解析html将表格插入word文档

用到依赖包

 <dependency><groupId>org.jsoup</groupId><artifactId>jsoup</artifactId><version>1.15.2</version></dependency><dependency><groupId>org.aspectj</groupId><artifactId>aspectjweaver</artifactId><version>1.9.7</version></dependency><dependency><groupId>org.apache.poi</groupId><artifactId>poi-ooxml</artifactId><version>3.15</version></dependency>

代码

import lombok.extern.slf4j.Slf4j;
import org.apache.commons.collections4.CollectionUtils;
import org.apache.poi.xwpf.usermodel.*;
import org.jsoup.Jsoup;
import org.jsoup.nodes.Document;
import org.jsoup.nodes.Element;
import org.jsoup.nodes.Node;
import org.jsoup.select.Elements;
import org.openxmlformats.schemas.wordprocessingml.x2006.main.*;
import org.springframework.util.StringUtils;import java.io.File;
import java.io.FileInputStream;
import java.io.FileOutputStream;
import java.math.BigInteger;
import java.util.List;
import java.util.Objects;@Slf4j
public class Html2WordTest1 {/*** 每厘米567缇*/private static double teePerCM = 567;/*** 每厘米28.35磅*/private static double poundPerCm = 28.35;private static String rgb_black = "000000";/*** 根据磅，获取缇** @return*/private static BigInteger getTwipFromPound(double pound) {Double d = pound * teePerCM / poundPerCm;return BigInteger.valueOf(d.longValue());}/*** 根据厘米，获取缇** @param cm* @return*/private static BigInteger getTwipFromCM(double cm) {Double d = cm * teePerCM;return BigInteger.valueOf(d.longValue());}public static void main(String[] args) throws Exception {
//        buildTable1();buildTable2();log.info("通过html构建docx文件成功！");}public static void buildTable2() throws Exception {String htmlPath = "D:\\work\\project\\jar\\wordtopdf-server\\createWordBytemplate\\src\\test\\java\\simple.html";String templatePath = "D:\\work\\project\\jar\\wordtopdf-server\\createWordBytemplate\\src\\test\\java\\template.docx";String targetPath = "D:\\work\\project\\jar\\wordtopdf-server\\createWordBytemplate\\src\\test\\java\\simple.docx";Document htmlDocument = Jsoup.parse(new File(htmlPath));List<Node> nodeList = htmlDocument.body().childNodes();XWPFDocument document = new XWPFDocument(new FileInputStream(new File(templatePath)));for (Node node : nodeList) {
//            获取出table标签if (node.nodeName().equalsIgnoreCase("table")) {XWPFParagraph paragraph = document.getLastParagraph();parseTableToWord(document, node, paragraph);}}FileOutputStream fileOutputStream = new FileOutputStream(new File(targetPath));document.write(fileOutputStream);fileOutputStream.close();}private static void parseTableToWord(XWPFDocument doc, Node node, XWPFParagraph paragraph) throws Exception {//简化表格htmlString string = node.toString();org.jsoup.nodes.Document tableDoc = Jsoup.parse(string);Elements trList = tableDoc.getElementsByTag("tr");//创建表格XWPFTable xwpfTable = doc.insertNewTbl(paragraph.getCTP().newCursor());if (null == xwpfTable) {return;}//设置样式CTTblPr ctTblPr = xwpfTable.getCTTbl().getTblPr();if (ctTblPr == null) {ctTblPr = xwpfTable.getCTTbl().addNewTblPr();}CTJc jc = ctTblPr.getJc();if (jc == null) {jc = ctTblPr.addNewJc();}//表格位置居中显示jc.setVal(STJc.CENTER);ctTblPr.setJc(jc);//设置表格宽度ctTblPr.getTblW().setType(STTblWidth.PCT);
//        ctTblPr.getTblW().setType(STTblWidth.DXA);//设置表格宽度10cm转换成缇  1cm=567缇ctTblPr.getTblW().setW(getTwipFromCM(10));//设置单元格边框
//        CTTcBorders borders = CTTcBorders.Factory.newInstance();
//        CTBorder border = CTBorder.Factory.newInstance();
//        border.setVal(STBorder.SINGLE);
//        加粗
//        border.setVal(STBorder.THICK);
//       虚线边框
//        border.setVal(STBorder.DOTTED);
//        三层边框
//        border.setVal(STBorder.TRIPLE);
//        波浪
//        border.setVal(STBorder.WAVE);
//        border.setColor(rgb_black);
//        border.setSz(BigInteger.ONE);
//        border.setSpace(BigInteger.ZERO);
//        borders.setLeft(border);
//        borders.setBottom(border);
//        borders.setRight(border);
//        borders.setTop(border);
//        CTTcPr ctTcPr = xwpfTable.getRow(0).getCell(0).getCTTc().getTcPr();
//        if (ctTcPr == null) ctTcPr = xwpfTable.getRow(0).getCell(0).getCTTc().addNewTcPr();
//        ctTcPr.setTcBorders(borders);//写入表格行和列内容for (int row = 0; row < trList.size(); row++) {XWPFTableRow tableRow = xwpfTable.getRow(row);if (null == tableRow) {tableRow = xwpfTable.createRow();}Element trElement = trList.get(row);
//            先后获取表头<th></th>Elements tds = trElement.getElementsByTag("th");if (CollectionUtils.isEmpty(tds)) {
//                 如果th为空说明该行是表内容 <td></td>tds = trElement.getElementsByTag("td");}double widthTotal = 0.0;
//            操作列for (int col = 0; col < tds.size(); col++) {
//                获取当前列对象XWPFTableCell tableCell = tableRow.getCell(col);if (null == tableCell) {tableCell = tableRow.createCell();}//设置单元格边框样式CTTcPr tcPr = tableCell.getCTTc().isSetTcPr() ? tableCell.getCTTc().getTcPr() : tableCell.getCTTc().addNewTcPr();CTTcBorders ctTcBorders = tcPr.addNewTcBorders();ctTcBorders.addNewLeft().setVal(STBorder.SINGLE);ctTcBorders.addNewRight().setVal(STBorder.SINGLE);ctTcBorders.addNewTop().setVal(STBorder.SINGLE);ctTcBorders.addNewBottom().setVal(STBorder.SINGLE);Element colElement = tds.get(col);List<Node> nodes = colElement.childNodes();for (Node tdNode : nodes) {if ("#text".equalsIgnoreCase(tdNode.nodeName())) {//将每一个的内容当成一个段落进行编辑之后再填充到table中parseHtmlToWordTable(colElement, doc, tableCell.getParagraphs().get(0));}}}}}static void parseHtmlToWordTable(Node node, XWPFDocument doc, XWPFParagraph xwpfParagraph) {List<Node> nodes = node.childNodes();if (CollectionUtils.isNotEmpty(nodes)) {for (Node childNode : nodes) {parseHtmlToWordTable(childNode, doc, xwpfParagraph);}}//处理table标签if ("table".equalsIgnoreCase(node.nodeName())) {return;}if (CollectionUtils.isNotEmpty(node.childNodes())) {return;}String nodeValue = node.toString();Node parent = node.parent();boolean boldFlag = false;String color = "";boolean subFlag = false;boolean supFlag = false;boolean ulFlag = false;if (null != parent) {String parentNodeName = parent.nodeName();if (parentNodeName.equalsIgnoreCase("strong") || parentNodeName.equalsIgnoreCase("b")) {boldFlag = true;} else if (parentNodeName.equalsIgnoreCase("font")) {if (Objects.requireNonNull(parent.parent()).nodeName().equalsIgnoreCase("strong")|| Objects.requireNonNull(parent.parent()).nodeName().equalsIgnoreCase("b")) {boldFlag = true;}String color1 = parent.attr("color");if (!StringUtils.isEmpty(color1)) {color = color1.substring(1);}Node parented = parent.parent();if (null != parented) {if (parented.nodeName().equalsIgnoreCase("li")) {if (Objects.requireNonNull(parented.parent()).nodeName().equalsIgnoreCase("ul")) {ulFlag = true;}}}} else if (parentNodeName.equalsIgnoreCase("sub")) {subFlag = true;} else if (parentNodeName.equalsIgnoreCase("sup")) {supFlag = true;} else if (parentNodeName.equalsIgnoreCase("li")) {if (Objects.requireNonNull(parent.parent()).nodeName().equalsIgnoreCase("ul")) {ulFlag = true;}}}if ("#text".equalsIgnoreCase(node.nodeName()) && !nodeValue.contains("<")) {XWPFRun run = xwpfParagraph.createRun();run.setFontFamily("Times New Roman");run.setFontSize(10);if (boldFlag) {run.setBold(true);}if (StringUtils.isEmpty(color)) {run.setColor(color);}if (supFlag) {run.setSubscript(VerticalAlign.SUPERSCRIPT);}if (subFlag) {run.setSubscript(VerticalAlign.SUBSCRIPT);}if (ulFlag && StringUtils.isEmpty(nodeValue.trim())) {XWPFParagraph paragraph = doc.createParagraph();paragraph.setIndentFromLeft(0);paragraph.setFirstLineIndent(0);
//                paragraph.setIndentationLeftChars(125);XWPFRun run1 = paragraph.createRun();run1.setFontFamily("Times New Roman");run1.setFontSize(8);run1.setText("●");run1.addTab();XWPFRun run2 = paragraph.createRun();run2.setText(nodeValue.trim());run2.setFontFamily("Times New Roman");run2.setFontSize(10);}if (StringUtils.isEmpty(nodeValue) && !ulFlag) {run.setText(nodeValue.trim());}}//        boolean enabledBreak = ReUtil.isMatch("(|h[12345]|li|img|br)", node.nodeName().toLowerCase());boolean enabledBreak = true;if (enabledBreak) {XWPFRun run = xwpfParagraph.createRun();run.addCarriageReturn();}}/****//*  public static void buildTable1() throws IOException {String htmlPath = "D:\\work\\project\\jar\\wordtopdf-server\\createWordBytemplate\\src\\test\\java\\simple.html";String templatePath = "D:\\work\\project\\jar\\wordtopdf-server\\createWordBytemplate\\src\\main\\resources\\template.docx";String targetPath = "D:\\work\\project\\jar\\wordtopdf-server\\createWordBytemplate\\src\\test\\java\\simple.docx";Document htmlDocument = Jsoup.parse(new File(htmlPath));List<Node> nodeList = htmlDocument.body().childNodes();List<Node> table = nodeList.stream().filter(o -> o.nodeName().equalsIgnoreCase("table")).collect(Collectors.toList());XWPFTemplate template = XWPFTemplate.compile(templatePath);for (Node node : table) {Map map = new HashMap<>();Elements rows = ((Element) node).getElementsByTag("tr");if (CollectionUtils.isEmpty(rows)) {continue;}RowRenderData[] rowArr = new RowRenderData[rows.size()];for (int row = 0; row < rows.size(); row++) {Elements cols = rows.get(row).getElementsByTag("th");if (CollectionUtils.isEmpty(cols)) {cols = rows.get(row).getElementsByTag("td");}String[] colsAtrr = new String[cols.size()];for (int col = 0; col < cols.size(); col++) {colsAtrr[col] = cols.get(col).text();}rowArr[row] = Rows.of(colsAtrr).create();}TableRenderData tableRenderData = Tables.create(rowArr);map.put("table", tableRenderData);template.render(map);}template.writeAndClose(Files.newOutputStream(Paths.get(targetPath)));}*/}