number;
}
//# sourceMappingURL=EmbeddingService.d.ts.map