Merge branch 'wyp/0108-similarStatusFilter' of Server/long-article-recommend into master

long-article-recommend-service/src/main/java/com/tzld/longarticle/recommend/server/model/entity/longArticle/LongArticlesText.java

@@ -31,4 +31,6 @@ public class LongArticlesText {
     private String kimiKeys;
     @Column(name = "kimi_status")
     private Integer kimiStatus;
+    @Column(name = "similarity_status")
+    private Integer similarityStatus;
 }

long-article-recommend-service/src/main/java/com/tzld/longarticle/recommend/server/service/recommend/recall/RecallService.java

@@ -2,6 +2,8 @@ package com.tzld.longarticle.recommend.server.service.recommend.recall;
 
 import com.alibaba.fastjson.JSONArray;
 import com.alibaba.fastjson.JSONObject;
+import com.google.common.cache.Cache;
+import com.google.common.cache.CacheBuilder;
 import com.google.common.collect.Lists;
 import com.tzld.longarticle.recommend.server.common.CostMonitor;
 import com.tzld.longarticle.recommend.server.common.ThreadPoolFactory;
@@ -96,11 +98,15 @@ public class RecallService implements ApplicationContextAware {
     ArticleTitleHisCacheRepository articleTitleHisCacheRepository;
     @Autowired
     LongArticleBaseMapper longArticleBaseMapper;
+    @Autowired
+    LongArticlesTextRepository longArticlesTextRepository;
 
     private final Map<String, RecallStrategy> strategyMap = new HashMap<>();
     private ApplicationContext applicationContext;
     private final ExecutorService pool = ThreadPoolFactory.recallPool();
 
+    Cache<String, Integer> textCache = CacheBuilder.newBuilder().expireAfterWrite(60 * 10, TimeUnit.SECONDS).build();
+
     @Value("${recall.content.his.feishu.enable:false}")
     private Boolean contentHisFeishuEnable;
     @Value("${morning.noon.fission.rate:0.64}")
@@ -174,6 +180,8 @@ public class RecallService implements ApplicationContextAware {
         List<Content> content = aigcWaitingPublishContentService.getAllContent(param);
         long t2 = System.currentTimeMillis();
         CostMonitor.logCost("Recall", "GetAllContents", t2 - t1);
+        // 临时过滤文章视频不匹配content
+        filterNotMatchContent(content);
         if (CollectionUtils.isEmpty(content)) {
             FeishuMessageSender.sendWebHookMessage(FeishuRobotIdEnum.RECOMMEND.getRobotId(),
                     "内容召回失败\n"
@@ -192,6 +200,25 @@ public class RecallService implements ApplicationContextAware {
         return content;
     }
 
+    private void filterNotMatchContent(List<Content> content) {
+        if (CollectionUtils.isEmpty(content)) {
+            return;
+        }
+        List<String> searchIds = content.stream()
+                .filter(o ->!Objects.equals(o.getSourceType(), PublishPlanInputSourceTypesEnum.longArticleVideoPoolSource.getVal()))
+                .map(Content::getSourceId).filter(o -> Objects.isNull(textCache.getIfPresent(o)))
+                .collect(Collectors.toList());
+        if (CollectionUtils.isNotEmpty(searchIds)) {
+            List<LongArticlesText> textList = longArticlesTextRepository.getByContentIdIn(searchIds);
+            Map<String, Integer> textMap = textList.stream()
+                    .collect(Collectors.toMap(LongArticlesText::getContentId, LongArticlesText::getSimilarityStatus));
+            textCache.putAll(textMap);
+        }
+        content.removeIf(o -> !Objects.equals(o.getSourceType(), PublishPlanInputSourceTypesEnum.longArticleVideoPoolSource.getVal())
+                && (Objects.isNull(textCache.getIfPresent(o.getSourceId()))
+                || textCache.getIfPresent(o.getSourceId()) == 0));
+    }
+
     private void setVideoContent(List<Content> contentList) {
         List<String> contentTraceIds = contentList.stream()
                 .filter(o -> PublishPlanInputSourceTypesEnum.longArticleVideoPoolSource.getVal().equals(o.getSourceType()))