diff --git a/laws-modules/src/main/java/com/jero/modules/document/utils/ReadPdfUtil.java b/laws-modules/src/main/java/com/jero/modules/document/utils/ReadPdfUtil.java index 62ec1a74..dcb9f7bc 100644 --- a/laws-modules/src/main/java/com/jero/modules/document/utils/ReadPdfUtil.java +++ b/laws-modules/src/main/java/com/jero/modules/document/utils/ReadPdfUtil.java @@ -2,35 +2,47 @@ package com.jero.modules.document.utils; import com.jero.common.exception.JeroBootException; import com.jero.common.util.MinioUtil; +import lombok.extern.slf4j.Slf4j; import org.apache.pdfbox.pdmodel.PDDocument; import org.apache.pdfbox.text.PDFTextStripper; import org.apache.pdfbox.text.PDFTextStripperByArea; import java.io.IOException; import java.io.InputStream; +import java.util.Objects; /** * @description * @date 2022/3/21 17:04 * @auth zhn */ +@Slf4j public class ReadPdfUtil { public static String readPdf(String path) { if(MinioUtil.doesObjectExist(path)){ - try (InputStream in = MinioUtil.download(path); - PDDocument document = PDDocument.load(in)) { - document.getClass(); - PDFTextStripperByArea stripper = new PDFTextStripperByArea(); - stripper.setSortByPosition(true); - PDFTextStripper tStripper = new PDFTextStripper(); + try (InputStream in = MinioUtil.download(path)) { + if(Objects.isNull(in) || in.available() == 0){ + return null; + } + try (PDDocument document = PDDocument.load(in)) { + document.getClass(); + PDFTextStripperByArea stripper = new PDFTextStripperByArea(); + stripper.setSortByPosition(true); + PDFTextStripper tStripper = new PDFTextStripper(); - String pdfFileInText = tStripper.getText(document); - return pdfFileInText; + String pdfFileInText = tStripper.getText(document); + return pdfFileInText; - }catch (IOException e) { - throw new JeroBootException("文件内容读取失败"); + }catch (Exception e) { + log.error("文件内容读取失败",e); +// throw new JeroBootException("文件内容读取失败"); + } + }catch (Exception e) { + log.error("文件内容读取失败",e); +// throw new JeroBootException("文件内容读取失败"); } + } return null; } diff --git a/laws-modules/src/main/java/com/jero/modules/laws/home/job/HomeSearchCheckEsPlanJob.java b/laws-modules/src/main/java/com/jero/modules/laws/home/job/HomeSearchCheckEsPlanJob.java index 38715cb3..920a4357 100644 --- a/laws-modules/src/main/java/com/jero/modules/laws/home/job/HomeSearchCheckEsPlanJob.java +++ b/laws-modules/src/main/java/com/jero/modules/laws/home/job/HomeSearchCheckEsPlanJob.java @@ -56,6 +56,7 @@ public class HomeSearchCheckEsPlanJob implements Job { @Override public void execute(JobExecutionContext context) throws JobExecutionException { + log.info("============ es同步任务开始=============="); // 获取外部传来的参数 JobDataMap map = context.getJobDetail().getJobDataMap(); String resourceType = map.getString("parameter"); @@ -92,6 +93,7 @@ public class HomeSearchCheckEsPlanJob implements Job { default:break; } }); + log.info("============ es同步任务结束=============="); } private void checkLawsNewsFeed(String resourceType, Map mapHash) { @@ -102,8 +104,6 @@ public class HomeSearchCheckEsPlanJob implements Job { m4 = listLawsNewsFeed.size(); } if(count4 != m4 || !StringUtils.isBlank(resourceType)){ - // 清理数据重新更新 - delete(resourceType); if(!CollectionUtils.isEmpty(listLawsNewsFeed)){ List listIds = listLawsNewsFeed.stream().map(LawsNewsFeed::getId).collect(Collectors.toList()); addES(listIds, LawsHomeSearchCommon.LAWS_NEWS_FEED); @@ -119,8 +119,6 @@ public class HomeSearchCheckEsPlanJob implements Job { m3 = listLawsEnterpriseStandard.size(); } if(count3 != m3 || !StringUtils.isBlank(resourceType)){ - // 清理数据重新更新 - delete(resourceType); if(!CollectionUtils.isEmpty(listLawsEnterpriseStandard)){ List listIds = listLawsEnterpriseStandard.stream().map(LawsEnterpriseStandard::getId).collect(Collectors.toList()); addES(listIds, LawsHomeSearchCommon.LAWS_ENTERPRISE_STANDARD); @@ -136,8 +134,6 @@ public class HomeSearchCheckEsPlanJob implements Job { m2 = listLawsOverseasStandard.size(); } if(count2 != m2 || !StringUtils.isBlank(resourceType)){ - // 清理数据重新更新 - delete(resourceType); if(!CollectionUtils.isEmpty(listLawsOverseasStandard)){ List listIds = listLawsOverseasStandard.stream().map(LawsOverseasStandard::getId).collect(Collectors.toList()); addES(listIds, LawsHomeSearchCommon.LAWS_OVERSEAS_STANDARD); @@ -154,8 +150,6 @@ public class HomeSearchCheckEsPlanJob implements Job { m1 = listLawsDomesticStandard.size(); } if(count1 != m1 || !StringUtils.isBlank(resourceType)){ - // 清理数据重新更新 - delete(resourceType); if(!CollectionUtils.isEmpty(listLawsDomesticStandard)){ List listIds = listLawsDomesticStandard.stream().map(LawsDomesticStandard::getId).collect(Collectors.toList()); addES(listIds, LawsHomeSearchCommon.LAWS_DOMESTIC_STANDARD); diff --git a/laws-modules/src/main/java/com/jero/modules/laws/home/service/impl/LawsHomeSearchServiceImpl.java b/laws-modules/src/main/java/com/jero/modules/laws/home/service/impl/LawsHomeSearchServiceImpl.java index 6399bd0a..f9e31f6a 100644 --- a/laws-modules/src/main/java/com/jero/modules/laws/home/service/impl/LawsHomeSearchServiceImpl.java +++ b/laws-modules/src/main/java/com/jero/modules/laws/home/service/impl/LawsHomeSearchServiceImpl.java @@ -1044,7 +1044,7 @@ public class LawsHomeSearchServiceImpl implements ILawsHomeSearchService { // 批量操作 // 从第0个下标开始 int startIndex = 0; - int batchCount = 2000; + int batchCount = 500; List listIds = Arrays.asList(ids.split(",")); ForkJoinPool pool = new ForkJoinPool(); if(CollectionUtils.isEmpty(listIds)) { @@ -1368,6 +1368,7 @@ public class LawsHomeSearchServiceImpl implements ILawsHomeSearchService { public void updateElasticsearch(LawsHomeNormalSearchVO obj,List listFile){ String idTemp = obj.getId(); + System.out.println("--------------" + obj); StringBuilder fileTextAll = getStringBuilder(listFile); addES(obj, idTemp, fileTextAll); }