Merge ac7b720b5b into 48469bd8ca

Update README.md

.gitignore

@@ -49,3 +49,4 @@ masks.json
 
 # mcp config
 app/mcp/mcp_config.json
+Dockerfile.local

README.md

@@ -7,7 +7,7 @@
 
 
 
-<h1 align="center">NextChat (ChatGPT Next Web)</h1>
+<h1 align="center">NextChat</h1>
 
 English / [简体中文](./README_CN.md)
 
@@ -22,7 +22,6 @@ English / [简体中文](./README_CN.md)
 [![MacOS][MacOS-image]][download-url]
 [![Linux][Linux-image]][download-url]
 
-[NextChatAI](https://nextchat.dev/chat?utm_source=readme) / [Web App Demo](https://app.nextchat.dev) / [Desktop App](https://github.com/Yidadaa/ChatGPT-Next-Web/releases) 
 [NextChatAI](https://nextchat.club?utm_source=readme) / [Web App Demo](https://app.nextchat.dev) / [Desktop App](https://github.com/Yidadaa/ChatGPT-Next-Web/releases) / [Discord](https://discord.gg/YCkeafCafC) / [Enterprise Edition](#enterprise-edition) / [Twitter](https://twitter.com/NextChatDev)
 
 
@@ -41,24 +40,6 @@ English / [简体中文](./README_CN.md)
 
 </div>
 
-## 👋 Hey, NextChat is going to develop a native app!
-
-> This week we are going to start working on iOS and Android APP, and we want to find some reliable friends to do it together!
-
-
-✨ Several key points:
-
-- Starting from 0, you are a veteran
-- Completely open source, not hidden
-- Native development, pursuing the ultimate experience
-
-Will you come and do something together? 😎
-
-https://github.com/ChatGPTNextWeb/NextChat/issues/6269
-
-#Seeking for talents is thirsty #lack of people"
-
-
 ## 🥳 Cheer for DeepSeek, China's AI star!
  > Purpose-Built UI for DeepSeek Reasoner Model
  

app/client/api.ts

@@ -40,6 +40,11 @@ export interface MultimodalContent {
   };
 }
 
+export interface MultimodalContentForAlibaba {
+  text?: string;
+  image?: string;
+}
+
 export interface RequestMessage {
   role: MessageRole;
   content: string | MultimodalContent[];

app/client/platforms/alibaba.ts

@@ -7,7 +7,10 @@ import {
   ChatMessageTool,
   usePluginStore,
 } from "@/app/store";
-import { streamWithThink } from "@/app/utils/chat";
+import {
+  preProcessImageContentForAlibabaDashScope,
+  streamWithThink,
+} from "@/app/utils/chat";
 import {
   ChatOptions,
   getHeaders,
@@ -15,12 +18,14 @@ import {
   LLMModel,
   SpeechOptions,
   MultimodalContent,
+  MultimodalContentForAlibaba,
 } from "../api";
 import { getClientConfig } from "@/app/config/client";
 import {
   getMessageTextContent,
   getMessageTextContentWithoutThinking,
   getTimeoutMSByModel,
+  isVisionModel,
 } from "@/app/utils";
 import { fetch } from "@/app/utils/stream";
 
@@ -89,14 +94,6 @@ export class QwenApi implements LLMApi {
   }
 
   async chat(options: ChatOptions) {
-    const messages = options.messages.map((v) => ({
-      role: v.role,
-      content:
-        v.role === "assistant"
-          ? getMessageTextContentWithoutThinking(v)
-          : getMessageTextContent(v),
-    }));
-
     const modelConfig = {
       ...useAppConfig.getState().modelConfig,
       ...useChatStore.getState().currentSession().mask.modelConfig,
@@ -105,6 +102,21 @@ export class QwenApi implements LLMApi {
       },
     };
 
+    const visionModel = isVisionModel(options.config.model);
+
+    const messages: ChatOptions["messages"] = [];
+    for (const v of options.messages) {
+      const content = (
+        visionModel
+          ? await preProcessImageContentForAlibabaDashScope(v.content)
+          : v.role === "assistant"
+          ? getMessageTextContentWithoutThinking(v)
+          : getMessageTextContent(v)
+      ) as any;
+
+      messages.push({ role: v.role, content });
+    }
+
     const shouldStream = !!options.config.stream;
     const requestPayload: RequestPayload = {
       model: modelConfig.model,
@@ -129,7 +141,7 @@ export class QwenApi implements LLMApi {
         "X-DashScope-SSE": shouldStream ? "enable" : "disable",
       };
 
-      const chatPath = this.path(Alibaba.ChatPath);
+      const chatPath = this.path(Alibaba.ChatPath(modelConfig.model));
       const chatPayload = {
         method: "POST",
         body: JSON.stringify(requestPayload),
@@ -162,7 +174,7 @@ export class QwenApi implements LLMApi {
             const json = JSON.parse(text);
             const choices = json.output.choices as Array<{
               message: {
-                content: string | null;
+                content: string | null | MultimodalContentForAlibaba[];
                 tool_calls: ChatMessageTool[];
                 reasoning_content: string | null;
               };
@@ -212,7 +224,9 @@ export class QwenApi implements LLMApi {
             } else if (content && content.length > 0) {
               return {
                 isThinking: false,
-                content: content,
+                content: Array.isArray(content)
+                  ? content.map((item) => item.text).join(",")
+                  : content,
               };
             }
 

app/client/platforms/anthropic.ts

@@ -71,8 +71,6 @@ const ClaudeMapper = {
   system: "user",
 } as const;
 
-const keys = ["claude-2, claude-instant-1"];
-
 export class ClaudeApi implements LLMApi {
   speech(options: SpeechOptions): Promise<ArrayBuffer> {
     throw new Error("Method not implemented.");

app/client/platforms/deepseek.ts

@@ -75,6 +75,25 @@ export class DeepSeekApi implements LLMApi {
       }
     }
 
+    // 检测并修复消息顺序，确保除system外的第一个消息是user
+    const filteredMessages: ChatOptions["messages"] = [];
+    let hasFoundFirstUser = false;
+
+    for (const msg of messages) {
+      if (msg.role === "system") {
+        // Keep all system messages
+        filteredMessages.push(msg);
+      } else if (msg.role === "user") {
+        // User message directly added
+        filteredMessages.push(msg);
+        hasFoundFirstUser = true;
+      } else if (hasFoundFirstUser) {
+        // After finding the first user message, all subsequent non-system messages are retained.
+        filteredMessages.push(msg);
+      }
+      // If hasFoundFirstUser is false and it is not a system message, it will be skipped.
+    }
+
     const modelConfig = {
       ...useAppConfig.getState().modelConfig,
       ...useChatStore.getState().currentSession().mask.modelConfig,
@@ -85,7 +104,7 @@ export class DeepSeekApi implements LLMApi {
     };
 
     const requestPayload: RequestPayload = {
-      messages,
+      messages: filteredMessages,
       stream: options.config.stream,
       model: modelConfig.model,
       temperature: modelConfig.temperature,

app/client/platforms/google.ts

@@ -197,8 +197,6 @@ export class GeminiProApi implements LLMApi {
         signal: controller.signal,
         headers: getHeaders(),
       };
-
-      const isThinking = options.config.model.includes("-thinking");
       // make a fetch request
       const requestTimeoutId = setTimeout(
         () => controller.abort(),

app/client/platforms/openai.ts

@@ -67,6 +67,8 @@ export interface RequestPayload {
   top_p: number;
   max_tokens?: number;
   max_completion_tokens?: number;
+  reasoning_effort?: string;
+  // O3 only
 }
 
 export interface DalleRequestPayload {
@@ -196,9 +198,9 @@ export class ChatGPTApi implements LLMApi {
     let requestPayload: RequestPayload | DalleRequestPayload;
 
     const isDalle3 = _isDalle3(options.config.model);
-    const isO1OrO3 =
-      options.config.model.startsWith("o1") ||
-      options.config.model.startsWith("o3");
+    const isO1 = options.config.model.startsWith("o1");
+    const isO3 = options.config.model.startsWith("o3");
+    const isO1OrO3 = isO1 || isO3;
     if (isDalle3) {
       const prompt = getMessageTextContent(
         options.messages.slice(-1)?.pop() as any,
@@ -242,9 +244,18 @@ export class ChatGPTApi implements LLMApi {
         requestPayload["max_completion_tokens"] = modelConfig.max_tokens;
       }
 
+      if (isO3) {
+        requestPayload["reasoning_effort"] = "high";
+        // make o3-mini defaults to high reasoning effort
+      }
+
       // add max_tokens to vision model
       if (visionModel) {
-        requestPayload["max_tokens"] = Math.max(modelConfig.max_tokens, 4000);
+        if (isO1) {
+          requestPayload["max_completion_tokens"] = modelConfig.max_tokens;
+        } else {
+          requestPayload["max_tokens"] = Math.max(modelConfig.max_tokens, 4000);
+        }
       }
     }
 
@@ -286,6 +297,11 @@ export class ChatGPTApi implements LLMApi {
           isDalle3 ? OpenaiPath.ImagePath : OpenaiPath.ChatPath,
         );
       }
+      // make a fetch request
+      const requestTimeoutId = setTimeout(
+        () => controller.abort(),
+        getTimeoutMSByModel(options.config.model),
+      );
       if (shouldStream) {
         let index = -1;
         const [tools, funcs] = usePluginStore
@@ -393,12 +409,6 @@ export class ChatGPTApi implements LLMApi {
           headers: getHeaders(),
         };
 
-        // make a fetch request
-        const requestTimeoutId = setTimeout(
-          () => controller.abort(),
-          getTimeoutMSByModel(options.config.model),
-        );
-
         const res = await fetch(chatPath, chatPayload);
         clearTimeout(requestTimeoutId);
 

app/components/emoji.tsx

@@ -66,11 +66,11 @@ export function Avatar(props: { model?: ModelType; avatar?: string }) {
       LlmIcon = BotIconGemma;
     } else if (modelName.startsWith("claude")) {
       LlmIcon = BotIconClaude;
-    } else if (modelName.toLowerCase().includes("llama")) {
+    } else if (modelName.includes("llama")) {
       LlmIcon = BotIconMeta;
-    } else if (modelName.startsWith("mixtral")) {
+    } else if (modelName.startsWith("mixtral") || modelName.startsWith("codestral")) {
       LlmIcon = BotIconMistral;
-    } else if (modelName.toLowerCase().includes("deepseek")) {
+    } else if (modelName.includes("deepseek")) {
       LlmIcon = BotIconDeepseek;
     } else if (modelName.startsWith("moonshot")) {
       LlmIcon = BotIconMoonshot;
@@ -85,7 +85,7 @@ export function Avatar(props: { model?: ModelType; avatar?: string }) {
     } else if (modelName.startsWith("doubao") || modelName.startsWith("ep-")) {
       LlmIcon = BotIconDoubao;
     } else if (
-      modelName.toLowerCase().includes("glm") ||
+      modelName.includes("glm") ||
       modelName.startsWith("cogview-") ||
       modelName.startsWith("cogvideox-")
     ) {

app/constant.ts

@@ -221,7 +221,12 @@ export const ByteDance = {
 
 export const Alibaba = {
   ExampleEndpoint: ALIBABA_BASE_URL,
-  ChatPath: "v1/services/aigc/text-generation/generation",
+  ChatPath: (modelName: string) => {
+    if (modelName.includes("vl") || modelName.includes("omni")) {
+      return "v1/services/aigc/multimodal-generation/generation";
+    }
+    return `v1/services/aigc/text-generation/generation`;
+  },
 };
 
 export const Tencent = {
@@ -403,35 +408,20 @@ You are an AI assistant with access to system tools. Your role is to help users
    
 `;
 
-export const SUMMARIZE_MODEL = "gpt-4o-mini";
-export const GEMINI_SUMMARIZE_MODEL = "gemini-pro";
+export const SUMMARIZE_MODEL = "gpt-4.1-mini";
+export const GEMINI_SUMMARIZE_MODEL = "gemini-2.0-flash";
 export const DEEPSEEK_SUMMARIZE_MODEL = "deepseek-chat";
 
 export const KnowledgeCutOffDate: Record<string, string> = {
-  default: "2021-09",
-  "gpt-4-turbo": "2023-12",
-  "gpt-4-turbo-2024-04-09": "2023-12",
-  "gpt-4-turbo-preview": "2023-12",
-  "gpt-4o": "2023-10",
-  "gpt-4o-2024-05-13": "2023-10",
-  "gpt-4o-2024-08-06": "2023-10",
-  "gpt-4o-2024-11-20": "2023-10",
-  "chatgpt-4o-latest": "2023-10",
-  "gpt-4o-mini": "2023-10",
-  "gpt-4o-mini-2024-07-18": "2023-10",
-  "gpt-4-vision-preview": "2023-04",
-  "o1-mini-2024-09-12": "2023-10",
-  "o1-mini": "2023-10",
-  "o1-preview-2024-09-12": "2023-10",
-  "o1-preview": "2023-10",
-  "o1-2024-12-17": "2023-10",
-  o1: "2023-10",
-  "o3-mini-2025-01-31": "2023-10",
-  "o3-mini": "2023-10",
+  default: "2023-10",
   // After improvements,
   // it's now easier to add "KnowledgeCutOffDate" instead of stupid hardcoding it, as was done previously.
-  "gemini-pro": "2023-12",
-  "gemini-pro-vision": "2023-12",
+  "gemini-2.5-pro-exp-03-25": "2025-01",
+  "gemini-2.0-flash": "2024-08",
+  "claude-3-7-sonnet-latest": "2024-10",
+  "claude-3-5-haiku-latest": "2024-10",
+  "gpt-4.1": "2024-06",
+  "gpt-4.1-mini": "2024-06",
   "deepseek-chat": "2024-07",
   "deepseek-coder": "2024-07",
 };
@@ -452,11 +442,12 @@ export const DEFAULT_TTS_VOICES = [
 
 export const VISION_MODEL_REGEXES = [
   /vision/,
-  /gpt-4o/,
+  /gpt-4\.1/,
   /claude-3/,
   /gemini-1\.5/,
   /gemini-exp/,
   /gemini-2\.0/,
+  /gemini-2\.5-pro/,
   /learnlm/,
   /qwen-vl/,
   /qwen2-vl/,
@@ -464,77 +455,24 @@ export const VISION_MODEL_REGEXES = [
   /^dall-e-3$/, // Matches exactly "dall-e-3"
   /glm-4v/,
   /vl/i,
+  /o1/,
 ];
 
 export const EXCLUDE_VISION_MODEL_REGEXES = [/claude-3-5-haiku-20241022/];
 
-const openaiModels = [
-  // As of July 2024, gpt-4o-mini should be used in place of gpt-3.5-turbo,
-  // as it is cheaper, more capable, multimodal, and just as fast. gpt-3.5-turbo is still available for use in the API.
-  "gpt-3.5-turbo",
-  "gpt-3.5-turbo-1106",
-  "gpt-3.5-turbo-0125",
-  "gpt-4",
-  "gpt-4-0613",
-  "gpt-4-32k",
-  "gpt-4-32k-0613",
-  "gpt-4-turbo",
-  "gpt-4-turbo-preview",
-  "gpt-4o",
-  "gpt-4o-2024-05-13",
-  "gpt-4o-2024-08-06",
-  "gpt-4o-2024-11-20",
-  "chatgpt-4o-latest",
-  "gpt-4o-mini",
-  "gpt-4o-mini-2024-07-18",
-  "gpt-4-vision-preview",
-  "gpt-4-turbo-2024-04-09",
-  "gpt-4-1106-preview",
-  "dall-e-3",
-  "o1-mini",
-  "o1-preview",
-  "o3-mini",
-];
+const openaiModels = ["dall-e-3", "o1", "o3-mini", "gpt-4.1", "gpt-4.1-mini"];
 
 const googleModels = [
-  "gemini-1.0-pro", // Deprecated on 2/15/2025
-  "gemini-1.5-pro-latest",
-  "gemini-1.5-pro",
-  "gemini-1.5-pro-002",
-  "gemini-1.5-pro-exp-0827",
-  "gemini-1.5-flash-latest",
-  "gemini-1.5-flash-8b-latest",
-  "gemini-1.5-flash",
-  "gemini-1.5-flash-8b",
-  "gemini-1.5-flash-002",
-  "gemini-1.5-flash-exp-0827",
-  "learnlm-1.5-pro-experimental",
-  "gemini-exp-1114",
-  "gemini-exp-1121",
-  "gemini-exp-1206",
   "gemini-2.0-flash",
-  "gemini-2.0-flash-exp",
-  "gemini-2.0-flash-lite-preview-02-05",
-  "gemini-2.0-flash-thinking-exp",
-  "gemini-2.0-flash-thinking-exp-1219",
-  "gemini-2.0-flash-thinking-exp-01-21",
-  "gemini-2.0-pro-exp",
-  "gemini-2.0-pro-exp-02-05",
+  "gemini-2.0-flash-lite",
+  "gemini-2.5-pro-exp-03-25",
 ];
 
 const anthropicModels = [
-  "claude-instant-1.2",
-  "claude-2.0",
-  "claude-2.1",
-  "claude-3-sonnet-20240229",
-  "claude-3-opus-20240229",
   "claude-3-opus-latest",
-  "claude-3-haiku-20240307",
-  "claude-3-5-haiku-20241022",
   "claude-3-5-haiku-latest",
-  "claude-3-5-sonnet-20240620",
-  "claude-3-5-sonnet-20241022",
   "claude-3-5-sonnet-latest",
+  "claude-3-7-sonnet-latest",
 ];
 
 const baiduModels = [
@@ -568,6 +506,9 @@ const alibabaModes = [
   "qwen-max-0403",
   "qwen-max-0107",
   "qwen-max-longcontext",
+  "qwen-omni-turbo",
+  "qwen-vl-plus",
+  "qwen-vl-max",
 ];
 
 const tencentModels = [

app/store/config.ts

@@ -66,14 +66,14 @@ export const DEFAULT_CONFIG = {
   modelConfig: {
     model: "gpt-4o-mini" as ModelType,
     providerName: "OpenAI" as ServiceProvider,
-    temperature: 0.5,
+    temperature: 0.2,
     top_p: 1,
     max_tokens: 4000,
     presence_penalty: 0,
     frequency_penalty: 0,
     sendMemory: true,
-    historyMessageCount: 4,
-    compressMessageLengthThreshold: 1000,
+    historyMessageCount: 20,
+    compressMessageLengthThreshold: 5000,
     compressModel: "",
     compressProviderName: "",
     enableInjectSystemPrompts: true,

app/utils.ts

@@ -304,9 +304,18 @@ export function getTimeoutMSByModel(model: string) {
     model.startsWith("o1") ||
     model.startsWith("o3") ||
     model.includes("deepseek-r") ||
-    model.includes("-thinking")
-  )
+    model.includes("-thinking") ||
+    model.includes("pro")
+  ) {
+    console.log(
+      "thinking model is " +
+        model +
+        " timeout is " +
+        REQUEST_TIMEOUT_MS_FOR_THINKING,
+    );
     return REQUEST_TIMEOUT_MS_FOR_THINKING;
+  }
+  console.log("normal model is " + model + " timeout is " + REQUEST_TIMEOUT_MS);
   return REQUEST_TIMEOUT_MS;
 }
 

app/utils/chat.ts

@@ -3,7 +3,7 @@ import {
   UPLOAD_URL,
   REQUEST_TIMEOUT_MS,
 } from "@/app/constant";
-import { RequestMessage } from "@/app/client/api";
+import { MultimodalContent, RequestMessage } from "@/app/client/api";
 import Locale from "@/app/locales";
 import {
   EventStreamContentType,
@@ -70,8 +70,9 @@ export function compressImage(file: Blob, maxSize: number): Promise<string> {
   });
 }
 
-export async function preProcessImageContent(
+export async function preProcessImageContentBase(
   content: RequestMessage["content"],
+  transformImageUrl: (url: string) => Promise<{ [key: string]: any }>,
 ) {
   if (typeof content === "string") {
     return content;
@@ -81,7 +82,7 @@ export async function preProcessImageContent(
     if (part?.type == "image_url" && part?.image_url?.url) {
       try {
         const url = await cacheImageToBase64Image(part?.image_url?.url);
-        result.push({ type: part.type, image_url: { url } });
+        result.push(await transformImageUrl(url));
       } catch (error) {
         console.error("Error processing image URL:", error);
       }
@@ -92,6 +93,23 @@ export async function preProcessImageContent(
   return result;
 }
 
+export async function preProcessImageContent(
+  content: RequestMessage["content"],
+) {
+  return preProcessImageContentBase(content, async (url) => ({
+    type: "image_url",
+    image_url: { url },
+  })) as Promise<MultimodalContent[] | string>;
+}
+
+export async function preProcessImageContentForAlibabaDashScope(
+  content: RequestMessage["content"],
+) {
+  return preProcessImageContentBase(content, async (url) => ({
+    image: url,
+  }));
+}
+
 const imageCaches: Record<string, string> = {};
 export function cacheImageToBase64Image(imageUrl: string) {
   if (imageUrl.includes(CACHE_URL_PREFIX)) {

jest.config.ts

@@ -15,6 +15,8 @@ const config: Config = {
   moduleNameMapper: {
     "^@/(.*)$": "<rootDir>/$1",
   },
+  extensionsToTreatAsEsm: [".ts", ".tsx"],
+  injectGlobals: true,
 };
 
 // createJestConfig is exported this way to ensure that next/jest can load the Next.js config which is async

jest.setup.ts

@@ -1,24 +1,22 @@
 // Learn more: https://github.com/testing-library/jest-dom
 import "@testing-library/jest-dom";
+import { jest } from "@jest/globals";
 
 global.fetch = jest.fn(() =>
   Promise.resolve({
     ok: true,
     status: 200,
-    json: () => Promise.resolve({}),
+    json: () => Promise.resolve([]),
     headers: new Headers(),
     redirected: false,
     statusText: "OK",
     type: "basic",
     url: "",
-    clone: function () {
-      return this;
-    },
     body: null,
     bodyUsed: false,
     arrayBuffer: () => Promise.resolve(new ArrayBuffer(0)),
     blob: () => Promise.resolve(new Blob()),
     formData: () => Promise.resolve(new FormData()),
     text: () => Promise.resolve(""),
-  }),
+  } as Response),
 );

package.json

@@ -17,8 +17,8 @@
     "prompts": "node ./scripts/fetch-prompts.mjs",
     "prepare": "husky install",
     "proxy-dev": "sh ./scripts/init-proxy.sh && proxychains -f ./scripts/proxychains.conf yarn dev",
-    "test": "jest --watch",
-    "test:ci": "jest --ci"
+    "test": "node --no-warnings --experimental-vm-modules $(yarn bin jest) --watch",
+    "test:ci": "node --no-warnings --experimental-vm-modules $(yarn bin jest) --ci"
   },
   "dependencies": {
     "@fortaine/fetch-event-source": "^3.0.6",

test/model-available.test.ts

@@ -3,7 +3,7 @@ import { isModelNotavailableInServer } from "../app/utils/model";
 describe("isModelNotavailableInServer", () => {
   test("test model will return false, which means the model is available", () => {
     const customModels = "";
-    const modelName = "gpt-4";
+    const modelName = "gpt-4.1";
     const providerNames = "OpenAI";
     const result = isModelNotavailableInServer(
       customModels,

test/vision-model-checker.test.ts

@@ -1,3 +1,4 @@
+import { jest } from "@jest/globals";
 import { isVisionModel } from "../app/utils";
 
 describe("isVisionModel", () => {
@@ -14,10 +15,11 @@ describe("isVisionModel", () => {
 
   test("should identify vision models using regex patterns", () => {
     const visionModels = [
-      "gpt-4-vision",
+      "gpt-4.1",
       "claude-3-opus",
       "gemini-1.5-pro",
       "gemini-2.0",
+      "gemini-2.5-pro",
       "gemini-exp-vision",
       "learnlm-vision",
       "qwen-vl-max",
@@ -50,7 +52,7 @@ describe("isVisionModel", () => {
 
   test("should identify models from VISION_MODELS env var", () => {
     process.env.VISION_MODELS = "custom-vision-model,another-vision-model";
-    
+
     expect(isVisionModel("custom-vision-model")).toBe(true);
     expect(isVisionModel("another-vision-model")).toBe(true);
     expect(isVisionModel("unrelated-model")).toBe(false);
@@ -64,4 +66,4 @@ describe("isVisionModel", () => {
     expect(isVisionModel("unrelated-model")).toBe(false);
     expect(isVisionModel("gpt-4-vision")).toBe(true);
   });
-});
+});