'commit'

src/main/java/com/dsideal/base/Tools/ReadCityDoc.java

@@ -1,93 +1,132 @@
 package com.dsideal.base.Tools;
 
-import cn.hutool.core.io.FileUtil;
+import com.dsideal.base.DataEase.Model.DataEaseModel;
+import com.dsideal.base.Tools.Util.LocalMysqlConnectUtil;
 import org.apache.poi.openxml4j.exceptions.InvalidFormatException;
 import org.apache.poi.openxml4j.util.ZipSecureFile;
-import org.apache.poi.ss.usermodel.CellType;
-import org.apache.poi.ss.usermodel.CellValue;
-import org.apache.poi.ss.usermodel.FormulaEvaluator;
 import org.apache.poi.xssf.usermodel.*;
 import org.apache.poi.xwpf.usermodel.XWPFChart;
 import org.apache.poi.xwpf.usermodel.XWPFDocument;
+import org.apache.poi.xwpf.usermodel.XWPFFootnote;
 import org.apache.poi.xwpf.usermodel.XWPFParagraph;
 
-import java.io.File;
-import java.io.FileInputStream;
-import java.io.IOException;
-import java.io.InputStream;
+import java.io.*;
+import java.util.ArrayList;
 import java.util.List;
 
+import com.jfinal.plugin.activerecord.Record;
+
 public class ReadCityDoc {
-    //https://blog.csdn.net/a346736962/article/details/123037797
+    public static DataEaseModel dm = new DataEaseModel();
+
+    /**
+     * 获取云南省下所有城市名称
+     *
+     * @return
+     */
+    public static List<String> getCityNameList() {
+        //获取云南省下的所有市州
+        String privinceName = "云南省";
+        String privinceId = dm.getAreaByName(privinceName).getStr("id");
+        List<Record> listCity = dm.getAreaList(privinceId);
+        List<String> listCityName = new ArrayList<>();
+        for (Record record : listCity) {
+            listCityName.add(record.getStr("area_name"));
+        }
+        return listCityName;
+    }
+
+    /**
+     * 根据文件名获取城市名称
+     *
+     * @param fileName
+     * @return
+     */
+    public static String getCityName(String fileName) {
+        //云南省下所有城市
+        List<String> listCity = getCityNameList();
+        //判断是哪个城市
+        boolean found = false;
+        String cityName = "";
+        for (String s : listCity) {
+            if (fileName.contains(s)) {
+                found = true;
+                cityName = s;
+            }
+        }
+        if (!found) {
+            System.out.println("未找到匹配的城市");
+            System.exit(0);
+        }
+        return cityName;
+    }
+
     public static void main(String[] args) throws IOException, InvalidFormatException {
-        String directoryPath = "D:\\dsWork\\YunNanDsBase\\Doc\\全省及州市县区人口与教育报告集20241023\\16个州市报告2022\\分析报告20240510";
-        directoryPath="D:\\dsWork\\YunNanDsBase\\Doc\\全省及州市县区人口与教育报告集20241023\\133个县区报告2022";
-        //遍历workingPath下的所有文件,注意，需要递归所有子目录下的所有文件
-        System.out.println("开始遍历目录下的所有文件");
-        // 调用 Hutool 的 FileUtil.loopFiles 方法递归获取所有文件
-        List<File> files = FileUtil.loopFiles(directoryPath, file -> {
-            // 这里可以添加你的过滤条件，如果不需要过滤，返回 true 即可
-            return true;
-        });
-        int cnt = 0;
-        // 打印所有文件的路径
-        for (File file : files) {
-            String fileName = file.getName();
-            //判断是否为docx文件
-            if (fileName.endsWith(".docx") && !fileName.startsWith("~")) {
-                cnt++;
-                //读取文件
-                String inputUrl = file.getAbsolutePath();
-                InputStream is = new FileInputStream(inputUrl);
-                //读取excel报错 Zip bomb detected! The file would exceed the max.
-                //https://blog.csdn.net/baidu_19473529/article/details/109601558
-                ZipSecureFile.setMinInflateRatio(-1.0d);
-                XWPFDocument doc = new XWPFDocument(is);
-                //图表
-                List<XWPFChart> charts = doc.getCharts();
+        //初始化数据库连接
+        LocalMysqlConnectUtil.Init();
 
-                System.out.println("图表数量=" + charts.size());
 
-                for (XWPFChart chart : charts) {
-                    XSSFWorkbook workbook = chart.getWorkbook();
-                    XSSFSheet sheet = workbook.getSheetAt(0);
-                    //遍历一下sheet
-                    for (int i = 0; i < sheet.getPhysicalNumberOfRows(); i++) {
-                        //遍历行
-                        XSSFRow row = sheet.getRow(i);
-                        if (row == null) continue;
-                        for (int j = 0; j < row.getPhysicalNumberOfCells(); j++) {
-                            //遍历列
-                            XSSFCell cell = row.getCell(j);
-                            // 创建公式计算器
-                            FormulaEvaluator evaluator = new XSSFFormulaEvaluator(workbook);
+        //开始读取word文档
+        String parentPath = "D:\\dsWork\\YunNanDsBase\\Doc\\全省及州市县区人口与教育报告集20241023\\16个州市报告2022";
+        //在parentPath下检查是不是存在Excel的目录,如果不存在则创建
+        File excelDir = new File(parentPath + "\\Excel");
+        if (!excelDir.exists()) {
+            excelDir.mkdir();
+        }
 
-                            // 检查单元格是否包含公式
-                            if (cell!=null && cell.getCellType() == CellType.FORMULA) {
-                                // 计算公式并获取结果
-                                CellValue evaluatedValue = evaluator.evaluate(cell);
-                                System.out.println("Calculated Value: " + evaluatedValue.formatAsString());
-                            }
-                            if (cell != null)
-                                System.out.print(cell + " ");
-                            else {
-                                System.out.print("null ");
-                            }
-                        }
-                        System.out.println();
-                    }
-                }
-                //段落
-                List<XWPFParagraph> paragraphs = doc.getParagraphs();
-                for (XWPFParagraph paragraph : paragraphs) {
-                    String text = paragraph.getText();
-                    if (text.startsWith("图")) {
-                        System.out.println(text);
-                    }
+        String fileStr = parentPath + "\\分析报告20240510\\昆明市人口变化及其对教育的影响20240419.docx";
+        File file = new File(fileStr);
+        String fileName = file.getName();
+        //城市名称
+        String cityName = getCityName(fileName);
+        //在excelDir下检查是不是存在cityName的文件夹，不存在则创建
+        File cityDir = new File(excelDir.getAbsolutePath() + "\\" + cityName);
+        if (!cityDir.exists()) {
+            cityDir.mkdir();
+        }
 
+        //判断是否为docx文件
+        if (fileName.endsWith(".docx") && !fileName.startsWith("~")) {
+            //读取文件
+            String inputUrl = file.getAbsolutePath();
+            InputStream is = new FileInputStream(inputUrl);
+            ZipSecureFile.setMinInflateRatio(-1.0d);
+            XWPFDocument doc = new XWPFDocument(is);
+            //段落
+            int num = 0;
+            List<String> tbList = new ArrayList<>();
+            List<XWPFParagraph> paragraphs = doc.getParagraphs();
+            for (XWPFParagraph paragraph : paragraphs) {
+                String text = paragraph.getText();
+                // 检查段落是否包含脚注
+                String footnotes = paragraph.getFootnoteText();
+                //删除掉text中脚注的文字
+                text = text.replace(footnotes, "");
+                text=text.replace("[footnoteRef:0]","");
+                if (text.startsWith("图")) {
+                    num++;
+                    String[] a = text.split(" ");
+                    //最后一个字符串
+                    String lastStr = "【" + num + "】" + a[a.length - 1];
+                    tbList.add(lastStr);
                 }
             }
+            //图表
+            List<XWPFChart> charts = doc.getCharts();
+            //System.out.println("图表数量=" + charts.size() + ",图例个数=" + num);
+            for (int i = 0; i < charts.size(); i++) {
+                XWPFChart chart = charts.get(i);
+                XSSFWorkbook workbook = chart.getWorkbook();
+                // 创建FileOutputStream对象，用于写入新的Excel文件
+                FileOutputStream outputStream = new FileOutputStream(cityDir.getAbsolutePath() + "\\" + tbList.get(i) + ".xlsx");
+                // 将工作簿写入到输出流中
+                workbook.write(outputStream);
+                // 关闭输出流和工作簿
+                outputStream.close();
+                workbook.close();
+                //输出完成信息
+                System.out.println("输出" + tbList.get(i) + ".xlsx完成！");
+            }
         }
-        System.out.println("共读取" + cnt + "个文件");
     }
 }