java按顺序读取docx文件中的内容（包括文字、表格、图片）

近期在工作中需要解析word，里面包括文字、表格、图片等，在网上找了很多关于解析word的，都是单独处理的，由于我需要按顺序解析内容，而且要清晰的指导每个内容的位置在哪儿，所以单独处理显然不行，最开始使用python的docx类库，这种解决起来有点小问题，后来切换到了java来处理这个，没想到POI有这么好用的类库

Alex_81D

1689人浏览 · 2024-11-14 23:15:00

Alex_81D · 2024-11-14 23:15:00 发布

背景：

近期在做大模型文章解析时需要解析word，从而进行精准回答，遇到了一些问题：里面包括文字、表格、图片等，在网上找了很多关于解析word的，都是单独处理的，由于我需要按顺序解析内容，而且要清晰的指导每个内容的位置在哪儿，所以单独处理显然不行，最开始使用python的docx类库，这种解决起来有点小问题，后来切换到了java来处理这个，没想到POI有这么好用的类库，所以来分享一下，话不多说，开整：

（java按顺序读取docx文档中的文本与表格？）

先看看原文档长啥样子，的确有点麻烦：

1.复杂文本

2.复杂表格

解决问题思路

一、使用ApachePOI读取docx文件，引入maven:

<!-- Apache POI -->
<dependency>
	<groupId>org.apache.poi</groupId>
	<artifactId>poi</artifactId>
	<version>4.1.0</version>
</dependency>

<!--处理word文档需要的额外的jar包-->
<dependency>
	<groupId>org.apache.poi</groupId>
	<artifactId>poi-ooxml</artifactId>
	<version>4.1.0</version>
</dependency>

<!--处理word文档需要的额外的jar包-->
<dependency>
	<groupId>org.apache.poi</groupId>
	<artifactId>poi-ooxml-schemas</artifactId>
	<version>4.1.0</version>
</dependency>

注意：请使用这个4.1.0的版本的，我之前用的是5.1.0和poi-ooxml-full这个依赖包，总是会报这个错：java.lang.NoSuchMethodException: org.openxmlformats.schemas.wordprocessingml.x2006.main.impl.CTPictureBaseImpl.<init>

java.lang.NoSuchMethodException: org.openxmlformats.schemas.wordprocessingml.x2006.main.impl.CTPictureBaseImpl.<init>(org.apache.xmlbeans.SchemaType, boolean)
    at java.lang.Class.getConstructor0(Class.java:3082)
    at java.lang.Class.getDeclaredConstructor(Class.java:2178)
    at org.apache.xmlbeans.impl.schema.SchemaTypeImpl.getJavaImplConstructor2(SchemaTypeImpl.java:1817)
    at org.apache.xmlbeans.impl.schema.SchemaTypeImpl.createUnattachedSubclass(SchemaTypeImpl.java:1961)
    at org.apache.xmlbeans.impl.schema.SchemaTypeImpl.createUnattachedNode(SchemaTypeImpl.java:1950)
    at org.apache.xmlbeans.impl.schema.SchemaTypeImpl.createElementType(SchemaTypeImpl.java:1051)
    at org.apache.xmlbeans.impl.values.XmlObjectBase.create_element_user(XmlObjectBase.java:938)
    at org.apache.xmlbeans.impl.store.Xobj.getUser(Xobj.java:1675)
    at org.apache.xmlbeans.impl.store.Cur.getUser(Cur.java:2659)
    at org.apache.xmlbeans.impl.store.Cur.getObject(Cur.java:2652)
    at org.apache.xmlbeans.impl.store.Cursor._getObject(Cursor.java:995)
    at org.apache.xmlbeans.impl.store.Cursor.getObject(Cursor.java:2904)
    at org.apache.poi.xwpf.usermodel.XWPFDocument.onDocumentRead(XWPFDocument.java:162)
    at org.apache.poi.POIXMLDocument.load(POIXMLDocument.java:169)
    at org.apache.poi.xwpf.usermodel.XWPFDocument.<init>(XWPFDocument.java:112)
    at org.apache.poi.xwpf.extractor.XWPFWordExtractor.<init>(XWPFWordExtractor.java:60)
    at com.programmersupermarket.autocreateproject.common.util.poi.testRead.main(testRead.java:17)

https://blog.csdn.net/qq_16307345/article/details/79076657

https://zhuanlan.zhihu.com/p/556486563

二、获取内容-方法一

1.段落和表格

public static void parseFile(String path){
	try {
		//String path = "C:\\Users\\admin\\Desktop\\first.docx";
		FileInputStream fis = new FileInputStream(path);
		XWPFDocument document = new XWPFDocument(fis);

		// 遍历文档中的所有元素（段落和表格）
		List<IBodyElement> bodyElements = document.getBodyElements();
		for (IBodyElement bodyElement : bodyElements) {
			if (bodyElement instanceof XWPFParagraph) {
				XWPFParagraph paragraph = (XWPFParagraph) bodyElement;
				System.out.println(paragraph.getStyleID() + ":" + paragraph.getText());
			} else if (bodyElement instanceof XWPFTable) {
				System.out.println(((XWPFTable) bodyElement).getText());
			} else if (bodyElement instanceof XWPFPicture) {
				System.out.println(Arrays.toString(((XWPFPicture) bodyElement).getPictureData().getData()));
			}
		}

		fis.close();
	} catch (IOException e) {
		e.printStackTrace();
	}
}

对元素进行遍历，获取内容(段落和表格)，发现此遍历方式不能读取到图片，只能通过getAllPictures方法得到所有图片，但失去了段落和图片之间的顺序。

2.段落、表格和图片

public static void parseFile2(String path){
	try {
		//String path = "C:\\Users\\admin\\Desktop\\first.docx";
		FileInputStream fis = new FileInputStream(path);
		XWPFDocument document = new XWPFDocument(fis);

		// 遍历文档中的所有元素（段落和表格）
		List<IBodyElement> bodyElements = document.getBodyElements();
		for (IBodyElement element : bodyElements) {
			if (element instanceof XWPFParagraph) {
				XWPFParagraph paragraph = (XWPFParagraph) element;
				String text = paragraph.getText();
				if (text != null && !text.isEmpty()) {
				  //处理段落或正文
				  System.out.println(text);
				} else {
					// 顺序遍历图片
					paragraph.getIRuns().forEach(run -> {
						if (run instanceof XWPFRun) {
							XWPFRun xWPFRun = (XWPFRun) run;
							for (XWPFPicture picture : xWPFRun.getEmbeddedPictures()) {
								XWPFPictureData pictureData = picture.getPictureData();
								String base64Image = "<img src='data:image/png;base64," + Base64.getEncoder().encodeToString((pictureData.getData())) + "'/>";
								System.out.println(base64Image);								
							}
						}
					});
				}
			} else if (element instanceof XWPFTable) {
			//处理表格
				XWPFTable table = (XWPFTable) element;
				String text = table.getText();
				System.out.println(text);
			}
		}
		fis.close();
	} catch (IOException e) {
		e.printStackTrace();
	}
}

这样遍历就可以按顺序读取docx文件的内容了。

参考：stackoverflow

https://blog.csdn.net/gloriaied/article/details/135405896

三、获取内容-方法二

1.读取word文本

package docx;
 
import java.io.FileInputStream;
import java.io.InputStream;
import java.util.List;
 
import org.apache.commons.lang3.StringUtils;
import org.apache.poi.xwpf.usermodel.BodyElementType;
import org.apache.poi.xwpf.usermodel.IBodyElement;
import org.apache.poi.xwpf.usermodel.XWPFDocument;
import org.apache.poi.xwpf.usermodel.XWPFParagraph;
import org.apache.poi.xwpf.usermodel.XWPFRun;
/**
 * 读文本
 */
public class TextRead {
    
    public static void main(String[] args) throws Exception {
        readDocx();
    }
    public static void readDocx() throws Exception{
        InputStream is;
        is = new FileInputStream("test.docx");
        XWPFDocument xwpf = new XWPFDocument(is);
        
        List<IBodyElement> ibs = xwpf.getBodyElements();
        for(IBodyElement ib : ibs) {
            BodyElementType elementType = ib.getElementType();
            //表格
            if(elementType == BodyElementType.TABLE) {
                System.out.println("table"+ib.getPart());
            }else {
                //段落
                XWPFParagraph para = (XWPFParagraph) ib;
                System.out.println("It is a new paragraph...The indention is "+para.getFirstLineIndent());
                List<XWPFRun> runs = para.getRuns();
                System.out.println("run");
                if(runs.size() <= 0) {
                    System.out.println("empty line");
                }
                for(XWPFRun run : runs) {
                    //如果片段没有文字，可能该片段是图片
                    if(StringUtils.isEmpty(run.text())) {
                        //该片段是图片时
                        if(run.getEmbeddedPictures().size() > 0) {
                            System.out.println("image***"+run.getEmbeddedPictures());
                        }else {
                            System.out.println("objects:"+run.getCTR().getObjectList());
                            //公式
                            if(run.getCTR().xmlText().indexOf("instrText") > 0) {
                                System.out.println("there is an equation field");
                            }
                        }
                    }else {
                        System.out.println("==="+run.getCharacterSpacing()+run.text());
                    }
                }
            }
        }
        is.close();
    }
 
}

注：这里使用的是这个标签做的判断：

BodyElementType elementType = ib.getElementType();
//表格
if(elementTypeequals(BodyElementType.TABLE)) //表格

else if (elementType.equals(BodyElementType.PARAGRAPH)) { //段落

https://blog.csdn.net/qq_38759137/article/details/130083388

2.word内容段落、公式xml标签

import lombok.extern.slf4j.Slf4j;
import org.apache.poi.xwpf.usermodel.*;
import org.dom4j.Document;
import org.dom4j.DocumentException;
import org.dom4j.Element;
import org.dom4j.io.SAXReader;
import org.openxmlformats.schemas.wordprocessingml.x2006.main.CTP;
import org.springframework.web.multipart.MultipartFile;


import javax.xml.transform.*;
import javax.xml.transform.stream.StreamResult;
import javax.xml.transform.stream.StreamSource;
import java.io.*;
import java.util.ArrayList;
import java.util.HashMap;
import java.util.List;
import java.util.Map;


/**
 * word内容段落、公式xml标签
 * 正常文本：
 * <w:r> 标签对应一个 XWPFRun对象
 * <w:t xml:space=“preserve”> 标签对应一段在 Word中的字符（也可以是一个字符）
 * 公式：
 * <w:object> 标签对应一个公式（当然我们这里只讲公式，此标签中也可以是一个 Excel也可以是一个 PPT等等）
 * <v:shape> 标签中有个 style属性，这里 style就是图片在 Word中显示的宽高
 * <v:imagedata> 标签关联着显示的图片（ <v:imagedata>为 <v:shape>子标签）
 * <o:OLEObject>标签关联着图片显示公式对应的二进制文件（二进制文件也是最重要的文件，没有这个文件当你在word中双击时，是打不开第三方公式插件的）
 *
 */
/**
 * @description:
 * @author: Han LiDong
 * @create: 2021/5/7 10:33
 * @update: 2021/5/7 10:33
 */
@Slf4j
public class WordUtil {

    public static void main(String[] args) throws Exception {
//        MultipartFile file = null;
//        InputStream inputStream = file.getInputStream();
        // 所有公式latex表达式集合,借助mathJax可以在页面上进行展示
        String path = "C:\\Users\\allen_sun\\Desktop\\first.docx";
        List <String> formulas = getFormulaMap(new FileInputStream(path));
        log.info("解析到{}个公式",formulas.size());
        // word解析 公式+文本
        wordAnalysis(new FileInputStream("C:\\D\\算法-β系数.docx"));
    }



    /**
     * word所有内容解析(公式、文本）
     * @param inputStream
     * @throws Exception
     */
    public static void wordAnalysis(InputStream inputStream) throws Exception {
        XWPFDocument word = new XWPFDocument(inputStream);
        try {
            for (IBodyElement ibodyelement : word.getBodyElements()) {
                if (ibodyelement.getElementType().equals(BodyElementType.PARAGRAPH)) {  //段落
                    XWPFParagraph paragraph = (XWPFParagraph) ibodyelement;
                    //段落解析
                    String paragraphStr = parseParagraph(paragraph);
                } else if (ibodyelement.getElementType().equals(BodyElementType.TABLE)) {   //表格
                    XWPFTable table = (XWPFTable) ibodyelement;
                    for (XWPFTableRow row : table.getRows()) {  //行
                        for (XWPFTableCell cell : row.getTableCells()) {    //cell
                            List<String> cellMath = new ArrayList<>(16);
                            for (XWPFParagraph paragraph : cell.getParagraphs()) {  //段落
                                //cell段落解析
                                String paragraphStr = parseParagraph(paragraph);
                                if (!"".equals(paragraphStr.trim())){
                                    cellMath.add(paragraphStr);
                                }
                            }
                            log.info("当前cell有{}个公式",cellMath.size());
                        }
                    }
                }
            }

        } finally {
            word.close();
        }
    }


    /**
     * 解析word中公式（转换成latex表达式）
     *
     * @param inputStream  文件流
     * @return
     */
    public static List<String> getFormulaMap(InputStream inputStream) throws IOException, DocumentException {
        //XWPFDocument xwpfDocument = new XWPFDocument(inputStream);
        Map<Integer, String> result = new HashMap<>();
        XWPFDocument word = new XWPFDocument(inputStream);
        //storing the found MathML in a AllayList of strings
        List<String> mathMLList = new ArrayList<String>(16);
        try {
            for (IBodyElement ibodyelement : word.getBodyElements()) {
                if (ibodyelement.getElementType().equals(BodyElementType.PARAGRAPH)) {  //段落
                    XWPFParagraph paragraph = (XWPFParagraph) ibodyelement;
                    //段落解析
                    List<String> mathList = parseMathParagraph(paragraph);
                    mathMLList.addAll(mathList);
                } else if (ibodyelement.getElementType().equals(BodyElementType.TABLE)) { //woed表格
                    XWPFTable table = (XWPFTable) ibodyelement;
                    for (XWPFTableRow row : table.getRows()) {
                        for (XWPFTableCell cell : row.getTableCells()) {
                            List<String> cellMath = new ArrayList<>(16);
                            for (XWPFParagraph paragraph : cell.getParagraphs()) {
                                //cell段落解析
                                List<String> mathList = parseMathParagraph(paragraph);
                                mathMLList.addAll(mathList);
                                //cellMath.addAll(mathList);
                            }
                        }
                    }
                }
            }
        } finally {
            word.close();
        }
        log.info("当前文档一共有{}个公式",mathMLList.size());
        return mathMLList;
    }

    /**
     * 公式段落解析
     * @param xwpfParagraph
     * @throws DocumentException
     */
    public static List<String> parseMathParagraph(XWPFParagraph xwpfParagraph) throws DocumentException {
        CTP ctp = xwpfParagraph.getCTP();
        String xmlText = ctp.xmlText();
        List<String > mathList = new ArrayList<>();
        if (xmlText.contains("<m:oMath>")) {
            //得到根节点的值
            SAXReader saxReader = new SAXReader();
            //将String类型的字符串转换成XML文本对象
            Document doc = saxReader.read(new ByteArrayInputStream(xmlText.getBytes()));
            Element root = doc.getRootElement();
            // 一个段落多个表达式解析
            List<Element> omMaths = root.selectNodes("//m:oMath");    //用xpath得到OMML节点
            for (Element ele : omMaths) {
                /**
                 * OMML -> MathML -> LaTex
                 * Office在安装目录中提供了将OMML转为MathML的xsl工具：MML2OMML.XSL
                 * MathML转LaTex使用网上找到另一个xsl工具mmltex.xsl。
                 */
                String xml = ele.asXML();
                //xml转 mathml
                String mml = convertOMML2MML(xml);
                //mathml转latx
                String latex = convertMML2Latex(mml);
                mathList.add(latex);
                log.info("late表达式：{}" , latex);
            }
        }
        return mathList;
    }

    /**
     * 段落解析
     * @param xwpfParagraph
     * @throws DocumentException
     */
    public static String parseParagraph(XWPFParagraph xwpfParagraph) throws DocumentException {
        CTP ctp = xwpfParagraph.getCTP();
        String xmlText = ctp.xmlText();
        StringBuilder sb = new StringBuilder();
//        if (xmlText.contains("<m:oMath>")) {

        //段落文本内容
        sb.append(xwpfParagraph.getParagraphText());
        //段落公式解析
        //得到根节点的值
        SAXReader saxReader = new SAXReader();
        //将String类型的字符串转换成XML文本对象
        Document doc = saxReader.read(new ByteArrayInputStream(xmlText.getBytes()));
        Element root = doc.getRootElement();
        // 一个段落多个表达式解析
        List<Element> omMaths = root.selectNodes("//m:oMath");    //用xpath得到OMML节点
        if (omMaths != null && !omMaths.isEmpty()) {
            for (Element ele : omMaths) {
                /**
                 * OMML -> MathML -> LaTex
                 * Office在安装目录中提供了将OMML转为MathML的xsl工具：MML2OMML.XSL
                 * MathML转LaTex使用网上找到另一个xsl工具mmltex.xsl。
                 */
                String xml = ele.asXML();
                //xml转 mathml
                String mathml = convertOMML2MML(xml);
                //mathml转latx
                String latex = convertMML2Latex(mathml);
                sb.append(latex);
                log.info("latex表达式：{}",latex);
            }
        }
        log.info("公式个数：{},解析内容：{}",omMaths.size(),sb.toString());
        return sb.toString();
    }

    /**
     * Description: xsl转换器</p>
     * @param s  公式xml字符串
     * @param xslpath   转换器路径
     * @param uriResolver xls依赖文件
     * @return
     */
    public static String xslConvert(String s, String xslpath, URIResolver uriResolver){
        TransformerFactory tFac = TransformerFactory.newInstance();
        if(uriResolver != null) {
            tFac.setURIResolver(uriResolver);
        }
        StreamSource xslSource = new StreamSource(WordUtil.class.getResourceAsStream(xslpath));
        StringWriter writer = new StringWriter();
        try {
            Transformer t = tFac.newTransformer(xslSource);
            Source source = new StreamSource(new StringReader(s));
            Result result = new StreamResult(writer);
            t.transform(source, result);
        } catch (TransformerException e) {
            log.error(e.getMessage(), e);
        }
        return writer.getBuffer().toString();
    }

    /**
     * <p>Description: 将mathml转为latx </p>
     * @param mml mathml字符串
     * @return
     */
    public static String convertMML2Latex(String mml){
        mml = mml.substring(mml.indexOf("?>")+2, mml.length()); //去掉xml的头节点
        URIResolver r = new URIResolver(){  //设置xls依赖文件的路径
            @Override
            public Source resolve(String href, String base) throws TransformerException {
                File f = new File("/conventer/mml2tex/" + href);
                InputStream inputStream = WordUtil.class.getResourceAsStream("/conventer/mml2tex/" + href);
                return new StreamSource(inputStream);
            }
        };
        String latex = xslConvert(mml, "/conventer/mml2tex/mmltex.xsl", r);
        if(latex != null && latex.length() > 1){
            latex = latex.substring(1, latex.length() - 1);
        }
        return latex;
    }
    /**
     * <p>Description: office xml转为mathml </p>
     * @param xml  公式xml
     * @return
     */
    public static String convertOMML2MML(String xml){
        // 进行转换的过程中需要借助这个文件,一般来说本机安装office就会有这个文件,找到就可以
        String result = xslConvert(xml, "C:\\Program Files (x86)\\Microsoft Office\\root\\Office16\\OMML2MML.XSL", null);
        return result;
    }
}

https://blog.csdn.net/han949417140/article/details/116521013

四、扩展变种

由于我的word数据不仅要预处理后保存服务器上，而且还要在前端页面展示，所以做个后端简单效果展示，给有需要的同学使用！

备注：此处方法没有做封装拆分处理，请勿介意

1.处理方法

public static void parseFile3(String path){
	try {
		FileInputStream fis = new FileInputStream(path);
		XWPFDocument document = new XWPFDocument(fis);

		// 遍历文档中的所有元素（段落和表格）
		List<IBodyElement> bodyElements = document.getBodyElements();
		for (IBodyElement element : bodyElements) {
			if (element instanceof XWPFParagraph) {
				XWPFParagraph paragraph = (XWPFParagraph) element;
				String text = paragraph.getText();
				if (text != null && !text.isEmpty()) {
					//处理段落或正文
					System.out.println(paragraph.getText());
				} else {
					// 顺序遍历图片
					paragraph.getIRuns().forEach(run -> {
						if (run instanceof XWPFRun) {
							XWPFRun xWPFRun = (XWPFRun) run;
							//如果片段没有文字，可能该片段为图片
							if (StringUtils.isEmpty(xWPFRun.text())){
								//该片段为图片时
								if (xWPFRun.getEmbeddedPictures().size()>0){
									for (XWPFPicture picture : xWPFRun.getEmbeddedPictures()) {
										XWPFPictureData pictureData = picture.getPictureData();
										String base64Image = "<img src='data:image/png;base64," + Base64.getEncoder().encodeToString((pictureData.getData())) + "'/>";
										System.out.println(base64Image);
									}
								}else {
									if(xWPFRun.getCTR().xmlText().indexOf("instrText") > 0) {
										System.out.println("there is an equation field");
									}
								}
							}
						}else if(run instanceof XWPFFieldRun){ //公式
							XWPFFieldRun xWPFRun = (XWPFFieldRun) run;


						}
					});
				}
			} else if (element instanceof XWPFTable) {
				//处理表格
				XWPFTable table = (XWPFTable) element;
				StringBuilder tableBuilder = new StringBuilder();
				tableBuilder.append("<table border ='1' >");
				//tableBuilder.append("<table style='border: 1px solid #000000' >");
				for(XWPFTableRow row: table.getRows()){
					tableBuilder.append("<tr>");
					for(XWPFTableCell cell:  row.getTableCells()){
						tableBuilder.append("<td");
						String text = cell.getText();
//                            cell.getParagraphs().get(0).getIRuns()
						String color = cell.getColor();
						color= StringUtils.isEmpty(color) ? "#FFFFFF" : "#"+color;
						tableBuilder.append(" style='border: 1px solid #000000 ; background-color:"+color+"' >");
//                            System.out.println(text);
						if (StringUtils.isEmpty(text)){
							List<XWPFParagraph> cellParagraphs = cell.getParagraphs();
							//System.out.println(cellParagraphs.size());
							for(XWPFParagraph cellsTemp:cellParagraphs){
								cellsTemp.getIRuns().forEach(run -> {
									if (run instanceof XWPFRun) {
										XWPFRun xWPFRun = (XWPFRun) run;
										for (XWPFPicture picture : xWPFRun.getEmbeddedPictures()) {
											XWPFPictureData pictureData = picture.getPictureData();
											String base64Image = "<img src='data:image/png;base64," + Base64.getEncoder().encodeToString((pictureData.getData())) + "'/>";
											tableBuilder.append(base64Image);
											//System.out.println(base64Image);
										}
									}
								});
							}
						}else {
							tableBuilder.append(text);
						}
						tableBuilder.append("</td>");
					}
					tableBuilder.append("</tr>");
				}
				tableBuilder.append("</table>");
				System.out.println(tableBuilder.toString());
//                    String text = table.getText();
//                    System.out.println(text);
			}
			System.out.println("</br>");
		}
		fis.close();
	} catch (IOException e) {
		e.printStackTrace();
	}
}

2.来看转化效果

原文：

转换后的效果：

这里面有很多方法，比如表格类型有每个cell的颜色，高度、宽度、是否居中等多个属性参数，如果要做的更加精细的同学自行拼装html标签信息，处理后回合原文一样的效果。

3.重点方法

此处重点类包括XWPFParagraph（文字）、XWPFRun（图片）、XWPFTable（表格）、XWPFFieldRun（公式）等

XWPFDocument：代表一个docx文档
XWPFParagraph：代表文档、表格、标题等种的段落，由多个XWPFRun组成
XWPFRun：代表具有同样风格的一段文本
XWPFTable：代表一个表格
XWPFTableRow：代表表格的一行
XWPFTableCell：代表表格的一个单元格
XWPFChar：表示.docx文件中的图表
XWPFHyperlink：表示超链接
XWPFPicture：代表图片
XWPFComment ：代表批注
XWPFFooter：代表页脚
XWPFHeader：代表页眉

可参考此文：

https://blog.csdn.net/xl_21/article/details/125205284

https://www.cnblogs.com/unruly/p/7483858.html

五、python处理思路

核心类库是python-docx

1.安装python-docx

pip install python-docx

2.python-docx-入门级

先看看这个方法处理，入门级别的，不能完全解决问题，文字和表格、图片全都分开了，明显是有问题的

from docx import Document
from docx.shared import Inches
 
# 加载现有Word文档
doc = Document('example_with_table_and_image.docx')
 
# 读取文档中的表格和图片
for table in doc.tables:
    for row in table.rows:
        for cell in row.cells:
            print(cell.text)
 
for image in doc.inline_shapes:
    if image.type == WD_INLINE_SHAPE_TYPE.IMAGE:
        print(f"Image name: {image.image.filename}")
        # 可以在这里修改图片
        # image.width = Inches(2)  # 修改图片宽度为2英寸
        # image.height = Inches(2)  # 修改图片高度为2英寸
 
# 保存修改后的文档
doc.save('modified_document.docx')

3.python-docx-高级做法

# docx
import docx
from docx.document import Document
from docx.text.paragraph import Paragraph
from docx.parts.image import ImagePart
from docx.table import _Cell, Table
 
from docx.oxml.table import CT_Tbl
from docx.oxml.text.paragraph import CT_P
    
# 该行只能有一个图片
def is_image(graph:Paragraph,doc:Document):
    images = graph._element.xpath('.//pic:pic')  # 获取所有图片
    for image in images:
        for img_id in image.xpath('.//a:blip/@r:embed'):  # 获取图片id
            part = doc.part.related_parts[img_id]  # 根据图片id获取对应的图片
            if isinstance(part, ImagePart):
                return True
    return False
 
# 获取图片（该行只能有一个图片）
def get_ImagePart(graph:Paragraph,doc:Document):
    images = graph._element.xpath('.//pic:pic')  # 获取所有图片
    for image in images:
        for img_id in image.xpath('.//a:blip/@r:embed'):  # 获取图片id
            part = doc.part.related_parts[img_id]  # 根据图片id获取对应的图片
            if isinstance(part, ImagePart):
                return part 
    return None
 
def iter_block_items(parent):
    """
    Yield each paragraph and table child within *parent*, in document order.
    Each returned value is an instance of either Table or Paragraph. *parent*
    would most commonly be a reference to a main Document object, but
    also works for a _Cell object, which itself can contain paragraphs and tables.
    """
    if isinstance(parent, Document):
        parent_elm = parent.element.body
    elif isinstance(parent, _Cell):
        parent_elm = parent._tc
    else:
        raise ValueError("something's not right")
 
    for child in parent_elm.iterchildren():
        if isinstance(child, CT_P):
            paragraph=Paragraph(child, parent)
            if is_image(paragraph,parent):
                print('[Image] ')
                yield get_ImagePart(paragraph, parent) 
            print('[Text] ')
            yield Paragraph(child, parent)
        elif isinstance(child, CT_Tbl):
            print('[Table] ')
            yield Table(child, parent)
 
 
def parse_word(word_path):
    doc = docx.Document(word_path)
    for part in iter_block_items(doc):
        print(part)

https://blog.csdn.net/qq_39600166/article/details/101537368

https://blog.csdn.net/GstGxf/article/details/133944790

https://blog.csdn.net/weixin_47663607/article/details/128669041

Python: Extract Text and Images from Word Documents