/**
 * Text reconstruction from positioned text fragments.
 *
 * Assembles raw text fragments extracted from PDF content streams into
 * coherent, human-readable text with proper reading order, line breaks,
 * and paragraph detection.
 *
 * Challenges addressed:
 * - PDF text has no semantic structure (only "draw char at (x,y)")
 * - Text fragments may be out of order
 * - Word and line boundaries must be inferred from positions
 * - Columns and tables need proper handling
 * - Different fonts/sizes affect spacing thresholds
 * - Multi-column layouts need column detection
 * - RTL (Arabic, Hebrew) text needs right-to-left sorting
 * - Vertical CJK text needs column-based grouping
 *
 * @see PDF Reference 1.7, Chapter 5 - Text
 */
import type { TextFragment } from "./content-interpreter.js";
/**
 * Reconstruct readable text from positioned text fragments.
 *
 * @param fragments - Raw text fragments with positions from content stream
 * @returns Reconstructed text with proper line breaks and spacing
 */
export declare function reconstructText(fragments: TextFragment[]): string;
/**
 * Detailed text extraction result preserving position information.
 */
export interface TextLine {
    /** The text content of this line */
    text: string;
    /** Y position (PDF coordinate, origin = bottom-left) */
    y: number;
    /** X position of the start of the line */
    x: number;
    /** Font size of the first fragment */
    fontSize: number;
}
/**
 * Extract text as structured lines.
 */
export declare function reconstructTextLines(fragments: TextFragment[]): TextLine[];