切换到kvstore

ai-chat-service/chat-server/server/app.go

@@ -157,6 +157,10 @@ func (a *app) countMessageTokens(role, content string) (int, error) {
 	return tokenizer.GetTokens(&message, a.openaiConf.Model)
 }
 
+func (a *app) buildZeroUsage() *proto.Usage {
+	return &proto.Usage{}
+}
+
 func (a *app) buildUsage(promptTokens int, answer string) (*proto.Usage, error) {
 	completionTokens := 0
 	if answer != "" {

ai-chat-service/chat-server/server/server.go

@@ -83,12 +83,7 @@ func (s *chatService) ChatCompletion(ctx context.Context, in *proto.ChatCompleti
 	questionEmbedding, cachedRecord := s.searchCachedAnswer(ctx, in.Message)
 	if cachedRecord != nil {
 		res := app.buildChatCompletionResponse(cachedRecord.Answer)
-		usage, tokenErr := app.buildUsage(currTokens, cachedRecord.Answer)
-		if tokenErr != nil {
-			s.log.Error(tokenErr)
-		} else {
-			res.Usage = usage
-		}
+		res.Usage = app.buildZeroUsage()
 		res.Source = replySourceSemanticMatch
 		return res, nil
 	}
@@ -194,12 +189,7 @@ func (s *chatService) ChatCompletionStream(in *proto.ChatCompletionRequest, stre
 		}
 		final := app.buildChatCompletionStreamResponse(cachedRecord.ID, "", "stop")
 		final.Source = replySourceSemanticMatch
-		usage, tokenErr := app.buildUsage(currTokens, cachedRecord.Answer)
-		if tokenErr != nil {
-			s.log.Error(tokenErr)
-		} else {
-			final.Usage = usage
-		}
+		final.Usage = app.buildZeroUsage()
 		return stream.Send(final)
 	}