/**
 * Text Chunker for Gemini RAG Service
 *
 * Handles text chunking with configurable token limits,
 * overlap calculation, and line number tracking.
 */
/**
 * Configuration options for text chunking
 */
export interface ChunkOptions {
    /** Maximum number of tokens per chunk (default: 512) */
    maxTokens?: number;
    /** Number of overlapping tokens between chunks (default: 50) */
    overlapTokens?: number;
    /** Approximate characters per token (default: 4) */
    charsPerToken?: number;
    /** Preserve line boundaries when possible (default: true) */
    preserveLines?: boolean;
    /** Minimum chunk size in tokens (default: 100) */
    minChunkTokens?: number;
}
/**
 * Represents a single chunk of text with metadata
 */
export interface TextChunk {
    /** The chunk content */
    content: string;
    /** Zero-based chunk index */
    index: number;
    /** Starting line number (1-based) */
    startLine: number;
    /** Ending line number (1-based, inclusive) */
    endLine: number;
    /** Starting character offset in original text */
    startOffset: number;
    /** Ending character offset in original text */
    endOffset: number;
    /** Estimated token count */
    tokenCount: number;
    /** Whether this chunk overlaps with the previous */
    hasOverlap: boolean;
}
/**
 * Result of chunking operation
 */
export interface ChunkResult {
    /** Array of text chunks */
    chunks: TextChunk[];
    /** Total number of chunks */
    totalChunks: number;
    /** Total estimated tokens */
    totalTokens: number;
    /** Original text length */
    originalLength: number;
    /** Total lines in original text */
    totalLines: number;
}
/**
 * Text Chunker class for splitting text into manageable chunks
 */
export declare class TextChunker {
    private readonly options;
    constructor(options?: ChunkOptions);
    /**
     * Estimate token count for a string
     */
    estimateTokens(text: string): number;
    /**
     * Convert token count to approximate character count
     */
    private tokensToChars;
    /**
     * Get line number for a character offset
     */
    private getLineNumber;
    /**
     * Find the nearest line boundary for a position
     */
    private findLineBoundary;
    /**
     * Find a good split point near the target position
     */
    private findSplitPoint;
    /**
     * Chunk text into smaller pieces
     */
    chunk(text: string): ChunkResult;
    /**
     * Chunk text by semantic sections (code-aware)
     */
    chunkBySection(text: string, sectionDelimiters?: RegExp[]): ChunkResult;
}
/**
 * Create a chunker with default options
 */
export declare function createChunker(options?: ChunkOptions): TextChunker;
/**
 * Convenience function to chunk text with default settings
 */
export declare function chunkText(text: string, options?: ChunkOptions): ChunkResult;
/**
 * Convenience function to chunk text by semantic sections
 */
export declare function chunkTextBySection(text: string, options?: ChunkOptions): ChunkResult;
//# sourceMappingURL=chunker.d.ts.map