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)
+		res.Usage = app.buildZeroUsage()
-		if tokenErr != nil {
-			s.log.Error(tokenErr)
-		} else {
-			res.Usage = usage
-		}
 		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)
+		final.Usage = app.buildZeroUsage()
-		if tokenErr != nil {
-			s.log.Error(tokenErr)
-		} else {
-			final.Usage = usage
-		}
 		return stream.Send(final)
 	}
 

ai-chat-stack/compose.yaml

@@ -13,18 +13,18 @@ services:
       - FAISS_DIM=1024
       - APP_PORT=8000
 
-  redis:
+  # redis:
-    image: redis:7-alpine
+  #   image: redis:7-alpine
-    container_name: ai-chat-redis
+  #   container_name: ai-chat-redis
-    command:
+  #   command:
-      - redis-server
+  #     - redis-server
-      - --save
+  #     - --save
-      - ""
+  #     - ""
-      - --appendonly
+  #     - --appendonly
-      - "no"
+  #     - "no"
-    ports:
+  #   ports:
-      - "8888:6379"
+  #     - "8888:6379"
-    restart: unless-stopped
+  #   restart: unless-stopped
 
   tokenizer:
     build:

ai-chat-stack/configs/ai-chat-service.yaml

@@ -17,8 +17,8 @@ chat:
   bot_desc: "你是一个AI助手，我需要你模拟一名资深的软件工程师来回答我的问题"
   min_response_tokens: 600
 redis:
-  host: "redis"
+  host: "127.0.0.1"
-  port: 6379
+  port: 8888
   pwd: ""
 dependOn:
   sensitive: