service 修改 Redis 存储 KV

ai-chat-stack/compose.yaml

@@ -1,35 +1,4 @@
 services:
-  mysql:
-    image: mysql:8.0
-    container_name: ai-chat-mysql
-    restart: unless-stopped
-    environment:
-      MYSQL_ROOT_PASSWORD: root
-    command:
-      - --default-authentication-plugin=mysql_native_password
-    volumes:
-      - /data/mysql:/var/lib/mysql
-      - /home/lian/share/aichat/init/create_db.sql:/docker-entrypoint-initdb.d/create_db.sql:ro
-    healthcheck:
-      test: ["CMD", "mysqladmin", "ping", "-h", "127.0.0.1", "-proot"]
-      interval: 15s
-      timeout: 5s
-      retries: 10
-
-  pgvector:
-    image: pgvector/pgvector:pg16
-    container_name: ai-chat-pgvector
-    restart: unless-stopped
-    environment:
-      POSTGRES_DB: ai_chat
-      POSTGRES_USER: postgres
-      POSTGRES_PASSWORD: postgres
-    ports:
-      - "15432:5432"
-    volumes:
-      - /data/pgvector:/var/lib/postgresql/data
-      - /home/lian/share/aichat/init/pgvector-init.sql:/docker-entrypoint-initdb.d/pgvector-init.sql:ro
-
   tokenizer:
     build:
       context: ../tokenizer
@@ -83,11 +52,9 @@ services:
     ports:
       - "50055:50055"
     depends_on:
-      - mysql
       - tokenizer
       - sensitive-filter
       - keywords-filter
-      - pgvector
     healthcheck:
       test: ["CMD", "grpc_health_probe", "-addr=:50055"]
       interval: 15s

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

@@ -16,17 +16,10 @@ chat:
   frequency_penalty: 0
   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
-  context_ttl: 1800
-  context_len: 4
 redis:
   host: "host.docker.internal"
   port: 8888
   pwd: "123456"
-mysql:
-  dsn: "root:root@tcp(mysql:3306)/ai_chat?collation=utf8mb4_unicode_ci&charset=utf8mb4"
-  maxLifeTime: 3600
-  maxOpenConn: 10
-  maxIdleConn: 10
 dependOn:
   sensitive:
     address: "sensitive-filter:50053"
@@ -36,19 +29,14 @@ dependOn:
     accessToken: "ang1chubdev1ozhome256487d22sapguuv1ozhom"
   tokenizer:
     address: "http://tokenizer:3002"
-vector:
-  provider: "pgvector"
-  threshold: 0.99
-  pgvector:
-    dsn: "postgres://postgres:postgres@pgvector:5432/ai_chat?sslmode=disable"
-    table: "chat_record_vectors"
-    dimensions: 1024
-    maxLifeTime: 3600
-    maxOpenConn: 10
-    maxIdleConn: 10
 embedding:
   provider: "openai-compatible"
   base_url: "https://open.bigmodel.cn/api/paas/v4"
   api_key: "__SET_FROM_ENV__"
   model: "embedding-2"
   timeout: 10
+faiss:
+  base_url: "http://host.docker.internal:8451"
+  search_k: 1
+  similarity_threshold: 0.9
+  timeout: 10