/**
 * PDF Clean Composer
 *
 * Filters PDF content to include only main content area elements,
 * excluding headers, footers, page numbers, and cleaning up text elements.
 *
 * Main responsibilities:
 * - Detect content area boundaries for each page
 * - Filter elements within content area
 * - Clean text elements (fix spacing, validate characters)
 * - Remove non-content elements (headers, footers, page numbers)
 * - Validate and clean image elements
 */
import type { PdfPageContent } from '../models/PdfPageContent.js';
import type { PdfDocument } from './PdfDocument.js';
import type { PdfPageRenderer } from '../types/renderer.types.js';
export interface PdfCleanComposerOptions {
    /**
     * Margin from top to exclude headers (as percentage of page height)
     * Default: 0.1 (10%)
     */
    topMarginPercent?: number;
    /**
     * Margin from bottom to exclude footers (as percentage of page height)
     * Default: 0.1 (10%)
     */
    bottomMarginPercent?: number;
    /**
     * Margin from left and right to exclude side elements (as percentage of page width)
     * Default: 0.05 (5%)
     */
    sideMarginPercent?: number;
    /**
     * Minimum height for text elements (in points)
     * Elements smaller than this will be filtered out
     * Default: 8
     */
    minTextHeight?: number;
    /**
     * Minimum width for text elements (in points)
     * Elements smaller than this will be filtered out
     * Default: 10
     */
    minTextWidth?: number;
    /**
     * Maximum allowed spacing between words (as ratio of font size)
     * Text with excessive spacing will be cleaned
     * Default: 3.0
     */
    maxWordSpacingRatio?: number;
    /**
     * Remove elements with non-printable or control characters
     * Default: true
     */
    removeControlCharacters?: boolean;
    /**
     * Minimum meaningful text length
     * Text shorter than this will be filtered out
     * Default: 3
     */
    minTextLength?: number;
    /**
     * Remove isolated single characters or symbols
     * Default: true
     */
    removeIsolatedCharacters?: boolean;
    /**
     * Minimum width for image elements (in points/pixels)
     * Images smaller than this will be filtered out as decorative elements
     * Default: 50
     */
    minImageWidth?: number;
    /**
     * Minimum height for image elements (in points/pixels)
     * Images smaller than this will be filtered out as decorative elements
     * Default: 50
     */
    minImageHeight?: number;
    /**
     * Minimum area for image elements (width × height)
     * Images with smaller area will be filtered out
     * Default: 2500 (50×50)
     */
    minImageArea?: number;
    /**
     * Enable cover page detection and screenshot generation
     * If the first page is detected as a cover (full-page image), generate a screenshot instead
     * Default: true
     */
    coverPageDetection?: boolean;
    /**
     * Cover page threshold (percentage of page area that an image must cover)
     * Used to determine if a page is a cover page
     * Default: 0.8 (80% of page area)
     */
    coverPageThreshold?: number;
    /**
     * Screenshot quality for cover pages (1-100)
     * Default: 95
     */
    coverPageScreenshotQuality?: number;
    /**
     * Output directory path for cleaning image files
     * If provided, removed image files will be deleted from disk
     */
    outputDir?: string;
    /**
     * Pluggable page renderer. When set (e.g. PuppeteerRenderer), cleanComposer
     * rasterizes cover/page screenshots through it (Chromium) instead of
     * node-canvas. When null/undefined, the node-canvas PageRenderer path is used.
     * Mirrors how PdfDecomposer.screenshot() picks its rasterization path.
     */
    renderer?: PdfPageRenderer | null;
    /**
     * Target width (px) for the page/cover screenshot when rendering via `renderer`.
     * Default: 1024. Ignored by the node-canvas fallback (which uses scale 1.0).
     */
    coverPageScreenshotWidth?: number;
}
export declare class PdfCleanComposer {
    /**
     * Clean and filter PDF pages to include only main content
     *
     * @param pages Array of PDF page content to clean
     * @param options Cleaning configuration options
     * @param pdfDocument Optional PDF document for cover page detection and screenshot
     * @returns Cleaned array of PDF page content
     */
    static cleanPages(pages: PdfPageContent[], options?: PdfCleanComposerOptions, pdfDocument?: PdfDocument): Promise<PdfPageContent[]>;
    /**
     * Clean a single PDF page
     * @param skipScreenshot If true, skip screenshot conversion (used when limit reached)
     */
    private static cleanPage;
    /**
     * Calculate content area boundaries for a page
     */
    private static calculateContentArea;
    /**
     * Clean and filter elements based on content area and quality
     */
    private static cleanElements;
    /**
     * Check if element is within content area boundaries
     */
    private static isElementInContentArea;
    /**
     * Clean individual element based on its type
     */
    private static cleanElement;
    /**
     * Clean text element content and validate dimensions
     */
    private static cleanTextElement;
    /**
     * Clean image element and validate dimensions
     */
    private static cleanImageElement;
    /**
     * Remove control characters and non-printable characters
     */
    private static removeControlCharacters;
    /**
     * Fix excessive spacing in text
     */
    private static fixTextSpacing;
    /**
     * Check if text is an isolated character or symbol
     */
    private static isIsolatedCharacter;
    /**
     * Validate text element dimensions
     */
    private static validateTextElementDimensions;
    /**
     * Check if element is a text element
     */
    private static isTextElement;
    /**
     * Check if element is an image element
     */
    private static isImageElement;
    /**
     * Normalize bounding box to consistent format
     */
    private static normalizeBoundingBox;
    /**
     * Check if element was modified during cleaning
     */
    private static isElementModified;
    /**
     * Remove image file from output directory when element is filtered out
     * Only works in Node.js environment - gracefully degrades in browser
     */
    private static removeImageFile;
    /**
     * Detect if the page is a cover page and process it as screenshot
     * Cover page is detected by having a large image that covers most of the page area
     * OR multiple images that collectively cover most of the page (for tiled cover pages)
     */
    private static detectAndProcessCoverPage;
    /**
     * Determine if a page should be converted to screenshot based on its content
     */
    private static shouldConvertToScreenshot;
    /**
     * Calculate how well images are distributed across the page
     * Returns distribution score (0-1) indicating coverage of page dimensions
     */
    private static calculateImageDistribution;
    /**
     * Generate screenshot for cover page
     * Note: Uses scale 1.0 for memory efficiency on large documents
     */
    private static generatePageScreenshot;
}
//# sourceMappingURL=PdfCleanComposer.d.ts.map