实现html和word的相互转换（带图片）

项目后端使用了springboot，maven，前端使用了ckeditor富文本编辑器。目前从html转换的word为doc格式，而图片处理支持的是docx格式，所以需要手动把doc另存为docx，然后才可以进行图片替换。

2021年了，我回来填坑了，又碰到了需要把 html 转换成 word 的场景，目前对于格式不固定的带图片的导出，有以下解决方案：

后端只做部分样式处理，导出工作交给前端完成，代码实现见第四部分。

一.添加maven依赖

主要使用了以下和poi相关的依赖，为了便于获取html的图片元素，还使用了jsoup：

<dependency><groupId>org.apache.poi</groupId><artifactId>poi</artifactId><version>3.14</version>
</dependency><dependency><groupId>org.apache.poi</groupId><artifactId>poi-scratchpad</artifactId><version>3.14</version>
</dependency><dependency><groupId>org.apache.poi</groupId><artifactId>poi-ooxml</artifactId><version>3.14</version>
</dependency><dependency><groupId>fr.opensagres.xdocreport</groupId><artifactId>xdocreport</artifactId><version>1.0.6</version>
</dependency><dependency><groupId>org.apache.poi</groupId><artifactId>poi-ooxml-schemas</artifactId><version>3.14</version>
</dependency><dependency><groupId>org.apache.poi</groupId><artifactId>ooxml-schemas</artifactId><version>1.3</version>
</dependency><dependency><groupId>org.jsoup</groupId><artifactId>jsoup</artifactId><version>1.11.3</version>
</dependency>

二.word转换为html

在springboot项目的resources目录下新建static文件夹，将需要转换的word文件temp.docx粘贴进去，由于static是springboot的默认资源文件，所以不需要在配置文件里面另行配置了，如果改成其他名字，需要在application.yml进行相应配置。

doc格式转换为html：

public static String docToHtml() throws Exception {File path = new File(ResourceUtils.getURL("classpath:").getPath());String imagePathStr = path.getAbsolutePath() + "\\static\\image\\";String sourceFileName = path.getAbsolutePath() + "\\static\\test.doc";String targetFileName = path.getAbsolutePath() + "\\static\\test2.html";File file = new File(imagePathStr);if(!file.exists()) {file.mkdirs();}HWPFDocument wordDocument = new HWPFDocument(new FileInputStream(sourceFileName));org.w3c.dom.Document document = DocumentBuilderFactory.newInstance().newDocumentBuilder().newDocument();WordToHtmlConverter wordToHtmlConverter = new WordToHtmlConverter(document);//保存图片，并返回图片的相对路径wordToHtmlConverter.setPicturesManager((content, pictureType, name, width, height) -> {try (FileOutputStream out = new FileOutputStream(imagePathStr + name)) {out.write(content);} catch (Exception e) {e.printStackTrace();}return "image/" + name;});wordToHtmlConverter.processDocument(wordDocument);org.w3c.dom.Document htmlDocument = wordToHtmlConverter.getDocument();DOMSource domSource = new DOMSource(htmlDocument);StreamResult streamResult = new StreamResult(new File(targetFileName));TransformerFactory tf = TransformerFactory.newInstance();Transformer serializer = tf.newTransformer();serializer.setOutputProperty(OutputKeys.ENCODING, "utf-8");serializer.setOutputProperty(OutputKeys.INDENT, "yes");serializer.setOutputProperty(OutputKeys.METHOD, "html");serializer.transform(domSource, streamResult);return targetFileName;
}

docx格式转换为html

public static String docxToHtml() throws Exception {File path = new File(ResourceUtils.getURL("classpath:").getPath());String imagePath = path.getAbsolutePath() + "\\static\\image";String sourceFileName = path.getAbsolutePath() + "\\static\\test.docx";String targetFileName = path.getAbsolutePath() + "\\static\\test.html";OutputStreamWriter outputStreamWriter = null;try {XWPFDocument document = new XWPFDocument(new FileInputStream(sourceFileName));XHTMLOptions options = XHTMLOptions.create();// 存放图片的文件夹options.setExtractor(new FileImageExtractor(new File(imagePath)));// html中图片的路径options.URIResolver(new BasicURIResolver("image"));outputStreamWriter = new OutputStreamWriter(new FileOutputStream(targetFileName), "utf-8");XHTMLConverter xhtmlConverter = (XHTMLConverter) XHTMLConverter.getInstance();xhtmlConverter.convert(document, outputStreamWriter, options);} finally {if (outputStreamWriter != null) {outputStreamWriter.close();}}return targetFileName;
}

转换成功后会生成对应的html文件，如果想在前端展示，直接读取文件转换为String返回给前端即可。

public static String readfile(String filePath) {File file = new File(filePath);InputStream input = null;try {input = new FileInputStream(file);} catch (FileNotFoundException e) {e.printStackTrace();}StringBuffer buffer = new StringBuffer();byte[] bytes = new byte[1024];try {for (int n; (n = input.read(bytes)) != -1;) {buffer.append(new String(bytes, 0, n, "utf8"));}} catch (IOException e) {e.printStackTrace();}return buffer.toString();
}

在富文本编辑器ckeditor中的显示效果：

三.html转换为word

实现思路就是先把html中的所有图片元素提取出来，统一替换为变量字符"${imgReplace}“，如果多张图片，可以依序排列下去，之后生成对应的doc文件（之前试过直接生成docx文件发现打不开，这个问题尚未找到好的解决方法），我们将其另存为docx文件，之后就可以替换变量为图片了：

public static String writeWordFile(String content) {String path = "D:/wordFile";Map<String, Object> param = new HashMap<String, Object>();if (!"".equals(path)) {File fileDir = new File(path);if (!fileDir.exists()) {fileDir.mkdirs();}content = HtmlUtils.htmlUnescape(content);List<HashMap<String, String>> imgs = getImgStr(content);int count = 0;for (HashMap<String, String> img : imgs) {count++;//处理替换以“/>”结尾的img标签content = content.replace(img.get("img"), "${imgReplace" + count + "}");//处理替换以“>”结尾的img标签content = content.replace(img.get("img1"), "${imgReplace" + count + "}");Map<String, Object> header = new HashMap<String, Object>();try {File filePath = new File(ResourceUtils.getURL("classpath:").getPath());String imagePath = filePath.getAbsolutePath() + "\\static\\";imagePath += img.get("src").replaceAll("/", "\\\\");//如果没有宽高属性，默认设置为400*300if(img.get("width") == null || img.get("height") == null) {header.put("width", 400);header.put("height", 300);}else {header.put("width", (int) (Double.parseDouble(img.get("width"))));header.put("height", (int) (Double.parseDouble(img.get("height"))));}header.put("type", "jpg");header.put("content", OfficeUtil.inputStream2ByteArray(new FileInputStream(imagePath), true));} catch (FileNotFoundException e) {e.printStackTrace();}param.put("${imgReplace" + count + "}", header);}try {// 生成doc格式的word文档，需要手动改为docxbyte by[] = content.getBytes("UTF-8");ByteArrayInputStream bais = new ByteArrayInputStream(by);POIFSFileSystem poifs = new POIFSFileSystem();DirectoryEntry directory = poifs.getRoot();DocumentEntry documentEntry = directory.createDocument("WordDocument", bais);FileOutputStream ostream = new FileOutputStream("D:\\wordFile\\temp.doc");poifs.writeFilesystem(ostream);bais.close();ostream.close();// 临时文件（手动改好的docx文件）CustomXWPFDocument doc = OfficeUtil.generateWord(param, "D:\\wordFile\\temp.docx");//最终生成的带图片的word文件FileOutputStream fopts = new FileOutputStream("D:\\wordFile\\final.docx");doc.write(fopts);fopts.close();} catch (Exception e) {e.printStackTrace();}}return "D:/wordFile/final.docx";}//获取html中的图片元素信息public static List<HashMap<String, String>> getImgStr(String htmlStr) {List<HashMap<String, String>> pics = new ArrayList<HashMap<String, String>>();Document doc = Jsoup.parse(htmlStr);Elements imgs = doc.select("img");for (Element img : imgs) {HashMap<String, String> map = new HashMap<String, String>();if(!"".equals(img.attr("width"))) {map.put("width", img.attr("width").substring(0, img.attr("width").length() - 2));}if(!"".equals(img.attr("height"))) {map.put("height", img.attr("height").substring(0, img.attr("height").length() - 2));}map.put("img", img.toString().substring(0, img.toString().length() - 1) + "/>");map.put("img1", img.toString());map.put("src", img.attr("src"));pics.add(map);}return pics;}

OfficeUtil工具类，之前发现网上的写法只支持一张图片的修改，多张图片就会报错，是因为添加了图片，processParagraphs方法中的runs的大小改变了，会报ArrayList的异常，就和我们循环list中删除元素会报异常道理一样，解决方法就是复制一个新的Arraylist进行循环：

package com.example.demo.util;  import java.io.ByteArrayInputStream;
import java.io.FileInputStream;
import java.io.IOException;
import java.io.InputStream;
import java.util.ArrayList;
import java.util.Iterator;
import java.util.List;
import java.util.Map;
import java.util.Map.Entry;import org.apache.poi.POIXMLDocument;
import org.apache.poi.hwpf.extractor.WordExtractor;
import org.apache.poi.openxml4j.opc.OPCPackage;
import org.apache.poi.xwpf.usermodel.XWPFParagraph;
import org.apache.poi.xwpf.usermodel.XWPFRun;
import org.apache.poi.xwpf.usermodel.XWPFTable;
import org.apache.poi.xwpf.usermodel.XWPFTableCell;
import org.apache.poi.xwpf.usermodel.XWPFTableRow;  /** * 适用于word 2007*/
public class OfficeUtil {  /** * 根据指定的参数值、模板，生成 word 文档 * @param param 需要替换的变量 * @param template 模板 */  public static CustomXWPFDocument generateWord(Map<String, Object> param, String template) {  CustomXWPFDocument doc = null;try {  OPCPackage pack = POIXMLDocument.openPackage(template);  doc = new CustomXWPFDocument(pack);  if (param != null && param.size() > 0) {  //处理段落List<XWPFParagraph> paragraphList = doc.getParagraphs();  processParagraphs(paragraphList, param, doc);  //处理表格Iterator<XWPFTable> it = doc.getTablesIterator();  while (it.hasNext()) {XWPFTable table = it.next();  List<XWPFTableRow> rows = table.getRows();  for (XWPFTableRow row : rows) {  List<XWPFTableCell> cells = row.getTableCells();  for (XWPFTableCell cell : cells) {  List<XWPFParagraph> paragraphListTable =  cell.getParagraphs();  processParagraphs(paragraphListTable, param, doc);  }  }  }  }  } catch (Exception e) {  e.printStackTrace();  }  return doc;  }  /** * 处理段落 * @param paragraphList */  public static void processParagraphs(List<XWPFParagraph> paragraphList,Map<String, Object> param,CustomXWPFDocument doc){  if(paragraphList != null && paragraphList.size() > 0){  for(XWPFParagraph paragraph:paragraphList){//poi转换过来的行间距过大，需要手动调整if(paragraph.getSpacingBefore() >= 1000 || paragraph.getSpacingAfter() > 1000) {paragraph.setSpacingBefore(0);paragraph.setSpacingAfter(0);}//设置word中左右间距paragraph.setIndentationLeft(0);paragraph.setIndentationRight(0);List<XWPFRun> runs = paragraph.getRuns();//加了图片，修改了paragraph的runs的size，所以循环不能使用runsList<XWPFRun> allRuns = new ArrayList<XWPFRun>(runs);for (XWPFRun run : allRuns) {String text = run.getText(0);  if(text != null){boolean isSetText = false;  for (Entry<String, Object> entry : param.entrySet()) {  String key = entry.getKey();  if(text.indexOf(key) != -1){  isSetText = true;  Object value = entry.getValue();  if (value instanceof String) {//文本替换text = text.replace(key, value.toString());  } else if (value instanceof Map) {//图片替换text = text.replace(key, "");  Map pic = (Map)value;  int width = Integer.parseInt(pic.get("width").toString());  int height = Integer.parseInt(pic.get("height").toString());  int picType = getPictureType(pic.get("type").toString());  byte[] byteArray = (byte[]) pic.get("content");  ByteArrayInputStream byteInputStream = new ByteArrayInputStream(byteArray);  try {  String blipId = doc.addPictureData(byteInputStream,picType);  doc.createPicture(blipId,doc.getNextPicNameNumber(picType), width, height,paragraph);} catch (Exception e) {  e.printStackTrace();  }  }  }  }  if(isSetText){  run.setText(text,0);  }  }  }  }  }  }  /** * 根据图片类型，取得对应的图片类型代码 * @param picType * @return int */  private static int getPictureType(String picType){  int res = CustomXWPFDocument.PICTURE_TYPE_PICT;  if(picType != null){  if(picType.equalsIgnoreCase("png")){  res = CustomXWPFDocument.PICTURE_TYPE_PNG;  }else if(picType.equalsIgnoreCase("dib")){  res = CustomXWPFDocument.PICTURE_TYPE_DIB;  }else if(picType.equalsIgnoreCase("emf")){  res = CustomXWPFDocument.PICTURE_TYPE_EMF;  }else if(picType.equalsIgnoreCase("jpg") || picType.equalsIgnoreCase("jpeg")){  res = CustomXWPFDocument.PICTURE_TYPE_JPEG;  }else if(picType.equalsIgnoreCase("wmf")){  res = CustomXWPFDocument.PICTURE_TYPE_WMF;  }  }  return res;  }  /** * 将输入流中的数据写入字节数组 * @param in * @return */  public static byte[] inputStream2ByteArray(InputStream in,boolean isClose){  byte[] byteArray = null;  try {  int total = in.available();  byteArray = new byte[total];  in.read(byteArray);  } catch (IOException e) {  e.printStackTrace();  }finally{  if(isClose){  try {  in.close();  } catch (Exception e2) {  System.out.println("关闭流失败");  }  }  }  return byteArray;  }
}

我认为之所以word2003不支持图片替换，主要是因为处理2003版本的HWPFDocument对象被声明为了final，我们就无法重写他的方法了。

而处理2007版本的类为XWPFDocument，是可以继承的，通过继承XWPFDocument，重写createPicture方法即可实现图片替换，以下为对应的CustomXWPFDocument类：

package com.example.demo.util;    import java.io.IOException;
import java.io.InputStream;
import org.apache.poi.openxml4j.opc.OPCPackage;
import org.apache.poi.xwpf.usermodel.XWPFDocument;
import org.apache.poi.xwpf.usermodel.XWPFParagraph;
import org.apache.xmlbeans.XmlException;
import org.apache.xmlbeans.XmlToken;
import org.openxmlformats.schemas.drawingml.x2006.main.CTNonVisualDrawingProps;
import org.openxmlformats.schemas.drawingml.x2006.main.CTPositiveSize2D;
import org.openxmlformats.schemas.drawingml.x2006.wordprocessingDrawing.CTInline;  /** * 自定义 XWPFDocument，并重写 createPicture()方法 */
public class CustomXWPFDocument extends XWPFDocument {    public CustomXWPFDocument(InputStream in) throws IOException {    super(in);    }    public CustomXWPFDocument() {    super();    }    public CustomXWPFDocument(OPCPackage pkg) throws IOException {    super(pkg);    }    /** * @param ind * @param width 宽 * @param height 高 * @param paragraph  段落 */  public void createPicture(String blipId, int ind, int width, int height,XWPFParagraph paragraph) {    final int EMU = 9525;    width *= EMU;    height *= EMU;    CTInline inline = paragraph.createRun().getCTR().addNewDrawing().addNewInline();    String picXml = ""    + "<a:graphic xmlns:a=\"http://schemas.openxmlformats.org/drawingml/2006/main\">"    + "   <a:graphicData uri=\"http://schemas.openxmlformats.org/drawingml/2006/picture\">"    + "      <pic:pic xmlns:pic=\"http://schemas.openxmlformats.org/drawingml/2006/picture\">"    + "         <pic:nvPicPr>" + "            <pic:cNvPr id=\""    + ind    + "\" name=\"Generated\"/>"    + "            <pic:cNvPicPr/>"    + "         </pic:nvPicPr>"    + "         <pic:blipFill>"    + "            <a:blip r:embed=\""    + blipId    + "\" xmlns:r=\"http://schemas.openxmlformats.org/officeDocument/2006/relationships\"/>"    + "            <a:stretch>"    + "               <a:fillRect/>"    + "            </a:stretch>"    + "         </pic:blipFill>"    + "         <pic:spPr>"    + "            <a:xfrm>"    + "               <a:off x=\"0\" y=\"0\"/>"    + "               <a:ext cx=\""    + width    + "\" cy=\""    + height    + "\"/>"    + "            </a:xfrm>"    + "            <a:prstGeom prst=\"rect\">"    + "               <a:avLst/>"    + "            </a:prstGeom>"    + "         </pic:spPr>"    + "      </pic:pic>"    + "   </a:graphicData>" + "</a:graphic>";    inline.addNewGraphic().addNewGraphicData();    XmlToken xmlToken = null;    try {    xmlToken = XmlToken.Factory.parse(picXml);    } catch (XmlException xe) {    xe.printStackTrace();    }    inline.set(xmlToken);   inline.setDistT(0);      inline.setDistB(0);      inline.setDistL(0);      inline.setDistR(0);      CTPositiveSize2D extent = inline.addNewExtent();    extent.setCx(width);    extent.setCy(height);    CTNonVisualDrawingProps docPr = inline.addNewDocPr();      docPr.setId(ind);      docPr.setName("图片" + ind);      docPr.setDescr("测试");   }
}

以上就是通过POI实现html和word的相互转换，对于html无法转换为可读的docx这个问题，不是很好解决。

html和word的相互转换也可以通过 freemarker 模板的方式实现，这个是同事给的思路，我没有尝试，大家有兴趣的可以试试，缺点就是模板是固定的，不够灵活，不适用于经常变动的网页。

历经两年多又发现了一个新的解决方案，亲测可用，见第四部分。

四.html转换为word之前端实现

前端实现的方法比较灵活，也支持表格，echarts 的导出，使用 file-saver 插件实现。

在 vue 和 angular 下测试了都可以用，而后端只需要对图片和样式进行处理，转换和导出工作交给前端来做（后端同学露出了如释重负的笑容）。

java 代码实现

controller 实现：

    @PostMapping("/article/htmlFormat")public Ret html(String html) throws Exception {return articleService.formatHtmlStyle(html);}

service 实现：

 // 实现图片和样式处理的方法public Ret formatHtmlStyle(String html) {JSONArray picsArr = new JSONArray();// 缩小图片Document doc = Jsoup.parse(html);Elements elementsP = doc.getElementsByTag("p");for (int i = 0; i < elementsP.size(); i++) {Element element = elementsP.get(i);boolean hasImg = false;Elements elementsChildren = element.children();for (int j = 0; j < elementsChildren.size(); j++) {Element elementChild = elementsChildren.get(j);if (elementChild.nodeName().equals("img")) {hasImg = true;break;}}if (hasImg) {element.attr("style", "text-align: center ;");} else {element.attr("style","font-family: FangSong_GB2312 ;font-size:18px;text-indent: 2em ;line-height:34px;text-align:justify;");}}Elements elements = doc.getElementsByTag("img");for (int i = 0; i < elements.size(); i++) {Element element = elements.get(i);String src = element.attr("src");JSONObject picjo = new JSONObject();picjo.put("index", i);// 将网络图片转为 base64 格式picjo.put("src", CommonUtil.urlToBase64(src));picsArr.add(picjo);element.attr("src", "" + i);}Ret ret = Ret.create().setCodeAndMsg(200).set("html", doc.body().html()).set("pics", picsArr);return ret;  }

注意事项：

1.这里的Ret就是一个结果封装类，不是重点，只要把 html 和 pics 返回即可。

2.html 中的图片路径是网络图片，需要转换为 base64 才能在 word 中显示。

CommonUtil 中的转换代码如下：

public static String urlToBase64(String imgUrl) {InputStream inputStream = null;ByteArrayOutputStream outputStream = null;byte[] buffer = null;try {// 创建URLURL url = new URL(imgUrl);// 创建链接HttpURLConnection conn = (HttpURLConnection) url.openConnection();conn.setRequestMethod("GET");conn.setConnectTimeout(5000);inputStream = conn.getInputStream();outputStream = new ByteArrayOutputStream();// 将内容读取内存中buffer = new byte[1024];int len = -1;while ((len = inputStream.read(buffer)) != -1) {outputStream.write(buffer, 0, len);}buffer = outputStream.toByteArray();} catch (IOException e) {e.printStackTrace();} finally {if (inputStream != null) {try {// 关闭inputStream流inputStream.close();} catch (IOException e) {e.printStackTrace();}}if (outputStream != null) {try {// 关闭outputStream流outputStream.close();} catch (IOException e) {e.printStackTrace();}}}// 对字节数组Base64编码String base64 = "data:image/png;base64," + new BASE64Encoder().encode(buffer);return base64;}

vue 代码实现

1.安装 file-saver

npm install file-saver --save
bower install file-saver

如果提示 “bower不是内部或外部命令”，需要安装bower

npm install -g bower

安装 TypeScript 定义

npm install @types/file-saver --save-dev

2.使用

导入 saveAs

import { saveAs } from 'file-saver';

实现下载处理逻辑

downloadArticle: function() {var FileSaver = require('file-saver');var html = "这里是你的html内容";//var html = "<html>" + this.$refs.article.innerHTML + "</html>";var param = new URLSearchParams();param.append('html', html);// 调用上面的 java 接口获取处理后的html内容this.$axios({method: 'post',url: '/article/htmlFormat',data: param}).then((response) = > {const _static = {mhtml: {top: "Mime-Version: 1.0\nContent-Base: " + location.href + "\nContent-Type: Multipart/related; boundary=\"NEXT.ITEM-BOUNDARY\";type=\"text/html\"\n\n--NEXT.ITEM-BOUNDARY\nContent-Type: text/html; charset=\"utf-8\"\nContent-Location: " + location.href + "\n\n<!DOCTYPE html>\n<html>\n_html_</html>",head: "<head>\n<meta http-equiv=\"Content-Type\" content=\"text/html; charset=utf-8\">\n<style>\n_styles_\n</style>\n</head>\n",body: "<body>_body_</body>"}};const dealhtml = response.data.html;const img = response.data.pics;let mhtmlBottom = "\n";for (let i = 0; i < img.length; i++) {const uri = img[i].src;const index = img[i].index;mhtmlBottom += "--NEXT.ITEM-BOUNDARY\n";mhtmlBottom += "Content-Location: " + index + "\n";mhtmlBottom += "Content-Type: " + uri.substring(uri.indexOf(":") + 1, uri.indexOf(";")) + "\n";mhtmlBottom += "Content-Transfer-Encoding: " + uri.substring(uri.indexOf(";") + 1, uri.indexOf(",")) + "\n\n";mhtmlBottom += uri.substring(uri.indexOf(",") + 1) + "\n\n";}mhtmlBottom += "--NEXT.ITEM-BOUNDARY--";// 整合html代码片段const fileContent = _static.mhtml.top.replace("_html_", _static.mhtml.body.replace("_body_", dealhtml)) + mhtmlBottom;// 导出const blob = new Blob([fileContent], {type: "application/msword;charset=utf-8"});saveAs(blob, `testImage.doc`);});
}

注意事项：

不要图省事在后端把 base64 写在 image 标签的 src 里面，这样会导致导出的word文档只有在切换为编辑视图后才能显示图片。

最终效果：