import type { ChunkOptions, ChunkedDocument } from './types';
import type { ExtractedContent } from '../extraction/types';
/**
 * Hybrid chunking engine that splits documents into optimal chunks for RAG embeddings
 */
export declare class Chunker {
    private options;
    constructor(options: ChunkOptions);
    /**
     * Main entry point - dispatches to appropriate chunking strategy based on content type
     */
    chunkDocument(documentId: string, content: ExtractedContent): ChunkedDocument;
    /**
     * Semantic chunking by headers - preserves document structure
     */
    private chunkMarkdown;
    /**
     * Semantic chunking by functions/classes - preserves code structure
     */
    private chunkCode;
    /**
     * Fixed-size chunking with sliding window overlap
     */
    private chunkPlainText;
    /**
     * Helper to split text that exceeds maxTokens while preserving word boundaries
     */
    private splitLargeText;
    /**
     * Rough token estimation (~1.3 tokens per word, ~4 chars per token)
     */
    private estimateTokens;
}
//# sourceMappingURL=chunker.d.ts.map