word文档转html(只支持段落和表格)

李李李李李李等待，李李李李李李哒哒

于 2024-08-30 10:01:19 发布

阅读量602

点赞数 4

文章标签： word html java

本文链接：https://blog.csdn.net/m0_63159822/article/details/141710176

版权

maven依赖
 <dependency>  
        <groupId>org.apache.poi</groupId>  
        <artifactId>poi-ooxml</artifactId>  
        <version>5.2.3</version>  
</dependency>  


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

import java.io.*;


public class WordToHtmlConverter {



    public static String convertWordToHtml(String filePath) throws IOException {
        StringBuilder htmlContent = new StringBuilder("<html><body>");

        try (FileInputStream fis = new FileInputStream(filePath);
             XWPFDocument document = new XWPFDocument(fis)) {

            // 首先处理段落
            for (XWPFParagraph paragraph : document.getParagraphs()) {
                htmlContent.append("<p>").append(getParagraphText(paragraph)).append("</p>");
            }

            // 然后处理表格
            for (XWPFTable table : document.getTables()) {
                htmlContent.append("<table border='1'>");
                for (XWPFTableRow row : table.getRows()) {
                    htmlContent.append("<tr>");
                    for (XWPFTableCell cell : row.getTableCells()) {
                        htmlContent.append("<td>");
                        for (XWPFParagraph paragraph : cell.getParagraphs()) {
                            htmlContent.append(getParagraphText(paragraph));
                        }
                        htmlContent.append("</td>");
                    }
                    htmlContent.append("</tr>");
                }
                htmlContent.append("</table>");
            }
        }

        htmlContent.append("</body></html>");
        return htmlContent.toString();
    }

    public static String convertWordToHtml(InputStream inputStream) throws IOException {
        StringBuilder htmlContent = new StringBuilder("<html><body>");

        try (XWPFDocument document = new XWPFDocument(inputStream)) {

            // 处理段落
            for (XWPFParagraph paragraph : document.getParagraphs()) {
                htmlContent.append("<p>").append(getParagraphText(paragraph)).append("</p>");
            }

            // 处理表格
            for (XWPFTable table : document.getTables()) {
                htmlContent.append("<table border='1'>");
                for (XWPFTableRow row : table.getRows()) {
                    htmlContent.append("<tr>");
                    for (XWPFTableCell cell : row.getTableCells()) {
                        htmlContent.append("<td>");
                        // 这里简单地将单元格中的文本内容添加到HTML中
                        // 注意：单元格可能包含多个段落，需要遍历它们
                        for (XWPFParagraph paragraph : cell.getParagraphs()) {
                            htmlContent.append(getParagraphText(paragraph));
                        }
                        htmlContent.append("</td>");
                    }
                    htmlContent.append("</tr>");
                }
                htmlContent.append("</table>");
            }

        }

        htmlContent.append("</body></html>");
        return htmlContent.toString();
    }

    // 辅助方法，用于获取段落中的文本内容
    private static String getParagraphText(XWPFParagraph paragraph) {
        StringBuilder text = new StringBuilder();
        for (XWPFRun run : paragraph.getRuns()) {
            text.append(run.getText(0));
        }
        return text.toString();
    }

    public static void main(String[] args) {
        try {
            String html = convertWordToHtml("xxx.doc");
            System.out.println(html); 
        } catch (IOException e) {
            e.printStackTrace();
        }
    }


}