chore(core): remove useless enum AIActionType (#1725)

EAGzzyCSL · web-flow · commit 0e213a21ba9f · 2026-01-05T21:37:32.000+08:00
diff --git a/apps/chrome-extension/src/extension/recorder/utils.ts b/apps/chrome-extension/src/extension/recorder/utils.ts
@@ -1,5 +1,4 @@
 import {
-  AIActionType,
   type AIArgs,
   callAIWithObjectResponse,
   callAIWithStringResponse,
@@ -321,7 +320,7 @@ export const generateRecordTitle = async (
       const response = await callAIWithObjectResponse<{
         title: string;
         description: string;
-      }>([prompt[0], prompt[1]], AIActionType.EXTRACT_DATA, modelConfig);
+      }>([prompt[0], prompt[1]], modelConfig);
       if (response?.content) {
         return {
           title: response.content.title as string,
@@ -624,11 +623,7 @@ const generateAIMindmap = async (
       },
     ];
 
-    const response = await callAIWithStringResponse(
-      prompt,
-      AIActionType.EXTRACT_DATA,
-      modelConfig,
-    );
+    const response = await callAIWithStringResponse(prompt, modelConfig);
 
     if (response?.content && typeof response.content === 'string') {
       return response.content as string;
diff --git a/packages/core/src/ai-model/index.ts b/packages/core/src/ai-model/index.ts
@@ -31,7 +31,7 @@ export {
   type ConversationHistoryOptions,
 } from './conversation-history';
 
-export { AIActionType, type AIArgs } from '../common';
+export type { AIArgs } from '../common';
 
 export {
   getMidsceneLocationSchema,
diff --git a/packages/core/src/ai-model/inspect.ts b/packages/core/src/ai-model/inspect.ts
@@ -23,12 +23,7 @@ import type {
   ChatCompletionUserMessageParam,
 } from 'openai/resources/index';
 import type { TMultimodalPrompt, TUserPrompt } from '../common';
-import {
-  AIActionType,
-  adaptBboxToRect,
-  expandSearchArea,
-  mergeRects,
-} from '../common';
+import { adaptBboxToRect, expandSearchArea, mergeRects } from '../common';
 import {
   extractDataQueryPrompt,
   systemPromptToExtract,
@@ -200,7 +195,7 @@ export async function AiLocateElement(options: {
     msgs.push(...addOns);
   }
 
-  const res = await callAIFn(msgs, AIActionType.INSPECT_ELEMENT, modelConfig);
+  const res = await callAIFn(msgs, modelConfig);
 
   const rawResponse = JSON.stringify(res.content);
 
@@ -315,7 +310,6 @@ export async function AiLocateSection(options: {
 
   const result = await callAIWithObjectResponse<AISectionLocatorResponse>(
     msgs,
-    AIActionType.EXTRACT_DATA,
     modelConfig,
   );
 
@@ -436,7 +430,6 @@ export async function AiExtractElementInfo<T>(options: {
 
   const result = await callAIWithObjectResponse<AIDataExtractionResponse<T>>(
     msgs,
-    AIActionType.EXTRACT_DATA,
     modelConfig,
   );
   return {
@@ -465,11 +458,7 @@ export async function AiJudgeOrderSensitive(
     },
   ];
 
-  const result = await callAIFn(
-    msgs,
-    AIActionType.INSPECT_ELEMENT, // Reuse existing action type for now
-    modelConfig,
-  );
+  const result = await callAIFn(msgs, modelConfig);
 
   return {
     isOrderSensitive: result.content.isOrderSensitive ?? false,
diff --git a/packages/core/src/ai-model/llm-planning.ts b/packages/core/src/ai-model/llm-planning.ts
@@ -12,7 +12,6 @@ import { getDebug } from '@midscene/shared/logger';
 import { assert } from '@midscene/shared/utils';
 import type { ChatCompletionMessageParam } from 'openai/resources/index';
 import {
-  AIActionType,
   buildYamlFlowFromPlans,
   fillBboxParam,
   findAllMidsceneLocatorField,
@@ -133,7 +132,6 @@ export async function plan(
     reasoning_content,
   } = await callAIWithObjectResponse<RawResponsePlanningAIResponse>(
     msgs,
-    AIActionType.PLAN,
     modelConfig,
     {
       deepThink: opts.deepThink === 'unset' ? undefined : opts.deepThink,
diff --git a/packages/core/src/ai-model/prompt/playwright-generator.ts b/packages/core/src/ai-model/prompt/playwright-generator.ts
@@ -5,7 +5,7 @@ import type {
 import { PLAYWRIGHT_EXAMPLE_CODE } from '@midscene/shared/constants';
 import type { IModelConfig } from '@midscene/shared/env';
 import type { ChatCompletionMessageParam } from 'openai/resources/index';
-import { AIActionType, callAI, callAIWithStringResponse } from '../index';
+import { callAI, callAIWithStringResponse } from '../index';
 // Import shared utilities and types from yaml-generator
 import {
   type ChromeRecordedEvent,
@@ -126,11 +126,7 @@ ${PLAYWRIGHT_EXAMPLE_CODE}`;
     },
   ];
 
-  const response = await callAIWithStringResponse(
-    prompt,
-    AIActionType.TEXT,
-    modelConfig,
-  );
+  const response = await callAIWithStringResponse(prompt, modelConfig);
 
   if (response?.content && typeof response.content === 'string') {
     return response.content;
@@ -212,17 +208,13 @@ ${PLAYWRIGHT_EXAMPLE_CODE}`;
 
   if (options.stream && options.onChunk) {
     // Use streaming
-    return await callAI(prompt, AIActionType.TEXT, modelConfig, {
+    return await callAI(prompt, modelConfig, {
       stream: true,
       onChunk: options.onChunk,
     });
   } else {
     // Fallback to non-streaming
-    const response = await callAIWithStringResponse(
-      prompt,
-      AIActionType.TEXT,
-      modelConfig,
-    );
+    const response = await callAIWithStringResponse(prompt, modelConfig);
 
     if (response?.content && typeof response.content === 'string') {
       return {
diff --git a/packages/core/src/ai-model/prompt/yaml-generator.ts b/packages/core/src/ai-model/prompt/yaml-generator.ts
@@ -5,7 +5,6 @@ import type {
 import { YAML_EXAMPLE_CODE } from '@midscene/shared/constants';
 import type { IModelConfig } from '@midscene/shared/env';
 import {
-  AIActionType,
   type ChatCompletionMessageParam,
   callAI,
   callAIWithStringResponse,
@@ -342,11 +341,7 @@ Important: Return ONLY the raw YAML content. Do NOT wrap the response in markdow
       });
     }
 
-    const response = await callAIWithStringResponse(
-      prompt,
-      AIActionType.TEXT,
-      modelConfig,
-    );
+    const response = await callAIWithStringResponse(prompt, modelConfig);
 
     if (response?.content && typeof response.content === 'string') {
       return response.content;
@@ -433,17 +428,13 @@ Important: Return ONLY the raw YAML content. Do NOT wrap the response in markdow
 
     if (options.stream && options.onChunk) {
       // Use streaming
-      return await callAI(prompt, AIActionType.TEXT, modelConfig, {
+      return await callAI(prompt, modelConfig, {
         stream: true,
         onChunk: options.onChunk,
       });
     } else {
       // Fallback to non-streaming
-      const response = await callAIWithStringResponse(
-        prompt,
-        AIActionType.TEXT,
-        modelConfig,
-      );
+      const response = await callAIWithStringResponse(prompt, modelConfig);
 
       if (response?.content && typeof response.content === 'string') {
         return {
diff --git a/packages/core/src/ai-model/service-caller/index.ts b/packages/core/src/ai-model/service-caller/index.ts
@@ -21,13 +21,11 @@ import { jsonrepair } from 'jsonrepair';
 import OpenAI from 'openai';
 import type { ChatCompletionMessageParam } from 'openai/resources/index';
 import type { Stream } from 'openai/streaming';
-import type { AIActionType, AIArgs } from '../../common';
+import type { AIArgs } from '../../common';
 
 async function createChatClient({
-  AIActionTypeValue,
   modelConfig,
 }: {
-  AIActionTypeValue: AIActionType;
   modelConfig: IModelConfig;
 }): Promise<{
   completion: OpenAI.Chat.Completions;
@@ -204,7 +202,6 @@ async function createChatClient({
 
 export async function callAI(
   messages: ChatCompletionMessageParam[],
-  AIActionTypeValue: AIActionType,
   modelConfig: IModelConfig,
   options?: {
     stream?: boolean;
@@ -219,7 +216,6 @@ export async function callAI(
 }> {
   const { completion, modelName, modelDescription, uiTarsVersion, vlMode } =
     await createChatClient({
-      AIActionTypeValue,
       modelConfig,
     });
 
@@ -425,7 +421,6 @@ export async function callAI(
 
 export async function callAIWithObjectResponse<T>(
   messages: ChatCompletionMessageParam[],
-  AIActionTypeValue: AIActionType,
   modelConfig: IModelConfig,
   options?: {
     deepThink?: DeepThinkOption;
@@ -436,7 +431,7 @@ export async function callAIWithObjectResponse<T>(
   usage?: AIUsageInfo;
   reasoning_content?: string;
 }> {
-  const response = await callAI(messages, AIActionTypeValue, modelConfig, {
+  const response = await callAI(messages, modelConfig, {
     deepThink: options?.deepThink,
   });
   assert(response, 'empty response');
@@ -456,10 +451,9 @@ export async function callAIWithObjectResponse<T>(
 
 export async function callAIWithStringResponse(
   msgs: AIArgs,
-  AIActionTypeValue: AIActionType,
   modelConfig: IModelConfig,
 ): Promise<{ content: string; usage?: AIUsageInfo }> {
-  const { content, usage } = await callAI(msgs, AIActionTypeValue, modelConfig);
+  const { content, usage } = await callAI(msgs, modelConfig);
   return { content, usage };
 }
 
diff --git a/packages/core/src/ai-model/ui-tars-planning.ts b/packages/core/src/ai-model/ui-tars-planning.ts
@@ -10,7 +10,6 @@ import { getDebug } from '@midscene/shared/logger';
 import { transformHotkeyInput } from '@midscene/shared/us-keyboard-layout';
 import { assert } from '@midscene/shared/utils';
 import { actionParser } from '@ui-tars/action-parser';
-import { AIActionType } from '../common';
 import type { ConversationHistory } from './conversation-history';
 import { getSummary, getUiTarsPlanningPrompt } from './prompt/ui-tars-planning';
 import { callAIWithStringResponse } from './service-caller/index';
@@ -85,7 +84,6 @@ export async function uiTarsPlanning(
       },
       ...conversationHistory.snapshot(),
     ],
-    AIActionType.INSPECT_ELEMENT,
     modelConfig,
   );
   const convertedText = convertBboxToCoordinates(res.content);
diff --git a/packages/core/src/common.ts b/packages/core/src/common.ts
@@ -21,15 +21,6 @@ import { z } from 'zod';
 
 export type AIArgs = ChatCompletionMessageParam[];
 
-export enum AIActionType {
-  ASSERT = 0,
-  INSPECT_ELEMENT = 1,
-  EXTRACT_DATA = 2,
-  PLAN = 3,
-  DESCRIBE_ELEMENT = 4,
-  TEXT = 5,
-}
-
 const defaultBboxSize = 20; // must be even number
 const debugInspectUtils = getDebug('ai:common');
 type AdaptBboxInput = number[] | string[] | string | (number[] | string[])[];
diff --git a/packages/core/src/service/index.ts b/packages/core/src/service/index.ts
@@ -5,7 +5,7 @@ import {
 } from '@/ai-model/index';
 import { AiLocateSection } from '@/ai-model/inspect';
 import { elementDescriberInstruction } from '@/ai-model/prompt/describe';
-import { AIActionType, type AIArgs, expandSearchArea } from '@/common';
+import { type AIArgs, expandSearchArea } from '@/common';
 import type {
   AIDescribeElementResponse,
   AIUsageInfo,
@@ -343,11 +343,7 @@ export default class Service {
     const callAIFn = this
       .aiVendorFn as typeof callAIWithObjectResponse<AIDescribeElementResponse>;
 
-    const res = await callAIFn(
-      msgs,
-      AIActionType.DESCRIBE_ELEMENT,
-      modelConfig,
-    );
+    const res = await callAIFn(msgs, modelConfig);
 
     const { content } = res;
     assert(!content.error, `describe failed: ${content.error}`);
diff --git a/packages/core/tests/ai/connectivity.test.ts b/packages/core/tests/ai/connectivity.test.ts
@@ -1,6 +1,5 @@
 import { existsSync } from 'node:fs';
 import path from 'node:path';
-import { AIActionType } from '@/ai-model';
 import { callAI, callAIWithObjectResponse } from '@/ai-model/service-caller';
 import { localImg2Base64 } from '@/image';
 import { globalModelConfigManager } from '@midscene/shared/env';
@@ -58,7 +57,6 @@ vi.setConfig({
               '鲁迅认识周树人吗？回答我：1. 分析原因 2.回答：是/否/无效问题',
           },
         ],
-        AIActionType.EXTRACT_DATA,
         defaultModelConfig,
       );
 
@@ -77,7 +75,6 @@ vi.setConfig({
             content: '3 x 5 = ?',
           },
         ],
-        AIActionType.INSPECT_ELEMENT,
         defaultModelConfig,
       );
       expect(result.content).toEqual({ answer: 15 });
@@ -106,7 +103,6 @@ vi.setConfig({
             content: '3 x 5 = ?',
           },
         ],
-        AIActionType.INSPECT_ELEMENT,
         defaultModelConfig,
       );
       expect(result.content).toEqual({ answer: 15 });
@@ -133,7 +129,6 @@ vi.setConfig({
             ],
           },
         ],
-        AIActionType.EXTRACT_DATA,
         defaultModelConfig,
       );
 
diff --git a/packages/core/tests/ai/streaming.test.ts b/packages/core/tests/ai/streaming.test.ts
diff --git a/packages/core/tests/unit-test/prompt/playwright-generator.test.ts b/packages/core/tests/unit-test/prompt/playwright-generator.test.ts
diff --git a/packages/core/tests/unit-test/proxy-configuration.test.ts b/packages/core/tests/unit-test/proxy-configuration.test.ts
