/**
 * Neo4j GraphRAG Storage
 *
 * Extends Neo4jStorage with vector embedding and semantic search capabilities
 * using Neo4j 5.x's native vector index support.
 *
 * @packageDocumentation
 */
import type { SkillId, GraphRAGStorage, EmbeddingAdapter, SkillWithEmbedding, VectorSearchResult, VectorSearchOptions, GraphRAGSearchResult, GraphRAGSearchOptions, GraphRAGContext, LearningPathRAGRequest, LearningPathRAGResult, SimilarityMetric } from '../types/index.js';
import { Neo4jStorage } from './neo4j.js';
/**
 * Default GraphRAG configuration
 */
export declare const GRAPHRAG_DEFAULTS: {
    /** Default limit for vector search results */
    searchLimit: number;
    /** Default minimum similarity score */
    minScore: number;
    /** Default graph traversal depth */
    graphDepth: number;
    /** Default weight for vector similarity in combined score */
    vectorWeight: number;
    /** Default similarity metric */
    metric: SimilarityMetric;
    /** Batch size for embedding operations */
    embeddingBatchSize: number;
};
/**
 * Neo4j GraphRAG Storage
 *
 * Provides vector embedding storage and semantic search using Neo4j 5.x
 * native vector indexes combined with graph traversal for enhanced context.
 *
 * @example
 * ```typescript
 * import { Neo4jGraphRAGStorage, createOpenAIEmbeddingAdapter } from 'learngraph';
 *
 * const storage = new Neo4jGraphRAGStorage();
 * await storage.connect({
 *   backend: 'neo4j',
 *   uri: 'bolt://localhost:7687',
 *   username: 'neo4j',
 *   password: 'password',
 * });
 *
 * // Set embedding adapter
 * const embedding = createOpenAIEmbeddingAdapter();
 * storage.setEmbeddingAdapter(embedding);
 *
 * // Create vector index
 * await storage.createVectorIndex({ dimensions: 1536 });
 *
 * // Embed all skills
 * await storage.embedAllSkills();
 *
 * // Semantic search with graph context
 * const results = await storage.searchWithGraphContext('learn JavaScript functions');
 * ```
 */
export declare class Neo4jGraphRAGStorage extends Neo4jStorage implements GraphRAGStorage {
    private embeddingAdapter;
    private vectorIndexName;
    private vectorIndexDimensions;
    setEmbeddingAdapter(adapter: EmbeddingAdapter): void;
    getEmbeddingAdapter(): EmbeddingAdapter | null;
    private requireEmbeddingAdapter;
    /**
     * Generate text representation for embedding
     */
    private getEmbeddingText;
    embedSkill(skillId: SkillId): Promise<void>;
    embedAllSkills(options?: {
        batchSize?: number;
        onProgress?: (completed: number, total: number) => void;
    }): Promise<{
        embedded: number;
        failed: number;
    }>;
    hasEmbedding(skillId: SkillId): Promise<boolean>;
    getSkillWithEmbedding(skillId: SkillId): Promise<SkillWithEmbedding | null>;
    searchByVector(query: string, options?: VectorSearchOptions): Promise<VectorSearchResult[]>;
    searchByEmbedding(embedding: number[], options?: VectorSearchOptions): Promise<VectorSearchResult[]>;
    findSimilarSkills(skillId: SkillId, options?: VectorSearchOptions): Promise<VectorSearchResult[]>;
    searchWithGraphContext(query: string, options?: GraphRAGSearchOptions): Promise<GraphRAGSearchResult[]>;
    buildRAGContext(query: string, options?: GraphRAGSearchOptions): Promise<GraphRAGContext>;
    private formatContextForLLM;
    generateLearningPathRAG(request: LearningPathRAGRequest): Promise<LearningPathRAGResult>;
    createVectorIndex(options?: {
        dimensions?: number;
        metric?: SimilarityMetric;
    }): Promise<void>;
    dropVectorIndex(): Promise<void>;
    hasVectorIndex(): Promise<boolean>;
    getVectorIndexStats(): Promise<{
        exists: boolean;
        dimensions?: number;
        metric?: SimilarityMetric;
        indexedCount: number;
        totalSkills: number;
    }>;
}
/**
 * Create a Neo4j GraphRAG storage instance
 */
export declare function createNeo4jGraphRAGStorage(): Neo4jGraphRAGStorage;
//# sourceMappingURL=neo4j-graphrag.d.ts.map