java将html带图片转为.docx、 .pdf、 .image的工具类

可以使用Apache POI和iText库来实现将HTML带图片转为.docx、.pdf和.image格式的工具类。

1. 将HTML转为.docx

import java.io.*;
import org.apache.poi.xwpf.usermodel.*;

public class HTMLToDOCXConverter {
    public static void convert(String htmlFilePath, String docxFilePath) throws IOException {
        // Load HTML file
        File htmlFile = new File(htmlFilePath);
        FileInputStream fis = new FileInputStream(htmlFile);
        byte[] data = new byte[(int) htmlFile.length()];
        fis.read(data);
        fis.close();
        String htmlContent = new String(data, "UTF-8");

        // Create a new Word document
        XWPFDocument document = new XWPFDocument();

        // Add HTML content to the document
        XWPFParagraph paragraph = document.createParagraph();
        XWPFRun run = paragraph.createRun();
        run.setText(htmlContent);

        // Save the document
        FileOutputStream fos = new FileOutputStream(docxFilePath);
        document.write(fos);
        fos.close();
    }
}

2. 将HTML转为.pdf

import java.io.*;
import com.itextpdf.html2pdf.*;

public class HTMLToPDFConverter {
    public static void convert(String htmlFilePath, String pdfFilePath) throws IOException {
        // Load HTML file
        File htmlFile = new File(htmlFilePath);
        FileInputStream fis = new FileInputStream(htmlFile);
        byte[] data = new byte[(int) htmlFile.length()];
        fis.read(data);
        fis.close();
        String htmlContent = new String(data, "UTF-8");

        // Create a new PDF document
        PdfDocument pdfDocument = new PdfDocument(new PdfWriter(pdfFilePath));

        // Convert HTML to PDF
        ConverterProperties converterProperties = new ConverterProperties();
        HtmlConverter.convertToPdf(htmlContent, pdfDocument, converterProperties);

        // Close the document
        pdfDocument.close();
    }
}

3. 将HTML转为.image

import java.awt.*;
import java.awt.image.*;
import java.io.*;
import javax.imageio.*;
import org.jsoup.*;
import org.jsoup.nodes.*;
import org.jsoup.select.*;

public class HTMLToImageConverter {
    public static void convert(String htmlFilePath, String imageFilePath) throws IOException {
        // Load HTML file
        File htmlFile = new File(htmlFilePath);
        Document doc = Jsoup.parse(htmlFile, "UTF-8");

        // Get the body element
        Element body = doc.body();

        // Create a new BufferedImage
        BufferedImage image = new BufferedImage(body.outerWidth(), body.outerHeight(), BufferedImage.TYPE_INT_RGB);

        // Create a new Graphics2D object
        Graphics2D g2d = image.createGraphics();
        g2d.setColor(Color.WHITE);
        g2d.fillRect(0, 0, body.outerWidth(), body.outerHeight());

        // Render the HTML to the BufferedImage
        body.children().forEach((child) -> {
            g2d.translate(child.absLeft(), child.absTop());
            child.render(g2d);
            g2d.translate(-child.absLeft(), -child.absTop());
        });

        // Save the BufferedImage as an image file
        File imageFile = new File(imageFilePath);
        ImageIO.write(image, "png", imageFile);
    }
}

以上是将HTML带图片转为.docx、.pdf和.image格式的工具类，你可以根据自己的需要进行调整和改进。