Merge branch 'caihaohan'

adc-da-report/src/main/java/com/adc/da/report/service/impl/IReportContentServiceImpl.java

@@ -109,7 +109,7 @@ public class IReportContentServiceImpl extends ServiceImpl<ReportContentDao, Rep
         //将内容存入
         ReportContentEntity reportContent = query().eq("report_id", reportId).one();
         if (reportContent == null) {
-            ReportContentEntity reportContentEntity = new ReportContentEntity(UUID.randomUUID10(), reportId, reportContentFileIds,content.toString());
+            ReportContentEntity reportContentEntity = new ReportContentEntity(UUID.randomUUID10(), reportId, reportContentFileIds, content.toString());
             save(reportContentEntity);
         } else {
             reportContent.setReportContent(content.toString());

adc-da-report/src/main/java/com/adc/da/report/service/impl/IReportServiceImpl.java

@@ -262,8 +262,8 @@ public class IReportServiceImpl extends ServiceImpl<ReportDao, ReportEntity>
             //修改报表内容表(先删后增)
             iReportContentService.removeByReportId(reportId);
             try {
-                iReportContentService.insertReportContent(reportId);
                 log.info("抽取reportId为{}报告内容", reportId);
+                iReportContentService.insertReportContent(reportId);
                 count++;
             } catch (IOException e) {
                 log.error("获取reportId为{}的报告内容失败", reportId);

adc-da-report/src/main/java/com/adc/da/report/util/FileContentUtil.java

@@ -3,6 +3,8 @@ package com.adc.da.report.util;
 import java.io.File;
 import java.io.FileInputStream;
 import java.io.IOException;
+import java.nio.charset.CharsetEncoder;
+import java.nio.charset.StandardCharsets;
 
 import com.adc.da.report.constant.ReportConstants;
 import lombok.extern.slf4j.Slf4j;
@@ -92,9 +94,21 @@ public class FileContentUtil {
             fis.close();
             log.info("文件内容抽取失败，原因可能是上传的文件为空文件");
         }
-        return content.toString();
+        return cleanInvalidCharacters(content.toString());
     }
 
+    public static String cleanInvalidCharacters(String input) {
+        StringBuilder stringBuilder = new StringBuilder();
+        CharsetEncoder encoder = StandardCharsets.UTF_8.newEncoder();
+        for (char ch : input.toCharArray()) {
+            if (encoder.canEncode(ch)) {
+                stringBuilder.append(ch);
+            }
+        }
+        return stringBuilder.toString();
+    }
+
+
     private static String readExcelContent(org.apache.poi.ss.usermodel.Workbook workbook) {
         StringBuilder sb = new StringBuilder();
         int numberOfSheets = workbook.getNumberOfSheets();