pdf-ocr-cli/dist/renderPdfToPng.js

A CLI tool for OCR processing of PDF files using Mistral API with optional LLM verification

"use strict";
var __importDefault = (this && this.__importDefault) || function (mod) {
    return (mod && mod.__esModule) ? mod : { "default": mod };
};
Object.defineProperty(exports, "__esModule", { value: true });
exports.renderPdfToPng = renderPdfToPng;
const pdf_lib_1 = require("pdf-lib");
const fs_1 = __importDefault(require("fs"));
const path_1 = __importDefault(require("path"));
const os_1 = __importDefault(require("os"));
const pdf2pic_1 = require("pdf2pic");
const uuid_1 = require("uuid");
/**
 * Renders a single-page PDF to a PNG image
 * @param pdfBuffer - Buffer containing a single-page PDF
 * @param resolution - DPI resolution for rendering (default: 150)
 * @returns Buffer containing the PNG image
 * @throws Error if the PDF has multiple pages or is invalid
 */
async function renderPdfToPng(pdfBuffer, resolution = 150) {
    try {
        // Load the PDF document to check page count
        const pdfDoc = await pdf_lib_1.PDFDocument.load(pdfBuffer);
        // Ensure it's a single-page PDF
        if (pdfDoc.getPageCount() !== 1) {
            throw new Error('Input PDF must be a single page');
        }
        // Create a temporary directory
        const tempDir = path_1.default.join(os_1.default.tmpdir(), `pdf-ocr-${(0, uuid_1.v4)()}`);
        fs_1.default.mkdirSync(tempDir, { recursive: true });
        // Configure pdf2pic
        const pdf2picOptions = {
            density: resolution,
            savePath: tempDir,
            format: "png",
            width: 2000, // Max width
            height: 2000 // Max height
        };
        // Convert PDF to PNG
        const converter = (0, pdf2pic_1.fromBuffer)(pdfBuffer, pdf2picOptions);
        const result = await converter(1); // Convert page 1
        if (!result || !result.path) {
            throw new Error('Failed to render PDF to PNG');
        }
        // Read the PNG file
        const pngBuffer = fs_1.default.readFileSync(result.path);
        // Clean up temporary files
        try {
            fs_1.default.unlinkSync(result.path);
            fs_1.default.rmdirSync(tempDir);
        }
        catch (cleanupError) {
            console.warn('Failed to clean up temporary files:', cleanupError);
        }
        // Return the PNG buffer
        return pngBuffer;
    }
    catch (error) {
        // Handle errors
        if (error instanceof Error) {
            if (error.message === 'Input PDF must be a single page') {
                throw error; // Re-throw our own error
            }
            throw new Error(`Invalid PDF: ${error.message}`);
        }
        else {
            throw new Error('Invalid PDF: Unknown error');
        }
    }
}