Workflow deep and workflow connection performance (#2664)

* feat: Workflow dispatch deep

* perf: workflow connection

docSite/content/zh-cn/docs/development/upgrading/4811.md

@@ -22,3 +22,6 @@ weight: 813
 5. 新增 - 插件支持配置使用引导、全局变量和文件输入。
 6. 新增 - 简易模式支持新的版本管理方式。
 7. 新增 - 聊天记录滚动加载，不再只加载 30 条。
+8. 优化 - 工作流嵌套层级限制 20 层，避免因编排不合理导致的无限死循环。
+9. 优化 - 工作流 handler 性能优化。
+10. 修复 - 知识库选择权限问题。

packages/global/core/workflow/runtime/constants.ts

@@ -24,7 +24,8 @@ export enum DispatchNodeResponseKeyEnum {
   assistantResponses = 'assistantResponses', // assistant response
   rewriteHistories = 'rewriteHistories', // If have the response, workflow histories will be rewrite
 
-  interactive = 'INTERACTIVE' // is interactive
+  interactive = 'INTERACTIVE', // is interactive
+  runTimes = 'runTimes' // run times
 }
 
 export const needReplaceReferenceInputTypeList = [

packages/global/core/workflow/runtime/type.d.ts

@@ -45,6 +45,7 @@ export type ChatDispatchProps = {
   maxRunTimes: number;
   isToolCall?: boolean;
   workflowStreamResponse?: WorkflowResponseType;
+  workflowDispatchDeep?: number;
 };
 
 export type ModuleDispatchProps<T> = ChatDispatchProps & {
@@ -181,6 +182,7 @@ export type DispatchNodeResultType<T = {}> = {
   [DispatchNodeResponseKeyEnum.toolResponses]?: ToolRunResponseItemType; // Tool response
   [DispatchNodeResponseKeyEnum.assistantResponses]?: AIChatItemValueItemType[]; // Assistant response(Store to db)
   [DispatchNodeResponseKeyEnum.rewriteHistories]?: ChatItemType[];
+  [DispatchNodeResponseKeyEnum.runTimes]?: number;
 } & T;
 
 /* Single node props */

packages/service/core/workflow/constants.ts

+export const WORKFLOW_MAX_RUN_TIMES = process.env.WORKFLOW_MAX_RUN_TIMES
+  ? parseInt(process.env.WORKFLOW_MAX_RUN_TIMES)
+  : 500;

packages/service/core/workflow/dispatch/agent/runTool/functionCall.ts

@@ -298,7 +298,10 @@ export const runToolWithFunctionCall = async (
         dispatchFlowResponse,
         totalTokens: response?.totalTokens ? response.totalTokens + tokens : tokens,
         completeMessages: filterMessages,
-        assistantResponses: toolNodeAssistants
+        assistantResponses: toolNodeAssistants,
+        runTimes:
+          (response?.runTimes || 0) +
+          flatToolsResponseData.reduce((sum, item) => sum + item.runTimes, 0)
       };
     }
 
@@ -310,7 +313,10 @@ export const runToolWithFunctionCall = async (
       {
         dispatchFlowResponse,
         totalTokens: response?.totalTokens ? response.totalTokens + tokens : tokens,
-        assistantResponses: toolNodeAssistants
+        assistantResponses: toolNodeAssistants,
+        runTimes:
+          (response?.runTimes || 0) +
+          flatToolsResponseData.reduce((sum, item) => sum + item.runTimes, 0)
       }
     );
   } else {
@@ -330,7 +336,8 @@ export const runToolWithFunctionCall = async (
       dispatchFlowResponse: response?.dispatchFlowResponse || [],
       totalTokens: response?.totalTokens ? response.totalTokens + tokens : tokens,
       completeMessages,
-      assistantResponses: [...assistantResponses, ...toolNodeAssistant.value]
+      assistantResponses: [...assistantResponses, ...toolNodeAssistant.value],
+      runTimes: (response?.runTimes || 0) + 1
     };
   }
 };

packages/service/core/workflow/dispatch/agent/runTool/index.ts

@@ -125,7 +125,8 @@ export const dispatchRunTools = async (props: DispatchToolModuleProps): Promise<
     dispatchFlowResponse, // tool flow response
     totalTokens,
     completeMessages = [], // The actual message sent to AI(just save text)
-    assistantResponses = [] // FastGPT system store assistant.value response
+    assistantResponses = [], // FastGPT system store assistant.value response
+    runTimes
   } = await (async () => {
     const adaptMessages = chats2GPTMessages({ messages, reserveId: false });
 
@@ -195,6 +196,7 @@ export const dispatchRunTools = async (props: DispatchToolModuleProps): Promise<
   const previewAssistantResponses = filterToolResponseToPreview(assistantResponses);
 
   return {
+    [DispatchNodeResponseKeyEnum.runTimes]: runTimes,
     [NodeOutputKeyEnum.answerText]: previewAssistantResponses
       .filter((item) => item.text?.content)
       .map((item) => item.text?.content || '')

packages/service/core/workflow/dispatch/agent/runTool/promptCall.ts

@@ -180,7 +180,8 @@ export const runToolWithPromptCall = async (
       dispatchFlowResponse: response?.dispatchFlowResponse || [],
       totalTokens: response?.totalTokens ? response.totalTokens + tokens : tokens,
       completeMessages,
-      assistantResponses: [...assistantResponses, ...toolNodeAssistant.value]
+      assistantResponses: [...assistantResponses, ...toolNodeAssistant.value],
+      runTimes: (response?.runTimes || 0) + 1
     };
   }
 
@@ -318,7 +319,8 @@ ANSWER: `;
       dispatchFlowResponse,
       totalTokens: response?.totalTokens ? response.totalTokens + tokens : tokens,
       completeMessages: filterMessages,
-      assistantResponses: toolNodeAssistants
+      assistantResponses: toolNodeAssistants,
+      runTimes: (response?.runTimes || 0) + toolsRunResponse.moduleRunResponse.runTimes
     };
   }
 
@@ -330,7 +332,8 @@ ANSWER: `;
     {
       dispatchFlowResponse,
       totalTokens: response?.totalTokens ? response.totalTokens + tokens : tokens,
-      assistantResponses: toolNodeAssistants
+      assistantResponses: toolNodeAssistants,
+      runTimes: (response?.runTimes || 0) + toolsRunResponse.moduleRunResponse.runTimes
     }
   );
 };

packages/service/core/workflow/dispatch/agent/runTool/toolChoice.ts

@@ -325,7 +325,10 @@ export const runToolWithToolChoice = async (
           dispatchFlowResponse,
           totalTokens: response?.totalTokens ? response.totalTokens + tokens : tokens,
           completeMessages,
-          assistantResponses: toolNodeAssistants
+          assistantResponses: toolNodeAssistants,
+          runTimes:
+            (response?.runTimes || 0) +
+            flatToolsResponseData.reduce((sum, item) => sum + item.runTimes, 0)
         };
       }
 
@@ -338,7 +341,10 @@ export const runToolWithToolChoice = async (
         {
           dispatchFlowResponse,
           totalTokens: response?.totalTokens ? response.totalTokens + tokens : tokens,
-          assistantResponses: toolNodeAssistants
+          assistantResponses: toolNodeAssistants,
+          runTimes:
+            (response?.runTimes || 0) +
+            flatToolsResponseData.reduce((sum, item) => sum + item.runTimes, 0)
         }
       );
     } else {
@@ -358,7 +364,8 @@ export const runToolWithToolChoice = async (
         dispatchFlowResponse: response?.dispatchFlowResponse || [],
         totalTokens: response?.totalTokens ? response.totalTokens + tokens : tokens,
         completeMessages,
-        assistantResponses: [...assistantResponses, ...toolNodeAssistant.value]
+        assistantResponses: [...assistantResponses, ...toolNodeAssistant.value],
+        runTimes: (response?.runTimes || 0) + 1
       };
     }
   } catch (error) {

packages/service/core/workflow/dispatch/agent/runTool/type.d.ts

@@ -8,6 +8,7 @@ import type { RuntimeNodeItemType } from '@fastgpt/global/core/workflow/runtime/
 import { ChatNodeUsageType } from '@fastgpt/global/support/wallet/bill/type';
 import type { DispatchFlowResponse } from '../../type.d';
 import { AIChatItemValueItemType, ChatItemValueItemType } from '@fastgpt/global/core/chat/type';
+import { DispatchNodeResponseKeyEnum } from '@fastgpt/global/core/workflow/runtime/constants';
 
 export type DispatchToolModuleProps = ModuleDispatchProps<{
   [NodeInputKeyEnum.history]?: ChatItemType[];
@@ -25,6 +26,7 @@ export type RunToolResponse = {
   totalTokens: number;
   completeMessages?: ChatCompletionMessageParam[];
   assistantResponses?: AIChatItemValueItemType[];
+  [DispatchNodeResponseKeyEnum.runTimes]: number;
 };
 export type ToolNodeItemType = RuntimeNodeItemType & {
   toolParams: RuntimeNodeItemType['inputs'];

packages/service/core/workflow/dispatch/index.ts

@@ -119,6 +119,31 @@ export async function dispatchWorkFlow(data: Props): Promise<DispatchFlowRespons
     ...props
   } = data;
 
+  // 初始化深度和自动增加深度，避免无限嵌套
+  if (!props.workflowDispatchDeep) {
+    props.workflowDispatchDeep = 1;
+  } else {
+    props.workflowDispatchDeep += 1;
+  }
+
+  if (props.workflowDispatchDeep > 20) {
+    return {
+      flowResponses: [],
+      flowUsages: [],
+      debugResponse: {
+        finishedNodes: [],
+        finishedEdges: [],
+        nextStepRunNodes: []
+      },
+      [DispatchNodeResponseKeyEnum.runTimes]: 1,
+      [DispatchNodeResponseKeyEnum.assistantResponses]: [],
+      [DispatchNodeResponseKeyEnum.toolResponses]: null,
+      newVariables: removeSystemVariable(variables)
+    };
+  }
+
+  let workflowRunTimes = 0;
+
   // set sse response headers
   if (stream && res) {
     res.setHeader('Content-Type', 'text/event-stream;charset=utf-8');
@@ -154,7 +179,8 @@ export async function dispatchWorkFlow(data: Props): Promise<DispatchFlowRespons
       nodeDispatchUsages,
       toolResponses,
       assistantResponses,
-      rewriteHistories
+      rewriteHistories,
+      runTimes = 1
     }: Omit<
       DispatchNodeResultType<{
         [NodeOutputKeyEnum.answerText]?: string;
@@ -163,6 +189,10 @@ export async function dispatchWorkFlow(data: Props): Promise<DispatchFlowRespons
       'nodeResponse'
     >
   ) {
+    // Add run times
+    workflowRunTimes += runTimes;
+    props.maxRunTimes -= runTimes;
+
     if (responseData) {
       chatResponses.push(responseData);
     }
@@ -330,7 +360,6 @@ export async function dispatchWorkFlow(data: Props): Promise<DispatchFlowRespons
     const nodeRunResult = await (() => {
       if (status === 'run') {
         nodeRunBeforeHook(node);
-        props.maxRunTimes--;
         addLog.debug(`[dispatchWorkFlow] nodeRunWithActive: ${node.name}`);
         return nodeRunWithActive(node);
       }
@@ -565,6 +594,7 @@ export async function dispatchWorkFlow(data: Props): Promise<DispatchFlowRespons
       finishedEdges: runtimeEdges,
       nextStepRunNodes: debugNextStepRunNodes
     },
+    [DispatchNodeResponseKeyEnum.runTimes]: workflowRunTimes,
     [DispatchNodeResponseKeyEnum.assistantResponses]:
       mergeAssistantResponseAnswerText(chatAssistantResponse),
     [DispatchNodeResponseKeyEnum.toolResponses]: toolRunResponse,

packages/service/core/workflow/dispatch/plugin/run.ts

@@ -66,7 +66,7 @@ export const dispatchRunPlugin = async (props: RunPluginProps): Promise<RunPlugi
     appId: String(plugin.id)
   };
 
-  const { flowResponses, flowUsages, assistantResponses } = await dispatchWorkFlow({
+  const { flowResponses, flowUsages, assistantResponses, runTimes } = await dispatchWorkFlow({
     ...props,
     runningAppInfo: {
       id: String(plugin.id),
@@ -92,6 +92,7 @@ export const dispatchRunPlugin = async (props: RunPluginProps): Promise<RunPlugi
   return {
     assistantResponses,
     // responseData, // debug
+    [DispatchNodeResponseKeyEnum.runTimes]: runTimes,
     [DispatchNodeResponseKeyEnum.nodeResponse]: {
       moduleLogo: plugin.avatar,
       totalPoints: usagePoints,

packages/service/core/workflow/dispatch/plugin/runApp.ts

@@ -75,7 +75,7 @@ export const dispatchRunAppNode = async (props: Props): Promise<Response> => {
     appId: String(appData._id)
   };
 
-  const { flowResponses, flowUsages, assistantResponses } = await dispatchWorkFlow({
+  const { flowResponses, flowUsages, assistantResponses, runTimes } = await dispatchWorkFlow({
     ...props,
     runningAppInfo: {
       id: String(appData._id),
@@ -107,6 +107,7 @@ export const dispatchRunAppNode = async (props: Props): Promise<Response> => {
   const { text } = chatValue2RuntimePrompt(assistantResponses);
 
   return {
+    [DispatchNodeResponseKeyEnum.runTimes]: runTimes,
     [DispatchNodeResponseKeyEnum.nodeResponse]: {
       moduleLogo: appData.avatar,
       query: userChatInput,

packages/service/core/workflow/dispatch/type.d.ts

@@ -22,6 +22,7 @@ export type DispatchFlowResponse = {
   };
   [DispatchNodeResponseKeyEnum.toolResponses]: ToolRunResponseItemType;
   [DispatchNodeResponseKeyEnum.assistantResponses]: AIChatItemValueItemType[];
+  [DispatchNodeResponseKeyEnum.runTimes]: number;
   newVariables: Record<string, string>;
 };
 

projects/app/.env.template

@@ -36,3 +36,6 @@ HOME_URL=/
 # 日志等级: debug, info, warn, error
 LOG_LEVEL=debug
 STORE_LOG_LEVEL=warn
+
+# 工作流最大运行次数，避免极端的死循环情况
+WORKFLOW_MAX_RUN_TIMES=500
\ No newline at end of file

projects/app/src/pages/api/core/chat/chatTest.ts

@@ -27,6 +27,7 @@ import {
 import { StoreNodeItemType } from '@fastgpt/global/core/workflow/type/node';
 import { getWorkflowResponseWrite } from '@fastgpt/service/core/workflow/dispatch/utils';
 import { getNanoid } from '@fastgpt/global/common/string/tools';
+import { WORKFLOW_MAX_RUN_TIMES } from '@fastgpt/service/core/workflow/constants';
 
 export type Props = {
   messages: ChatCompletionMessageParam[];
@@ -120,7 +121,7 @@ async function handler(req: NextApiRequest, res: NextApiResponse) {
       chatConfig,
       histories: chatMessages,
       stream: true,
-      maxRunTimes: 200,
+      maxRunTimes: WORKFLOW_MAX_RUN_TIMES,
       workflowStreamResponse: workflowResponseWrite
     });
 

projects/app/src/pages/api/core/workflow/debug.ts

@@ -9,6 +9,7 @@ import { PostWorkflowDebugProps, PostWorkflowDebugResponse } from '@/global/core
 import { NextAPI } from '@/service/middleware/entry';
 import { ReadPermissionVal } from '@fastgpt/global/support/permission/constant';
 import { defaultApp } from '@/web/core/app/constants';
+import { WORKFLOW_MAX_RUN_TIMES } from '@fastgpt/service/core/workflow/constants';
 
 async function handler(
   req: NextApiRequest,
@@ -57,7 +58,7 @@ async function handler(
     chatConfig: defaultApp.chatConfig,
     histories: [],
     stream: false,
-    maxRunTimes: 200
+    maxRunTimes: WORKFLOW_MAX_RUN_TIMES
   });
 
   pushChatUsage({

projects/app/src/pages/api/v1/chat/completions.ts

@@ -59,6 +59,7 @@ import { getSystemTime } from '@fastgpt/global/common/time/timezone';
 import { rewriteNodeOutputByHistories } from '@fastgpt/global/core/workflow/runtime/utils';
 import { getWorkflowResponseWrite } from '@fastgpt/service/core/workflow/dispatch/utils';
 import { getPluginRunUserQuery } from '@fastgpt/service/core/workflow/utils';
+import { WORKFLOW_MAX_RUN_TIMES } from '@fastgpt/service/core/workflow/constants';
 
 type FastGptWebChatProps = {
   chatId?: string; // undefined: get histories from messages, '': new chat, 'xxxxx': get histories from db
@@ -264,7 +265,7 @@ async function handler(req: NextApiRequest, res: NextApiResponse) {
           chatConfig,
           histories: newHistories,
           stream,
-          maxRunTimes: 200,
+          maxRunTimes: WORKFLOW_MAX_RUN_TIMES,
           workflowStreamResponse: workflowResponseWrite
         });
       }

projects/app/src/pages/app/detail/components/WorkflowComponents/Flow/nodes/render/NodeCard.tsx

@@ -248,6 +248,14 @@ const NodeCard = (props: Props) => {
     onChangeNode,
     toast
   ]);
+  const RenderHandle = useMemo(() => {
+    return (
+      <>
+        <ConnectionSourceHandle nodeId={nodeId} />
+        <ConnectionTargetHandle nodeId={nodeId} />
+      </>
+    );
+  }, [nodeId]);
 
   return (
     <Box
@@ -283,8 +291,7 @@ const NodeCard = (props: Props) => {
       <NodeDebugResponse nodeId={nodeId} debugResult={debugResult} />
       {Header}
       {children}
-      <ConnectionSourceHandle nodeId={nodeId} />
-      <ConnectionTargetHandle nodeId={nodeId} />
+      {RenderHandle}
 
       <EditTitleModal maxLength={20} />
     </Box>

projects/app/src/service/core/app/utils.ts

@@ -13,6 +13,7 @@ import {
 import { UsageSourceEnum } from '@fastgpt/global/support/wallet/usage/constants';
 import { addLog } from '@fastgpt/service/common/system/log';
 import { MongoApp } from '@fastgpt/service/core/app/schema';
+import { WORKFLOW_MAX_RUN_TIMES } from '@fastgpt/service/core/workflow/constants';
 import { dispatchWorkFlow } from '@fastgpt/service/core/workflow/dispatch';
 
 export const getScheduleTriggerApp = async () => {
@@ -55,7 +56,7 @@ export const getScheduleTriggerApp = async () => {
           chatConfig: defaultApp.chatConfig,
           histories: [],
           stream: false,
-          maxRunTimes: 200
+          maxRunTimes: WORKFLOW_MAX_RUN_TIMES
         });
         pushChatUsage({
           appName: app.name,