...

src/main/kotlin/network/RagService.kt

@@ -19,6 +19,7 @@ import dev.langchain4j.model.openai.OpenAiEmbeddingModel
 import dev.langchain4j.service.AiServices
 import dev.langchain4j.service.SystemMessage
 import dev.langchain4j.store.embedding.EmbeddingSearchRequest
+import dev.langchain4j.store.embedding.EmbeddingSearchResult
 import dev.langchain4j.store.embedding.filter.MetadataFilterBuilder
 import kotlinx.coroutines.Dispatchers
 import kotlinx.coroutines.async
@@ -298,20 +299,23 @@ object RagService {
 
                 val question = "$stockName 실적 및 향후 전망"
                 val questionEmbedding = embeddingModel.embed(question).content()
+                var finalSearchResult : EmbeddingSearchResult<TextSegment>? = null
+                try {
+                    finalSearchResult = embeddingStore.search(
+                        EmbeddingSearchRequest.builder()
+                            .queryEmbedding(questionEmbedding)
+                            .filter(MetadataFilterBuilder.metadataKey("stockCode").isEqualTo(stockCode))
+                            .maxResults(10) // 최신 뉴스 3개 적정
+                            .minScore(0.2)
+                            .build()
+                    )
+                } catch (e: Exception) {}
 
-                val finalSearchResult = embeddingStore.search(
-                    EmbeddingSearchRequest.builder()
-                        .queryEmbedding(questionEmbedding)
-                        .filter(MetadataFilterBuilder.metadataKey("stockCode").isEqualTo(stockCode))
-                        .maxResults(10) // 최신 뉴스 3개 적정
-                        .minScore(0.2)
-                        .build()
-                )
 
                 // 3. 검색된 내용을 하나의 문자열로 합쳐서 전달
-                tradingDecision.newsContext = finalSearchResult.matches().distinct() // 중복 제거
-                    .take(4)    // 10개에서 4개로 축소
-                    .joinToString("\n\n")  {
+                tradingDecision.newsContext = finalSearchResult?.matches()?.distinct() // 중복 제거
+                    ?.take(4)    // 10개에서 4개로 축소
+                    ?.joinToString("\n\n")  {
                         it.embedded().text()
                     }
 
@@ -336,20 +340,25 @@ object RagService {
         val questionEmbedding = embeddingModel.embed(question).content()
 
         // 1. 벡터 DB에서 해당 종목의 뉴스 검색
-        val searchResult = embeddingStore.search(
-            EmbeddingSearchRequest.builder()
-                .queryEmbedding(questionEmbedding)
-                .filter(MetadataFilterBuilder.metadataKey("stockCode").isEqualTo(stockCode))
-                .maxResults(10)
-                .minScore(0.2)
-                .build()
-        )
+        var searchResult : EmbeddingSearchResult<TextSegment>? = null
+        try {
+             searchResult = embeddingStore.search(
+                EmbeddingSearchRequest.builder()
+                    .queryEmbedding(questionEmbedding)
+                    .filter(MetadataFilterBuilder.metadataKey("stockCode").isEqualTo(stockCode))
+                    .maxResults(10)
+                    .minScore(0.2)
+                    .build()
+            )
+        } catch (e: Exception) {
+
+        }
 
         // 2. 검색된 뉴스 중 1시간 이내(Very Recent) 데이터가 있는지 확인
-        val hasHotNews = searchResult.matches().any { match ->
+        val hasHotNews = searchResult?.matches()?.any { match ->
             val pubDate = match.embedded().metadata().getString("date")
             isVeryRecentNews(pubDate, maxHours = 1)
-        }
+        } ?: false
 
         // 3. 최신 뉴스가 없다면 네이버 API 및 Playwright 스크래핑 가동
         if (!hasHotNews) {