afpp/dist/core.d.ts

Async Fast PDF Parser for Node.js — dependency-light, TypeScript-first, production-ready.

import { Canvas, CanvasRenderingContext2D } from '@napi-rs/canvas';
export declare enum PROCESSING_TYPE {
    IMAGE = "IMAGE",
    MIXED = "MIXED",
    TEXT = "TEXT"
}
export interface AfppParseOptions {
    /**
     * Concurrency level for page processing. Defaults to 1.
     * Higher values may improve performance but increase memory usage.
     * Set to 'auto' to use the number of available CPU cores (capped at 8).
     * @default 1
     */
    concurrency?: number | 'auto';
    /**
     * Image encoding format when rendering non-text pages. Defaults to 'png'.
     * Supported formats: 'avif', 'jpeg', 'png', 'webp'.
     * @default 'png'
     */
    imageEncoding?: ImageEncoding;
    /**
     * Password for encrypted pdf files.
     */
    password?: string;
    /**
     * Scale factor for image rendering. Defaults to 1.0.
     * - 1.0: Standard quality (72 DPI equivalent)
     * - 2.0: High quality (144 DPI equivalent, 4x memory)
     * - 3.0: Print quality (216 DPI equivalent, 9x memory)
     * @default 1.0
     */
    scale?: number;
}
export interface CanvasAndContext {
    canvas: Canvas;
    context: CanvasRenderingContext2D;
}
export type ImageEncoding = 'avif' | 'jpeg' | 'png' | 'webp';
export type PageProcessor<T> = (content: Buffer | string, pageNumber: number, pageCount: number) => Promise<T> | T;
export interface PdfCanvasFactory {
    create(width: number, height: number): CanvasAndContext;
    destroy(canvasAndContext: CanvasAndContext): void;
    reset(canvasAndContext: CanvasAndContext, width: number, height: number): void;
}
/**
 * Result yielded by streaming PDF parser
 */
export interface StreamingResult<T> {
    data: T;
    pageCount: number;
    pageNumber: number;
}
export declare function parsePdfFile(type: PROCESSING_TYPE.IMAGE, input: Buffer | string | Uint8Array | URL, options?: AfppParseOptions, callback?: undefined): Promise<Buffer[]>;
export declare function parsePdfFile(type: PROCESSING_TYPE.TEXT, input: Buffer | string | Uint8Array | URL, options?: AfppParseOptions, callback?: undefined): Promise<string[]>;
export declare function parsePdfFile<T>(type: PROCESSING_TYPE.MIXED, input: Buffer | string | Uint8Array | URL, options: AfppParseOptions, callback: PageProcessor<T>): Promise<T[]>;
/**
 * Streaming PDF parser that yields results as pages are processed.
 * Useful for large PDFs where you want to process pages as they become available
 * rather than waiting for all pages to complete.
 *
 * @example
 * ```typescript
 * for await (const { pageNumber, data } of streamPdfFile(PROCESSING_TYPE.IMAGE, './large.pdf')) {
 *   await saveImage(data, `page-${pageNumber}.png`);
 * }
 * ```
 */
export declare function streamPdfFile(type: PROCESSING_TYPE.IMAGE, input: Buffer | string | Uint8Array | URL, options?: AfppParseOptions): AsyncGenerator<StreamingResult<Buffer>>;
export declare function streamPdfFile(type: PROCESSING_TYPE.TEXT, input: Buffer | string | Uint8Array | URL, options?: AfppParseOptions): AsyncGenerator<StreamingResult<string>>;