Refactor model provider and command

cmd/server_main.go

@@ -176,6 +176,7 @@ func startServer(config *server.Config) {
 	searchService := service.NewSearchService()
 	fileService := service.NewFileService()
 	memoryService := service.NewMemoryService()
+	modelProviderService := service.NewModelProviderService()
 
 	// Initialize handler layer
 	authHandler := handler.NewAuthHandler()
@@ -193,7 +194,7 @@ func startServer(config *server.Config) {
 	searchHandler := handler.NewSearchHandler(searchService, userService)
 	fileHandler := handler.NewFileHandler(fileService, userService)
 	memoryHandler := handler.NewMemoryHandler(memoryService)
-	providerHandler := handler.NewProviderHandler(userService)
+	providerHandler := handler.NewProviderHandler(userService, modelProviderService)
 
 	// Initialize router
 	r := router.NewRouter(authHandler, userHandler, tenantHandler, documentHandler, datasetsHandler, systemHandler, kbHandler, chunkHandler, llmHandler, chatHandler, chatSessionHandler, connectorHandler, searchHandler, fileHandler, memoryHandler, providerHandler)

conf/models/openai.json

@@ -2,6 +2,9 @@
   "name": "OpenAI",
   "tags": "LLM,TEXT EMBEDDING,TTS,TEXT RE-RANK,SPEECH2TEXT,MODERATION",
   "url": "https://api.openai.com/v1",
+  "url_suffix": {
+    "chat": "chat/completions"
+  },
   "models": [
     {
       "name": "gpt-5.2-pro",

conf/models/xai.json

@@ -2,6 +2,9 @@
   "name": "xAI",
   "tags": "LLM",
   "url": "https://api.x.ai/v1",
+  "url_suffix": {
+    "chat": "chat/completions"
+  },
   "models": [
     {
       "name": "grok-4",

conf/models/zhipu-ai.json

@@ -0,0 +1,182 @@
+{
+  "name": "ZHIPU-AI",
+  "tags": "LLM,TEXT EMBEDDING,SPEECH2TEXT,MODERATION",
+  "url": "https://open.bigmodel.cn/api/paas/v4",
+  "url_suffix": {
+    "chat": "chat/completions",
+    "async_chat": "async/chat/completions",
+    "async_result": "async-result",
+    "embedding": "embedding",
+    "rerank": "rerank"
+  },
+  "models": [
+    {
+      "name": "glm-4.7",
+      "max_tokens": 128000,
+      "model_types": [
+        "chat"
+      ],
+      "features": {}
+    },
+    {
+      "name": "glm-4.5",
+      "max_tokens": 128000,
+      "model_types": [
+        "chat"
+      ],
+      "features": {}
+    },
+    {
+      "name": "glm-4.5-x",
+      "max_tokens": 128000,
+      "model_types": [
+        "chat"
+      ],
+      "features": {}
+    },
+    {
+      "name": "glm-4.5-air",
+      "max_tokens": 128000,
+      "model_types": [
+        "chat"
+      ],
+      "features": {}
+    },
+    {
+      "name": "glm-4.5-airx",
+      "max_tokens": 128000,
+      "model_types": [
+        "chat"
+      ],
+      "features": {}
+    },
+    {
+      "name": "glm-4.5-flash",
+      "max_tokens": 128000,
+      "model_types": [
+        "chat"
+      ],
+      "features": {}
+    },
+    {
+      "name": "glm-4.5v",
+      "max_tokens": 64000,
+      "model_types": [
+        "image2text"
+      ],
+      "features": {}
+    },
+    {
+      "name": "glm-4-plus",
+      "max_tokens": 128000,
+      "model_types": [
+        "chat"
+      ],
+      "features": {}
+    },
+    {
+      "name": "glm-4-0520",
+      "max_tokens": 128000,
+      "model_types": [
+        "chat"
+      ],
+      "features": {}
+    },
+    {
+      "name": "glm-4",
+      "max_tokens": 128000,
+      "model_types": [
+        "chat"
+      ],
+      "features": {}
+    },
+    {
+      "name": "glm-4-airx",
+      "max_tokens": 8000,
+      "model_types": [
+        "chat"
+      ],
+      "features": {}
+    },
+    {
+      "name": "glm-4-air",
+      "max_tokens": 128000,
+      "model_types": [
+        "chat"
+      ],
+      "features": {}
+    },
+    {
+      "name": "glm-4-flash",
+      "max_tokens": 128000,
+      "model_types": [
+        "chat"
+      ],
+      "features": {}
+    },
+    {
+      "name": "glm-4-flashx",
+      "max_tokens": 128000,
+      "model_types": [
+        "chat"
+      ],
+      "features": {}
+    },
+    {
+      "name": "glm-4-long",
+      "max_tokens": 1000000,
+      "model_types": [
+        "chat"
+      ],
+      "features": {}
+    },
+    {
+      "name": "glm-3-turbo",
+      "max_tokens": 128000,
+      "model_types": [
+        "chat"
+      ],
+      "features": {}
+    },
+    {
+      "name": "glm-4v",
+      "max_tokens": 2000,
+      "model_types": [
+        "image2text"
+      ],
+      "features": {}
+    },
+    {
+      "name": "glm-4-9b",
+      "max_tokens": 8192,
+      "model_types": [
+        "chat"
+      ],
+      "features": {}
+    },
+    {
+      "name": "embedding-2",
+      "max_tokens": 512,
+      "model_types": [
+        "embedding"
+      ],
+      "features": {}
+    },
+    {
+      "name": "embedding-3",
+      "max_tokens": 512,
+      "model_types": [
+        "embedding"
+      ],
+      "features": {}
+    },
+    {
+      "name": "glm-asr",
+      "max_tokens": 4096,
+      "model_types": [
+        "speech2text"
+      ],
+      "features": {}
+    }
+  ]
+}

internal/cli/admin_parser.go

@@ -275,30 +275,6 @@ func (p *Parser) parseAdminListDefaultModels() (*Command, error) {
 	return NewCommand("list_user_default_models"), nil
 }
 
-func (p *Parser) parseListModelsOfProvider() (*Command, error) {
-	if p.curToken.Type != TokenModels {
-		return nil, fmt.Errorf("expected MODELS")
-	}
-
-	p.nextToken()
-	if p.curToken.Type != TokenFrom {
-		return nil, fmt.Errorf("expected FROM")
-	}
-	p.nextToken()
-	providerName, err := p.parseQuotedString()
-	if err != nil {
-		return nil, err
-	}
-	cmd := NewCommand("list_provider_models")
-	cmd.Params["provider_name"] = providerName
-	p.nextToken()
-	// Semicolon is optional for UNSET TOKEN
-	if p.curToken.Type == TokenSemicolon {
-		p.nextToken()
-	}
-	return cmd, nil
-}
-
 func (p *Parser) parseCommonListProviders() (*Command, error) {
 	p.nextToken() // consume AVAILABLE
 

internal/cli/cli.go

@@ -332,7 +332,7 @@ func looksLikeSQL(s string) bool {
 		"LIST ", "SHOW ", "CREATE ", "DROP ", "ALTER ",
 		"LOGIN ", "REGISTER ", "PING", "GRANT ", "REVOKE ",
 		"SET ", "UNSET ", "UPDATE ", "DELETE ", "INSERT ",
-		"SELECT ", "DESCRIBE ", "EXPLAIN ",
+		"SELECT ", "DESCRIBE ", "EXPLAIN ", "ADD ", "ENABLE ", "DISABLE ", "CHAT ", "USE",
 	}
 	for _, prefix := range sqlPrefixes {
 		if strings.HasPrefix(s, prefix) {
@@ -1008,15 +1008,19 @@ Commands (User Mode):
   LIST TOKENS;                                           - List API tokens
   LIST PROVIDERS;                                        - List available LLM providers
   CREATE TOKEN;                                          - Create new API token
-  CREATE PROVIDER 'name';                                - Create a provider without API key
-  CREATE PROVIDER 'name' 'api_key';                      - Create a provider with API key
+  ADD PROVIDER 'name';                                - Create a provider without API key
+  ADD PROVIDER 'name' 'api_key';                      - Create a provider with API key
   DROP TOKEN 'token_value';                              - Delete an API token
-  DROP PROVIDER 'name';                                  - Delete a provider
+  DELETE PROVIDER 'name';                                  - Delete a provider
   SET TOKEN 'token_value';                               - Set and validate API token
   SHOW TOKEN;                                            - Show current API token
   SHOW PROVIDER 'name';                                  - Show provider details
+  SHOW CURRENT MODEL;                                    - Show current model settings
   UNSET TOKEN;                                           - Remove current API token
   ALTER PROVIDER 'name' NAME 'new_name';                 - Rename a provider
+  USE MODEL 'provider/instance/model';                   - Set current model for chat
+  CHAT 'message';                                        - Chat using current model
+  CHAT 'provider/instance/model' 'message';              - Chat with specified model
 
 Context Engine Commands (no quotes):
   ls [path]                    - List resources

internal/cli/client.go

@@ -24,13 +24,21 @@ import (
 // PasswordPromptFunc is a function type for password input
 type PasswordPromptFunc func(prompt string) (string, error)
 
+// CurrentModel holds the current model configuration
+type CurrentModel struct {
+	Provider string
+	Instance string
+	Model    string
+}
+
 // RAGFlowClient handles API interactions with the RAGFlow server
 type RAGFlowClient struct {
 	HTTPClient     *HTTPClient
 	ServerType     string             // "admin" or "user"
 	PasswordPrompt PasswordPromptFunc // Function for password input
 	OutputFormat   OutputFormat       // Output format: table, plain, json
 	ContextEngine  *ce.Engine         // Context Engine for virtual filesystem
+	CurrentModel   *CurrentModel      // Current model configuration
 }
 
 // NewRAGFlowClient creates a new RAGFlow client
@@ -158,6 +166,8 @@ func (c *RAGFlowClient) ExecuteAdminCommand(cmd *Command) (ResponseIf, error) {
 		return c.ShowProvider(cmd)
 	case "list_provider_models":
 		return c.ListModels(cmd)
+	case "list_instance_models":
+		return c.ListInstanceModels(cmd)
 	case "show_model":
 		return c.ShowModel(cmd)
 	// TODO: Implement other commands
@@ -203,21 +213,44 @@ func (c *RAGFlowClient) ExecuteUserCommand(cmd *Command) (ResponseIf, error) {
 		return c.CreateDocMetaIndex(cmd)
 	case "drop_doc_meta_index":
 		return c.DropDocMetaIndex(cmd)
-	case "list_pool_providers":
+	case "list_available_providers":
 		return c.ListAvailableProviders(cmd)
 	case "show_provider":
 		return c.ShowProvider(cmd)
 	case "list_provider_models":
 		return c.ListModels(cmd)
+	case "list_instance_models":
+		return c.ListInstanceModels(cmd)
 	case "show_model":
 		return c.ShowModel(cmd)
 	// Provider commands
-	case "create_provider":
-		return c.CreateProvider(cmd)
+	case "add_provider":
+		return c.AddProvider(cmd)
 	case "list_providers":
 		return c.ListProviders(cmd)
-	case "drop_provider":
-		return c.DropProvider(cmd)
+	case "delete_provider":
+		return c.DeleteProvider(cmd)
+	// Provider instance commands
+	case "create_provider_instance":
+		return c.CreateProviderInstance(cmd)
+	case "list_provider_instances":
+		return c.ListProviderInstances(cmd)
+	case "show_provider_instance":
+		return c.ShowProviderInstance(cmd)
+	case "alter_provider_instance":
+		return c.AlterProviderInstance(cmd)
+	case "drop_provider_instance":
+		return c.DropProviderInstance(cmd)
+	case "enable_model":
+		return c.EnableOrDisableModel(cmd, "enable")
+	case "disable_model":
+		return c.EnableOrDisableModel(cmd, "disable")
+	case "chat_to_model":
+		return c.ChatToModel(cmd)
+	case "use_model":
+		return c.UseModel(cmd)
+	case "show_current_model":
+		return c.ShowCurrentModel(cmd)
 	// ContextEngine commands
 	case "ce_ls":
 		return c.CEList(cmd)