JACOB + POI 含网络图片和公式(Mathml格式)的Html转Word

一、需求

应用场景是这样。有一个HTML页面，内含公式和图片，需要转到Word并下载。

框架是Vue + Java.

其中公式分为两种，一种是Mathml标签：

<math>   
	<mrow>
		<msup><mi>a</mi><mn>3</mn></msup>
		<mo>+</mo>	   
		<msup><mi>b</mi><mn>2</mn></msup>
		<mo>=</mo>		   
		<msup><mi>c</mi><mn>2</mn></msup>
	</mrow>   
</math>

还有一种是MathType格式的img标签：

<img class="Wirisformula" style="max-width:none;vertical-align:-16px;" role="math" alt="x equals fraction numerator negative b plus-or-minus square root of b squared minus 4 a c end root over denominator 2 a end fraction" height="50" width="147" data-mathml="«math xmlns=¨http://www.w3.org/1998/Math/MathML¨»«mi»x«/mi»«mo»=«/mo»«mfrac»«mrow»«mo»-«/mo»«mi»b«/mi»«mo»§#177;«/mo»«msqrt»«msup»«mi»b«/mi»«mn»2«/mn»«/msup»«mo»-«/mo»«mn»4«/mn»«mi»a«/mi»«mi»c«/mi»«/msqrt»«/mrow»«mrow»«mn»2«/mn»«mi»a«/mi»«/mrow»«/mfrac»«/math»" src="data:image/svg+xml;charset=utf8,%3Csvg%20xmlns%3D%22http%3A%2F%2Fwww.w3.org%2F2000%2Fsvg%22%20xmlns%3Awrs%3D%22http%3A%2F%2Fwww.wiris.com%2Fxml%2Fmathml-extension%22%20height%3D%2250%22%20width%3D%22147%22%20wrs%3Abaseline%3D%2234%22%3E%3C!--MathML%3A%20%3Cmath%20xmlns%3D%22http%3A%2F%2Fwww.w3.org%2F1998%2FMath%2FMathML%22%3E%3Cmi%3Ex%3C%2Fmi%3E%3Cmo%3E%3D%3C%2Fmo%3E%3Cmfrac%3E%3Cmrow%3E%3Cmo%3E-%3C%2Fmo%3E%3Cmi%3Eb%3C%2Fmi%3E%3Cmo%3E%26%23xB1%3B%3C%2Fmo%3E%3Cmsqrt%3E%3Cmsup%3E%3Cmi%3Eb%3C%2Fmi%3E%3Cmn%3E2%3C%2Fmn%3E%3C%2Fmsup%3E%3Cmo%3E-%3C%2Fmo%3E%3Cmn%3E4%3C%2Fmn%3E%3Cmi%3Ea%3C%2Fmi%3E%3Cmi%3Ec%3C%2Fmi%3E%3C%2Fmsqrt%3E%3C%2Fmrow%3E%3Cmrow%3E%3Cmn%3E2%3C%2Fmn%3E%3Cmi%3Ea%3C%2Fmi%3E%3C%2Fmrow%3E%3C%2Fmfrac%3E%3C%2Fmath%3E--%3E%3Cdefs%3E%3Cstyle%20type%3D%22text%2Fcss%22%3E%40font-face%7Bfont-family%3A'ae2ef524fbf3d9fe611d5a8e90fefdc'%3Bsrc%3Aurl(data%3Afont%2Ftruetype%3Bcharset%3Dutf-8%3Bbase64%2CAAEAAAAMAIAAAwBAT1MvMjv%2FLJYAAADMAAAATmNtYXDgWxEdAAABHAAAADRjdnQgAAAABwAAAVAAAAAEZ2x5ZoYrxVAAAAFUAAAA0WhlYWQOdyayAAACKAAAADZoaGVhC0UVwQAAAmAAAAAkaG10eCg8AIUAAAKEAAAACGxvY2EAAAVKAAACjAAAAAxtYXhwBIoEWwAAApgAAAAgbmFtZXSF9ZsAAAK4AAABrXBvc3QDogHPAAAEaAAAACBwcmVwukanGAAABIgAAAANAAAGtAGQAAUAAAgACAAAAAAACAAIAAAAAAAAAQIAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAACAgICAAAAAg8AMGe%2F57AAAHPgGyAAAAAAACAAEAAQAAABQAAwABAAAAFAAEACAAAAAEAAQAAQAAAGH%2F%2FwAAAGH%2F%2F%2F%2BgAAEAAAAAAAAABwACAFUAAAMAA6sAAwAHAAAzESERJSERIVUCq%2F2rAgD%2BAAOr%2FFVVAwAAAwAt%2F3QEAwRZAAsAFwAdADsYAbAdELAD1LADELAU1LAUELAc1LAcELAJ1LAcELAOPLAJELAbPACwBhCwEdSwBhCwANSwABCwF9QwMQEiABEWEjMyEjcQJgYWAwIGIyImNTQ2MwE1BhMjEgIBs%2F7fFvWy07oDhYZwFgxOhVmysoUB7YwEslEEWf7f%2Ft71%2Ft8BM%2BMBp5yyLf6d%2FwBlyJzfsvxZjF0B5%2F1eAAAAAAEAAAABAACav9usXw889QADCAD%2F%2F%2F%2F%2F1a3uPf%2F%2F%2F%2F%2FVre49AAH%2B9QQDBkMAAAAKAAIAAQAAAAAAAQAABz7%2BTgAAF3AAAf%2F8BAMAAQAAAAAAAAAAAAAAAAAAAAIDUgBVBEwALQAAAAAAAAAoAAAA0QABAAAAAgAeAAMAAAAAAAIAgAQAAAAAAAQAADsAAAAAAAAAFQECAAAAAAAAAAEAFgAAAAAAAAAAAAIADgAWAAAAAAAAAAMANAAkAAAAAAAAAAQAFgBYAAAAAAAAAAUAFgBuAAAAAAAAAAYACwCEAAAAAAAAAAgAHACPAAEAAAAAAAEAFgAAAAEAAAAAAAIADgAWAAEAAAAAAAMANAAkAAEAAAAAAAQAFgBYAAEAAAAAAAUAFgBuAAEAAAAAAAYACwCEAAEAAAAAAAgAHACPAAMAAQQJAAEAFgAAAAMAAQQJAAIADgAWAAMAAQQJAAMANAAkAAMAAQQJAAQAFgBYAAMAAQQJAAUAFgBuAAMAAQQJAAYACwCEAAMAAQQJAAgAHACPAE0AYQB0AGgAIABGAG8AbgB0ACAAMgBSAGUAZwB1AGwAYQByAE0AYQB0AGgAcwAgAEYAbwByACAATQBvAHIAZQAgAE0AYQB0AGgAIABGAG8AbgB0ACAAMgBNAGEAdABoACAARgBvAG4AdAAgADIAVgBlAHIAcwBpAG8AbgAgADEALgAwTWF0aF9Gb250XzIATQBhAHQAaABzACAARgBvAHIAIABNAG8AcgBlAAAAAAMAAAAAAAADnwHPAAAAAAAAAAAAAAAAAAAAAAAAAAC5ByIAAI2FGACyAAAAAAAA)format('truetype')%3Bfont-weight%3Anormal%3Bfont-style%3Anormal%3B%7D%40font-face%7Bfont-family%3A'math19244194cbc38427b5aca056d4d'%3Bsrc%3Aurl(data%3Afont%2Ftruetype%3Bcharset%3Dutf-8%3Bbase64%2CAAEAAAAMAIAAAwBAT1MvMi7iBBMAAADMAAAATmNtYXDEvmKUAAABHAAAAERjdnQgDVUNBwAAAWAAAAA6Z2x5ZoPi2VsAAAGcAAABmGhlYWQQC2qxAAADNAAAADZoaGVhCGsXSAAAA2wAAAAkaG10eE2rRkcAAAOQAAAAEGxvY2EAHTwYAAADoAAAABRtYXhwBT0FPgAAA7QAAAAgbmFtZaBxlY4AAAPUAAABn3Bvc3QB9wD6AAAFdAAAACBwcmVwa1uragAABZQAAAAUAAADSwGQAAUAAAQABAAAAAAABAAEAAAAAAAAAQEAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAACAgICAAAAAg1UADev96AAAD6ACWAAAAAAACAAEAAQAAABQAAwABAAAAFAAEADAAAAAIAAgAAgAAAD0AsSIS%2F%2F8AAAA9ALEiEv%2F%2F%2F8T%2FUd3xAAEAAAAAAAAAAAAAAVQDLACAAQAAVgAqAlgCHgEOASwCLABaAYACgACgANQAgAAAAAAAAAArAFUAgACrANUBAAErAAcAAAACAFUAAAMAA6sAAwAHAAAzESERJSERIVUCq%2F2rAgD%2BAAOr%2FFVVAwAAAgCAAOsC1QIVAAMABwBlGAGwCBCwBtSwBhCwBdSwCBCwAdSwARCwANSwBhCwBzywBRCwBDywARCwAjywABCwAzwAsAgQsAbUsAYQsAfUsAcQsAHUsAEQsALUsAYQsAU8sAcQsAQ8sAEQsAA8sAIQsAM8MTATITUhHQEhNYACVf2rAlUBwFXVVVUAAgCA%2F%2F8CgAKrAAsADwBlGAGwEBCwD9SwDxCwADywABCwAdSwARCwBNSwBBCwBdSwARCwCjywBBCwBzywBRCwDjwAsBAQsA%2FUsA8QsAzUsAwQsAnUsAkQsArUsAoQsAHUsAEQsALUsAEQsAQ8sAoQsAc8MDETMzUzFTMVIxUjJwcRIRUhgNZV1dVVAdUCAP4AAdXW1lbU1QH%2B1VUAAQCAAVUC1QGrAAMAMBgBsAQQsQAD9rADPLECB%2FWwATyxBQPmALEAABMQsQAG5bEAARMQsAE8sQMF9bACPBMhFSGAAlX9qwGrVgABAAAAAQAA1XjOQV8PPPUAAwQA%2F%2F%2F%2F%2F9Y6E3P%2F%2F%2F%2F%2F1joTcwAA%2FyAEgAOrAAAACgACAAEAAAAAAAEAAAPo%2F2oAABdwAAD%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%2BgAAAAAAAAAAAAAAAAAAAAAAAAAAuQcRAACNhRgAsgAAABUUE7EAAT8%3D)format('truetype')%3Bfont-weight%3Anormal%3Bfont-style%3Anormal%3B%7D%3C%2Fstyle%3E%3C%2Fdefs%3E%3Ctext%20font-family%3D%22Arial%22%20font-size%3D%2216%22%20font-style%3D%22italic%22%20text-anchor%3D%22middle%22%20x%3D%224.5%22%20y%3D%2234%22%3Ex%3C%2Ftext%3E%3Ctext%20font-family%3D%22math19244194cbc38427b5aca056d4d%22%20font-size%3D%2216%22%20text-anchor%3D%22middle%22%20x%3D%2217.5%22%20y%3D%2234%22%3E%3D%3C%2Ftext%3E%3Cline%20stroke%3D%22%23000000%22%20stroke-linecap%3D%22square%22%20stroke-width%3D%221%22%20x1%3D%2228.5%22%20x2%3D%22143.5%22%20y1%3D%2228.5%22%20y2%3D%2228.5%22%2F%3E%3Ctext%20font-family%3D%22math19244194cbc38427b5aca056d4d%22%20font-size%3D%2216%22%20text-anchor%3D%22middle%22%20x%3D%2236.5%22%20y%3D%2222%22%3E%26%23x2212%3B%3C%2Ftext%3E%3Ctext%20font-family%3D%22Arial%22%20font-size%3D%2216%22%20font-style%3D%22italic%22%20text-anchor%3D%22middle%22%20x%3D%2247.5%22%20y%3D%2222%22%3Eb%3C%2Ftext%3E%3Ctext%20font-family%3D%22math19244194cbc38427b5aca056d4d%22%20font-size%3D%2216%22%20text-anchor%3D%22middle%22%20x%3D%2260.5%22%20y%3D%2222%22%3E%26%23xB1%3B%3C%2Ftext%3E%3Cpolyline%20fill%3D%22none%22%20points%3D%2212%2C-20%2011%2C-20%205%2C0%202%2C-8%22%20stroke%3D%22%23000000%22%20stroke-linecap%3D%22square%22%20stroke-width%3D%221%22%20transform%3D%22translate(67.5%2C24.5)%22%2F%3E%3Cpolyline%20fill%3D%22none%22%20points%3D%225%2C0%202%2C-8%200%2C-7%22%20stroke%3D%22%23000000%22%20stroke-linecap%3D%22square%22%20stroke-width%3D%221%22%20transform%3D%22translate(67.5%2C24.5)%22%2F%3E%3Cline%20stroke%3D%22%23000000%22%20stroke-linecap%3D%22square%22%20stroke-width%3D%221%22%20x1%3D%2279.5%22%20x2%3D%22141.5%22%20y1%3D%224.5%22%20y2%3D%224.5%22%2F%3E%3Ctext%20font-family%3D%22Arial%22%20font-size%3D%2216%22%20font-style%3D%22italic%22%20text-anchor%3D%22middle%22%20x%3D%2285.5%22%20y%3D%2222%22%3Eb%3C%2Ftext%3E%3Ctext%20font-family%3D%22Arial%22%20font-size%3D%2212%22%20text-anchor%3D%22middle%22%20x%3D%2294.5%22%20y%3D%2215%22%3E2%3C%2Ftext%3E%3Ctext%20font-family%3D%22math19244194cbc38427b5aca056d4d%22%20font-size%3D%2216%22%20text-anchor%3D%22middle%22%20x%3D%22105.5%22%20y%3D%2222%22%3E%26%23x2212%3B%3C%2Ftext%3E%3Ctext%20font-family%3D%22Arial%22%20font-size%3D%2216%22%20text-anchor%3D%22middle%22%20x%3D%22117.5%22%20y%3D%2222%22%3E4%3C%2Ftext%3E%3Ctext%20font-family%3D%22ae2ef524fbf3d9fe611d5a8e90fefdc%22%20font-size%3D%2216%22%20font-style%3D%22italic%22%20text-anchor%3D%22middle%22%20x%3D%22126.5%22%20y%3D%2222%22%3Ea%3C%2Ftext%3E%3Ctext%20font-family%3D%22Arial%22%20font-size%3D%2216%22%20font-style%3D%22italic%22%20text-anchor%3D%22middle%22%20x%3D%22135.5%22%20y%3D%2222%22%3Ec%3C%2Ftext%3E%3Ctext%20font-family%3D%22Arial%22%20font-size%3D%2216%22%20text-anchor%3D%22middle%22%20x%3D%2281.5%22%20y%3D%2245%22%3E2%3C%2Ftext%3E%3Ctext%20font-family%3D%22ae2ef524fbf3d9fe611d5a8e90fefdc%22%20font-size%3D%2216%22%20font-style%3D%22italic%22%20text-anchor%3D%22middle%22%20x%3D%2290.5%22%20y%3D%2245%22%3Ea%3C%2Ftext%3E%3C%2Fsvg%3E" align="middle" />

这种可以从data-mathml中获取Mathml格式的数据

图片就是标准的img标签，src属性可以是图片链接或者BASE64格式的数据。

需求就是把这4种东西转到word里，其中图片要缓存下来，脱机可以正常显示。

二、依赖包

主要用到JACOB和POI，以及Mathml转OpenXml的fmath-mathml。

还有辅助包Jsoup

这里注意下POI的版本。统一用4.1.2。因为poi-ooxml-schemas在maven上只有4.1.2的，如果其他的POI用5.0会报错。

        <dependency>
			<groupId>org.jsoup</groupId>
			<artifactId>jsoup</artifactId>
			<version>1.11.3</version>
		</dependency>
		<!-- https://mvnrepository.com/artifact/net.sf.jacob-project/jacob -->
		<dependency>
			<groupId>net.sf.jacob-project</groupId>
			<artifactId>jacob</artifactId>
			<version>cus-1.0.0</version>
			<scope>system</scope>
			<systemPath>${project.basedir}/lib/jacob.jar
           </systemPath>
		</dependency>
		<dependency>
		    <groupId>org.apache.poi</groupId>
		    <artifactId>poi-ooxml-schemas</artifactId>
		    <version>4.1.2</version>
		</dependency>

		<dependency>
			<groupId>org.apache.poi</groupId>
			<artifactId>poi</artifactId>
			<version>4.1.2</version>
		</dependency>
		<dependency>
			<groupId>org.apache.poi</groupId>
			<artifactId>poi-ooxml</artifactId>
			<version>4.1.2</version>
		</dependency>
		<!-- https://mvnrepository.com/artifact/org.apache.poi/poi-scratchpad -->
		<dependency>
			<groupId>org.apache.poi</groupId>
			<artifactId>poi-scratchpad</artifactId>
			<version>4.1.2</version>
		</dependency>
		<dependency>
			<groupId>com.jdom</groupId>
			<artifactId>jdom</artifactId>
			<version>cus-1.0.0</version>
			<scope>system</scope>
			<systemPath>${project.basedir}/lib/jdom-2.0.6.jar
        </systemPath>
		</dependency>
		<dependency>
			<groupId>com.mathml</groupId>
			<artifactId>mathml</artifactId>
			<version>cus-1.0.0</version>
			<scope>system</scope>
			<systemPath>${project.basedir}/lib/fmath-mathml-java.jar
           </systemPath>
		</dependency>

三、代码

思路就是先把图片和公式用特殊的字符标记一下，例如${imageReplace1}

然后图片缓存成文件，拷贝到Word文档里。

公式比较简单，直接替换成OpenXml格式就可以了。

话不多说，直接上代码

1.继承自XWPFDocument的核心处理类

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.poi.xwpf.usermodel.XWPFRun;
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.CTEffectExtent;
import org.openxmlformats.schemas.drawingml.x2006.wordprocessingDrawing.CTInline;

import fmath.conversion.c.a;

/**
 * 自定义 XWPFDocument
 */
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, XWPFRun run) {
		final int EMU = 9525;
		width *= EMU;
		height *= EMU;
		CTInline inline = run.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("图片");

		inline.addNewCNvGraphicFramePr();
	}

	public void createMath(String mathStr, XWPFRun run) {
		XmlToken xmlToken = null;
		try {
			String openXML = a.a(mathStr);
			String header = "<m:oMathPara xmlns:w=\"http://schemas.openxmlformats.org/wordprocessingml/2006/main\" xmlns:m=\"http://schemas.openxmlformats.org/officeDocument/2006/math\">";
			openXML = header + openXML + "</m:oMathPara>";
			xmlToken = XmlToken.Factory.parse(openXML);
		} catch (XmlException xe) {
			xe.printStackTrace();
		}
		run.getCTR().set(xmlToken);
	}
}

2.处理公式和图片的工具类。

遍历所有的XWPFParagraph，找到需要替换的位置

import java.io.ByteArrayInputStream;
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.ooxml.POIXMLDocument;
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
	 */
	@SuppressWarnings("rawtypes")
	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runs
				List<XWPFRun> allRuns = new ArrayList<XWPFRun>(runs);
				for (XWPFRun run : allRuns) {
					String text = run.getText(0);
					if (text != null) {
						for (Entry<String, Object> entry : param.entrySet()) {
							String key = entry.getKey();
							if (text.indexOf(key) != -1) {
								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
									Map pic = (Map) value;
									String mapType = (String) pic.get("mapType");

									if ("img".equals(mapType)) {
										text = text.replace(key, "");

										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) + 200, width,
													height, run);
										} catch (Exception e) {
											e.printStackTrace();
										}
										run.setText(text, 0);
									} else if ("math".equals(mapType)) {
										doc.createMath((String) pic.get("content"), run);
									}
								}
							}
						}
					}
				}
			}
		}
	}

	/**
	 * 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;
	}
}

3.JACOB HTML转Word

需求里是用模板，所有这里用的是Open命令不是Add

	public static void htmlToWord(String html, String wordFile) {
		ActiveXComponent app = new ActiveXComponent("Word.Application"); // 启动word
		try {
			app.setProperty("Visible", new Variant(false));
			Dispatch wordDoc = app.getProperty("Documents").toDispatch();
			//打开word文件
			wordDoc = Dispatch.invoke(wordDoc, "Open", Dispatch.Method, new Object[]{"d:/MathType/B4H.dotx", new Variant(true), new Variant(true)}, new int[1]).toDispatch(); 
			// wordDoc = Dispatch.invoke(wordDoc, "Add", Dispatch.Method, new Object[0], new int[1]).toDispatch();
			Dispatch selection = app.getProperty("Selection").toDispatch();
			Dispatch.call(selection, "EndKey", new Variant(6));
			Dispatch.invoke(selection, "InsertFile", Dispatch.Method,
					new Object[] { html, "", new Variant(false), new Variant(false), new Variant(true) }, new int[3]);
			Dispatch.invoke(wordDoc, "SaveAs", Dispatch.Method, new Object[] { wordFile, new Variant(16) }, new int[1]);
			Dispatch.call(wordDoc, "Close", new Variant(false));
		} catch (Exception e) {
			e.printStackTrace();
		} finally {
			app.invoke("Quit", new Variant[] {});
		}
	}

4.找到HTML里需要替换的元素

这里用到了JSoup

    //获取html中的需要替换的元素信息
	public static HashMap<String, List<HashMap<String, String>>> getReplaceStr(String htmlStr) throws Exception {
		HashMap<String, List<HashMap<String, String>>> replaceParams = new HashMap<String, List<HashMap<String, String>>>();
		
		// 图片
		List<HashMap<String, String>> pics = new ArrayList<HashMap<String, String>>();
		// 图片公式
		List<HashMap<String, String>> imgmaths = new ArrayList<HashMap<String, String>>();

		Document doc = Jsoup.parse(htmlStr);
		
		// 遍历img标签
		Elements imgs = doc.select("img");
		for (Element img : imgs) {
			HashMap<String, String> map = new HashMap<String, String>();
			if (img.hasAttr("data-mathml")) {
				// mathtype格式的公式
				String mathml = img.attr("data-mathml");
				mathml = mathml.replace("«", "<").replace("»", ">").replace("¨", "\"")
						.replace("§#177;", "&plusmn;").replace("xmlns=\"http://www.w3.org/1998/Math/MathML\"", "");
				map.put("tag", img.toString());
				map.put("content", mathml);
				imgmaths.add(map);
			} else {
				// 图片
				String src = img.attr("src");
				String type = "";
				
				// 保存图片
				String fileName = "";
				if (src.startsWith("data:image")) {
					type = src.substring(src.indexOf("/") + 1, src.indexOf(";"));
					String regex = "data:image/(png|gif|jpg|jpeg|bmp|tif|psd|ICO);base64,";
					fileName = saveBase64Img(src.replaceAll(regex, ""), type);
				} else {
					type = src.substring(src.lastIndexOf(".") + 1);
					fileName = saveNetImg(img.attr("src"));
				}

				if(!"".equals(img.attr("width"))) {
					map.put("width", img.attr("width"));
				}
				if(!"".equals(img.attr("height"))) {
					map.put("height", img.attr("height"));
				}
				map.put("img", img.toString().substring(0, img.toString().length() - 1) + "/>");
				map.put("img1", img.toString());
				map.put("src", img.attr("src"));
				map.put("fileName", fileName);
				map.put("type", type);
				
				pics.add(map);
			}
		}
		
		replaceParams.put("pics", pics);
		replaceParams.put("imgmaths", imgmaths);

		return replaceParams;
	}

5.入口调用方法

这里用到了HtmlUtils.htmlUnescape，是springweb里的一个方法，用来标准化html字符串

public static String writeWordFile(String content) throws Exception {
		String path = "D:\\MathType";
		Map<String, Object> param = new HashMap<String, Object>();
		Map<String, String> mathParam = new HashMap<String, String>();

		if (!"".equals(path)) {
			File fileDir = new File(path);
			if (!fileDir.exists()) {
				fileDir.mkdirs();
			}
			content = HtmlUtils.htmlUnescape(content);
			
			HashMap<String, List<HashMap<String, String>>> repalceMap = getReplaceStr(content);
			// 图片公式
			List<HashMap<String, String>> maths = repalceMap.get("imgmaths");
			int count = 0;
			for (HashMap<String, String> math : maths) {
				count++;
				String key = "${mathReplace" + count + "}";
				content = content.replace(math.get("tag"), key + "\r\n");
				
				Map<String, Object> header = new HashMap<String, Object>();
				header.put("content", math.get("content"));
				header.put("mapType", "math");

				param.put(key, header);
			}

			// mathml公式
			String regex = "(<math)(.*?)(</math>)";
			Pattern r = Pattern.compile(regex, Pattern.CASE_INSENSITIVE | Pattern.DOTALL);
			Matcher m = r.matcher(content);
	        while (m.find()) {
	        	count++;
				String key = "${mathReplace" + count + "}";
				String replaceStr = m.group(0);
				content = content.replace(replaceStr, key);
				
				Map<String, Object> header = new HashMap<String, Object>();
				header.put("content", replaceStr);
				header.put("mapType", "math");
				
				param.put(key, header);
	        }
			
			// 图片
			List<HashMap<String, String>> imgs = repalceMap.get("pics");
			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 {
					String imagePath = "d:/MathType/";
					String[] sep = img.get("src").replaceAll("/", "\\\\").split("\\\\");
					imagePath += sep[sep.length - 1];
					//如果没有宽高属性，默认设置为400*300
					if(img.get("width") == null || "".equals(img.get("width")) || img.get("height") == null || "".equals(img.get("height"))) {
						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", img.get("type"));
					header.put("mapType", "img");
					header.put("content", OfficeUtil.inputStream2ByteArray(new FileInputStream(imagePath), true));
				} catch (FileNotFoundException e) {
					e.printStackTrace();
				}
				param.put("${imgReplace" + count + "}", header);
			}
			try {

				FileOutputStream fos = new FileOutputStream("d:/MathType/temp.html");
				fos.write(content.getBytes("utf-8"));
				fos.close();
				
				htmlToWord("d:/MathType/temp.html", "d:/MathType/temp.docx");
				
				// 临时文件（手动改好的docx文件）
				CustomXWPFDocument doc = OfficeUtil.generateWord(param, mathParam, "D:\\MathType\\temp.docx");
				
				//最终生成的带图片的word文件
				FileOutputStream fopts = new FileOutputStream("D:\\MathType\\final.docx");
				doc.write(fopts);
				fopts.close();
			} catch (Exception e) {
				e.printStackTrace();
			}

		}
		return "D:/wordFile/final.docx";
	}

大致就是这样。欢迎斧正。

附：对下划线做处理。场景是这样，在HTML中下划线是用u标签做的，并且字体不统一。

            Document contentDoc = Jsoup.parse(content);

			// 遍历u标签
			Elements uList = contentDoc.select("u");
			for (Element u : uList) {
				if (u.selectFirst("span") != null) {
					// span标签添加字体
					String style = u.selectFirst("span").attr("style");
					if (!style.contains("font-family")) {
						style = "font-family: Calibri;" + style;
						u.selectFirst("span").attr("style", style);
					}
				} else {
					// u标签内没有span的情况
					String ustr = u.html();
					u.html("");
					u.append("<span style='font-family: Calibri;font-size:10.5pt'>" + ustr + "</span>");
				}
				
			}
			content = contentDoc.outerHtml();

邮箱：yushen6403@163.com