Upload 15 files

.gitattributes

@@ -1,35 +1,35 @@
-*.7z filter=lfs diff=lfs merge=lfs -text
-*.arrow filter=lfs diff=lfs merge=lfs -text
-*.bin filter=lfs diff=lfs merge=lfs -text
-*.bz2 filter=lfs diff=lfs merge=lfs -text
-*.ckpt filter=lfs diff=lfs merge=lfs -text
-*.ftz filter=lfs diff=lfs merge=lfs -text
-*.gz filter=lfs diff=lfs merge=lfs -text
-*.h5 filter=lfs diff=lfs merge=lfs -text
-*.joblib filter=lfs diff=lfs merge=lfs -text
-*.lfs.* filter=lfs diff=lfs merge=lfs -text
-*.mlmodel filter=lfs diff=lfs merge=lfs -text
-*.model filter=lfs diff=lfs merge=lfs -text
-*.msgpack filter=lfs diff=lfs merge=lfs -text
-*.npy filter=lfs diff=lfs merge=lfs -text
-*.npz filter=lfs diff=lfs merge=lfs -text
-*.onnx filter=lfs diff=lfs merge=lfs -text
-*.ot filter=lfs diff=lfs merge=lfs -text
-*.parquet filter=lfs diff=lfs merge=lfs -text
-*.pb filter=lfs diff=lfs merge=lfs -text
-*.pickle filter=lfs diff=lfs merge=lfs -text
-*.pkl filter=lfs diff=lfs merge=lfs -text
-*.pt filter=lfs diff=lfs merge=lfs -text
-*.pth filter=lfs diff=lfs merge=lfs -text
-*.rar filter=lfs diff=lfs merge=lfs -text
-*.safetensors filter=lfs diff=lfs merge=lfs -text
-saved_model/**/* filter=lfs diff=lfs merge=lfs -text
-*.tar.* filter=lfs diff=lfs merge=lfs -text
-*.tar filter=lfs diff=lfs merge=lfs -text
-*.tflite filter=lfs diff=lfs merge=lfs -text
-*.tgz filter=lfs diff=lfs merge=lfs -text
-*.wasm filter=lfs diff=lfs merge=lfs -text
-*.xz filter=lfs diff=lfs merge=lfs -text
-*.zip filter=lfs diff=lfs merge=lfs -text
-*.zst filter=lfs diff=lfs merge=lfs -text
-*tfevents* filter=lfs diff=lfs merge=lfs -text
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text

README.md

@@ -1,176 +1,176 @@
----
-title: PromptLayer API Proxy
-emoji: 🚀
-colorFrom: blue
-colorTo: green
-sdk: docker
-app_port: 3000
----
-<div align="center">
-
-# 🚀 PromptLayer API 代理服务
-
-[![GitHub stars](https://img.shields.io/github/stars/Rfym21/PromptlayerProxy?style=social)](https://github.com/Rfym21/PromptlayerProxy)
-[![Docker Pulls](https://img.shields.io/docker/pulls/rfym21/promptlayer-proxy)](https://hub.docker.com/r/rfym21/promptlayer-proxy)
-
-*一个强大的 PromptLayer API 代理服务，支持多种主流 AI 模型*
-
-**🔗 [交流群](https://t.me/nodejs_project) | 🐳 [Docker Hub](https://hub.docker.com/r/rfym21/promptlayer-proxy)**
-
-</div>
-
-## ✨ 功能特点
-
-<div align="center">
-
-| 功能 | 状态 | 描述 |
-|------|------|------|
-| 🔄 **OpenAI API 兼容** | ✅ | 完全兼容 OpenAI API 格式 |
-| 🌊 **流式输出** | ✅ | 支持实时流式响应 |
-| 🖼️ **图像处理** | ✅ | 支持图像上传和识别 |
-| ⚖️ **负载均衡** | ✅ | 多账户轮询负载均衡 |
-| 🐳 **容器化部署** | ✅ | Docker 一键部署 |
-| 🔄 **自动刷新** | ✅ | 智能 Token 自动刷新 |
-| 🛠️ **Tools 支持** | ✅ | 支持Tools参数 |
-| 🔌 **其他参数(温度,Max_Tokens)** | ✅ | 支持配置其他参数，设置的参数将覆盖默认参数 |
-
-</div>
-
----
-
-## 🤖 支持的模型
-
-<div align="center">
-
-| 🏷️ 模型名称 | 📊 最大输出长度 | 🧠 思考长度 | 📈 类型 |
-|-----------|-------------|---------|-------|
-| 🔮 `claude-3-7-sonnet-20250219` | `64,000` | `-` | Anthropic |
-| 🧠 `claude-3-7-sonnet-20250219-thinking` | `64,000` | `32,000` | Anthropic |
-| 🔮 `claude-sonnet-4-20250514` | `64,000` | `-` | Anthropic |
-| 🧠 `claude-sonnet-4-20250514-thinking` | `64,000` | `32,000` | Anthropic |
-| 🔮 `claude-opus-4-20250514` | `32,000` | `-` | Anthropic |
-| 🧠 `claude-opus-4-20250514-thinking` | `32,000` | `16,000` | Anthropic |
-| 🤖 `o4-mini` | `100,000` | `-` | OpenAI |
-| 🤖 `chatgpt-4o-latest` | `-` | `-` | OpenAI |
-| 🤖 `gpt-4.1` | `-` | `-` | OpenAI |
-| 🤖 `gpt-4.5-preview` | `-` | `-` | OpenAI |
-
-</div>
-
----
-
-## 🚀 快速开始
-
-### 方式一：🐳 Docker Compose（推荐）
-
-#### 📥 **Step 1**: 下载配置文件
-
-```bash
-curl -o docker-compose.yml https://raw.githubusercontent.com/Rfym21/PromptlayerProxy/refs/heads/main/docker-compose.yml
-```
-
-#### ⚙️ **Step 2**: 配置环境变量
-
-在 `docker-compose.yml` 文件中设置以下参数：
-
-```yaml
-services:
-  promptlayer-proxy:
-    image: rfym21/promptlayer-proxy:latest
-    container_name: promptlayer-proxy
-    restart: always
-    ports:
-      - "3000:3000"
-    environment:
-       # 🔐 PromptLayer 账号密码
-      - ACCOUNTS=your_account1:your_password1,your_account2:your_password2...
-       # 🔑 API 认证密钥
-      - AUTH_TOKEN=your_auth_token_here
-```
-
-#### 🚀 **Step 3**: 启动服务
-
-```bash
-docker-compose up -d
-```
-
----
-
-### 方式二：🐳 Docker CLI
-
-```bash
-docker run -d \
-  --name promptlayer-proxy \
-  -p 3000:3000 \
-  -e ACCOUNTS=your_account:your_password \
-  -e AUTH_TOKEN=your_auth_token_here \
-  rfym21/promptlayer-proxy:latest
-```
-
----
-
-### 方式三：🤗 Hugging Face Spaces
-
-1.  **创建 Space**:
-    *   访问 [Hugging Face Spaces](https://huggingface.co/new-space) 并创建一个新的 Space。
-    *   选择 "Docker" 作为 SDK。
-    *   选择适合的硬件（通常免费套餐已足够）。
-
-2.  **配置 Secrets**:
-    *   在 Space 的 "Settings" 标签页中，找到 "Secrets and variables" 部分。
-    *   添加以下 Secrets:
-        *   `ACCOUNTS`: 你的 PromptLayer 账号和密码，格式同 Docker 部署方式，例如 `your_account1:your_password1,your_account2:your_password2`。
-        *   `AUTH_TOKEN`: 你设置的 API 认证密钥。
-
-3.  **上传文件**:
-    *   将本项目的所有文件（包括 `Dockerfile`, `package.json`, `src` 目录等）上传到你的 Space Git 仓库中。
-    *   确保 `README.md` 文件头部的 YAML 配置如下（如果通过 Hugging Face 界面创建 Docker Space，会自动生成类似配置，请确保 `app_port` 为 `3000`）：
-        ```yaml
-        ---
-        title: PromptLayer API Proxy
-        emoji: 🚀
-        colorFrom: blue
-        colorTo: green
-        sdk: docker
-        app_port: 3000
-        ---
-        ```
-
-4.  **部署**:
-    *   Hugging Face Spaces 会自动根据 `Dockerfile` 构建并部署你的应用。
-    *   部署完成后，你的服务将在 Space 的 URL 上可用。
-
----
-
-### 方式四：💻 本地开发
-
-#### 📦 **Step 1**: 安装依赖
-
-```bash
-npm install
-```
-
-#### 📝 **Step 2**: 环境配置
-
-创建 `.env` 文件：
-
-```env
-ACCOUNTS=your_account:your_password
-AUTH_TOKEN=your_auth_token_here
-```
-
-#### 🏃 **Step 3**: 启动开发模式
-
-```bash
-npm run dev
-```
-
----
-
-<div align="center">
-
-## 💬 交流与支持
-
-[![Telegram](https://img.shields.io/badge/Telegram-2CA5E0?style=for-the-badge&logo=telegram&logoColor=white)](https://t.me/nodejs_project)
-
-</div>
+---
+title: PromptLayer API Proxy
+emoji: 🚀
+colorFrom: blue
+colorTo: green
+sdk: docker
+app_port: 3000
+---
+<div align="center">
+
+# 🚀 PromptLayer API 代理服务
+
+[![GitHub stars](https://img.shields.io/github/stars/Rfym21/PromptlayerProxy?style=social)](https://github.com/Rfym21/PromptlayerProxy)
+[![Docker Pulls](https://img.shields.io/docker/pulls/rfym21/promptlayer-proxy)](https://hub.docker.com/r/rfym21/promptlayer-proxy)
+
+*一个强大的 PromptLayer API 代理服务，支持多种主流 AI 模型*
+
+**🔗 [交流群](https://t.me/nodejs_project) | 🐳 [Docker Hub](https://hub.docker.com/r/rfym21/promptlayer-proxy)**
+
+</div>
+
+## ✨ 功能特点
+
+<div align="center">
+
+| 功能 | 状态 | 描述 |
+|------|------|------|
+| 🔄 **OpenAI API 兼容** | ✅ | 完全兼容 OpenAI API 格式 |
+| 🌊 **流式输出** | ✅ | 支持实时流式响应 |
+| 🖼️ **图像处理** | ✅ | 支持图像上传和识别 |
+| ⚖️ **负载均衡** | ✅ | 多账户轮询负载均衡 |
+| 🐳 **容器化部署** | ✅ | Docker 一键部署 |
+| 🔄 **自动刷新** | ✅ | 智能 Token 自动刷新 |
+| 🛠️ **Tools 支持** | ✅ | 支持Tools参数 |
+| 🔌 **其他参数(温度,Max_Tokens)** | ✅ | 支持配置其他参数，设置的参数将覆盖默认参数 |
+
+</div>
+
+---
+
+## 🤖 支持的模型
+
+<div align="center">
+
+| 🏷️ 模型名称 | 📊 最大输出长度 | 🧠 思考长度 | 📈 类型 |
+|-----------|-------------|---------|-------|
+| 🔮 `claude-3-7-sonnet-20250219` | `64,000` | `-` | Anthropic |
+| 🧠 `claude-3-7-sonnet-20250219-thinking` | `64,000` | `32,000` | Anthropic |
+| 🔮 `claude-sonnet-4-20250514` | `64,000` | `-` | Anthropic |
+| 🧠 `claude-sonnet-4-20250514-thinking` | `64,000` | `32,000` | Anthropic |
+| 🔮 `claude-opus-4-20250514` | `32,000` | `-` | Anthropic |
+| 🧠 `claude-opus-4-20250514-thinking` | `32,000` | `16,000` | Anthropic |
+| 🤖 `o4-mini` | `100,000` | `-` | OpenAI |
+| 🤖 `chatgpt-4o-latest` | `-` | `-` | OpenAI |
+| 🤖 `gpt-4.1` | `-` | `-` | OpenAI |
+| 🤖 `gpt-4.5-preview` | `-` | `-` | OpenAI |
+
+</div>
+
+---
+
+## 🚀 快速开始
+
+### 方式一：🐳 Docker Compose（推荐）
+
+#### 📥 **Step 1**: 下载配置文件
+
+```bash
+curl -o docker-compose.yml https://raw.githubusercontent.com/Rfym21/PromptlayerProxy/refs/heads/main/docker-compose.yml
+```
+
+#### ⚙️ **Step 2**: 配置环境变量
+
+在 `docker-compose.yml` 文件中设置以下参数：
+
+```yaml
+services:
+  promptlayer-proxy:
+    image: rfym21/promptlayer-proxy:latest
+    container_name: promptlayer-proxy
+    restart: always
+    ports:
+      - "3000:3000"
+    environment:
+       # 🔐 PromptLayer 账号密码
+      - ACCOUNTS=your_account1:your_password1,your_account2:your_password2...
+       # 🔑 API 认证密钥
+      - AUTH_TOKEN=your_auth_token_here
+```
+
+#### 🚀 **Step 3**: 启动服务
+
+```bash
+docker-compose up -d
+```
+
+---
+
+### 方式二：🐳 Docker CLI
+
+```bash
+docker run -d \
+  --name promptlayer-proxy \
+  -p 3000:3000 \
+  -e ACCOUNTS=your_account:your_password \
+  -e AUTH_TOKEN=your_auth_token_here \
+  rfym21/promptlayer-proxy:latest
+```
+
+---
+
+### 方式三：🤗 Hugging Face Spaces
+
+1.  **创建 Space**:
+    *   访问 [Hugging Face Spaces](https://huggingface.co/new-space) 并创建一个新的 Space。
+    *   选择 "Docker" 作为 SDK。
+    *   选择适合的硬件（通常免费套餐已足够）。
+
+2.  **配置 Secrets**:
+    *   在 Space 的 "Settings" 标签页中，找到 "Secrets and variables" 部分。
+    *   添加以下 Secrets:
+        *   `ACCOUNTS`: 你的 PromptLayer 账号和密码，格式同 Docker 部署方式，例如 `your_account1:your_password1,your_account2:your_password2`。
+        *   `AUTH_TOKEN`: 你设置的 API 认证密钥。
+
+3.  **上传文件**:
+    *   将本项目的所有文件（包括 `Dockerfile`, `package.json`, `src` 目录等）上传到你的 Space Git 仓库中。
+    *   确保 `README.md` 文件头部的 YAML 配置如下（如果通过 Hugging Face 界面创建 Docker Space，会自动生成类似配置，请确保 `app_port` 为 `3000`）：
+        ```yaml
+        ---
+        title: PromptLayer API Proxy
+        emoji: 🚀
+        colorFrom: blue
+        colorTo: green
+        sdk: docker
+        app_port: 3000
+        ---
+        ```
+
+4.  **部署**:
+    *   Hugging Face Spaces 会自动根据 `Dockerfile` 构建并部署你的应用。
+    *   部署完成后，你的服务将在 Space 的 URL 上可用。
+
+---
+
+### 方式四：💻 本地开发
+
+#### 📦 **Step 1**: 安装依赖
+
+```bash
+npm install
+```
+
+#### 📝 **Step 2**: 环境配置
+
+创建 `.env` 文件：
+
+```env
+ACCOUNTS=your_account:your_password
+AUTH_TOKEN=your_auth_token_here
+```
+
+#### 🏃 **Step 3**: 启动开发模式
+
+```bash
+npm run dev
+```
+
+---
+
+<div align="center">
+
+## 💬 交流与支持
+
+[![Telegram](https://img.shields.io/badge/Telegram-2CA5E0?style=for-the-badge&logo=telegram&logoColor=white)](https://t.me/nodejs_project)
+
+</div>

src/routes/chat.js

@@ -1,472 +1,476 @@
-const express = require('express')
-const axios = require('axios')
-const WebSocket = require('ws')
-const router = express.Router()
-const { v4: uuidv4 } = require('uuid')
-const { uploadFileBuffer } = require('../lib/upload')
-const verify = require('./verify')
-const modelMap = require('../lib/model-map')
-
-
-async function parseMessages(req, res, next) {
-  const messages = req.body.messages
-  if (!Array.isArray(messages)) {
-    req.processedMessages = []
-    return next()
-  }
-
-  try {
-    const transformedMessages = await Promise.all(messages.map(async (msg) => {
-      // Determine provider to conditionally convert system role for Anthropic
-      const modelName = req.body.model;
-      const modelData = modelMap[modelName]; // modelMap is required at the top
-      const provider = modelData?.provider; // Use optional chaining for safety
-
-      let message = {
-        role: (msg.role === "system" && provider === "anthropic") ? "user" : msg.role,
-        tool_calls: [],
-        template_format: "jinja2"
-      }
-
-      if (Array.isArray(msg.content)) {
-        const contentItems = await Promise.all(msg.content.map(async (item) => {
-          if (item.type === "text") {
-            return {
-              type: "text",
-              text: item.text
-            }
-          }
-          else if (item.type === "image_url") {
-            try {
-              const base64Match = item.image_url.url.match(/^data:image\/\w+;base64,(.+)$/)
-              if (base64Match) {
-                const base64 = base64Match[1]
-                const data = Buffer.from(base64, 'base64')
-                const uploadResult = await uploadFileBuffer(data)
-
-                return {
-                  type: "media",
-                  media: {
-                    "type": "image",
-                    "url": uploadResult.file_url,
-                    "title": `image_${Date.now()}.png`
-                  }
-                }
-              } else {
-                return {
-                  type: "media",
-                  media: {
-                    "type": "image",
-                    "url": item.image_url.url,
-                    "title": "external_image"
-                  }
-                }
-              }
-            } catch (error) {
-              console.error("处理图像时出错:", error)
-              return {
-                type: "text",
-                text: "[图像处理失败]"
-              }
-            }
-          } else {
-            return {
-              type: "text",
-              text: JSON.stringify(item)
-            }
-          }
-        }))
-
-        message.content = contentItems
-      } else {
-        message.content = [
-          {
-            type: "text",
-            text: msg.content || ""
-          }
-        ]
-      }
-
-      return message
-    }))
-
-    req.body.messages = transformedMessages
-    return next()
-  } catch (error) {
-    console.error("处理消息时出错:", error.status)
-    req.body.messages = []
-    return next(error)
-  }
-}
-
-async function getChatID(req, res) {
-  try {
-    const url = 'https://api.promptlayer.com/api/dashboard/v2/workspaces/' + req.account.workspaceId + '/playground_sessions'
-    const headers = { Authorization: "Bearer " + req.account.token }
-    const model_data = modelMap[req.body.model] ? modelMap[req.body.model] : modelMap["claude-3-7-sonnet-20250219"]
-    let data = {
-      "id": uuidv4(),
-      "name": "Not implemented",
-      "prompt_blueprint": {
-        "inference_client_name": null,
-        "metadata": {
-          "model": model_data
-        },
-        "prompt_template": {
-          "type": "chat",
-          "messages": req.body.messages,
-          "tools": req.body.tools || [],
-          "tool_choice": req.body.tool_choice || "none",
-          "input_variables": [],
-          "functions": [],
-          "function_call": null
-        },
-        "provider_base_url_name": null
-      },
-      "input_variables": []
-    }
-
-    for (const item in req.body) {
-      if (item === "messages" || item === "model" || item === "stream") {
-        continue
-      }  else if (model_data.parameters[item]) {
-        model_data.parameters[item] = req.body[item]
-      }
-    }
-    data.prompt_blueprint.metadata.model = model_data
-    console.log(`模型参数: ${data.prompt_blueprint.metadata.model}`)
-
-    const response = await axios.put(url, data, { headers })
-    if (response.data.success) {
-      console.log(`生成会话ID成功: ${response.data.playground_session.id}`)
-      req.chatID = response.data.playground_session.id
-      return response.data.playground_session.id
-    } else {
-      return false
-    }
-  } catch (error) {
-    // console.error("错误:", error.response?.data)
-    res.status(500).json({
-      "error": {
-        "message": error.message || "服务器内部错误",
-        "type": "server_error",
-        "param": null,
-        "code": "server_error"
-      }
-    })
-    return false
-  }
-}
-
-async function sentRequest(req, res) {
-  try {
-    const url = 'https://api.promptlayer.com/api/dashboard/v2/workspaces/' + req.account.workspaceId + '/run_groups'
-    const headers = { Authorization: "Bearer " + req.account.token }
-    const model_data = modelMap[req.body.model] ? modelMap[req.body.model] : modelMap["claude-3-7-sonnet-20250219"];
-    const provider = model_data?.provider; // Get provider
-
-    // Base prompt template structure
-    let prompt_template = {
-      "type": "chat",
-      "messages": req.body.messages,
-      "tools": req.body.tools || [], // Default value
-      "tool_choice": req.body.tool_choice || "none", // Default value
-      "input_variables": [],
-      "functions": [],
-      "function_call": null
-    };
-
-    // Conditionally modify for Mistral/Cohere
-    if (provider === 'mistral' || provider === 'cohere') {
-      prompt_template.tools = null;
-      delete prompt_template.tool_choice; // Remove tool_choice entirely
-      delete prompt_template.function_call;
-    }
-
-    let data = {
-      "id": uuidv4(),
-      "playground_session_id": req.chatID,
-      "shared_prompt_blueprint": {
-        "inference_client_name": null,
-        "metadata": {
-          "model": model_data // Keep original model_data here for metadata
-        },
-        "prompt_template": prompt_template, // Use the adjusted template
-        "provider_base_url_name": null
-      },
-      "individual_run_requests": [
-        {
-          "input_variables": {},
-          "run_group_position": 1
-        }
-      ]
-    };
-
-    console.log(JSON.stringify(data))
-
-    // Update model parameters (this loop remains the same)
-    for (const item in req.body) {
-      if (item === "messages" || item === "model" || item === "stream") {
-        continue
-      } else if (model_data.parameters && model_data.parameters.hasOwnProperty(item)) { // Check if parameters exist and has the property
-        model_data.parameters[item] = req.body[item]
-      }
-    }
-    // Ensure the potentially modified model_data (with updated parameters) is in metadata
-    data.shared_prompt_blueprint.metadata.model = model_data;
-
-    const response = await axios.post(url, data, { headers });
-    if (response.data.success) {
-      return response.data.run_group.individual_run_requests[0].id
-    } else {
-      return false
-    }
-  } catch (error) {
-    // console.error("错误:", error.response?.data)
-    res.status(500).json({
-      "error": {
-        "message": error.message || "服务器内部错误",
-        "type": "server_error",
-        "param": null,
-        "code": "server_error"
-      }
-    })
-  }
-}
-
-// 聊天完成路由
-router.post('/v1/chat/completions', verify, parseMessages, async (req, res) => {
-  // console.log(JSON.stringify(req.body))
-
-  try {
-
-    const setHeader = () => {
-      try {
-        if (req.body.stream === true) {
-          res.setHeader('Content-Type', 'text/event-stream')
-          res.setHeader('Cache-Control', 'no-cache')
-          res.setHeader('Connection', 'keep-alive')
-        } else {
-          res.setHeader('Content-Type', 'application/json')
-        }
-      } catch (error) {
-        // console.error("设置响应头时出错:", error)
-      }
-    }
-
-    const { access_token, clientId } = req.account
-    // 生成会话ID
-    await getChatID(req, res)
-
-    // 发送的数据
-    const sendAction = `{"action":10,"channel":"user:${clientId}","params":{"agent":"react-hooks/2.0.2"}}`
-    // 构建 WebSocket URL
-    const wsUrl = `wss://realtime.ably.io/?access_token=${encodeURIComponent(access_token)}&clientId=${clientId}&format=json&heartbeats=true&v=3&agent=ably-js%2F2.0.2%20browser`
-    // 创建 WebSocket 连接
-    const ws = new WebSocket(wsUrl)
-
-    // 状态详细
-    let ThinkingLastContent = ""
-    let TextLastContent = ""
-    let ThinkingStart = false
-    let ThinkingEnd = false
-    let RequestID = ""
-    let MessageID = "chatcmpl-" + uuidv4()
-    let streamChunk = {
-      "id": MessageID,
-      "object": "chat.completion.chunk",
-      "system_fingerprint": "fp_44709d6fcb",
-      "created": Math.floor(Date.now() / 1000),
-      "model": req.body.model,
-      "choices": [
-        {
-          "index": 0,
-          "delta": {
-            "content": null
-          },
-          "finish_reason": null
-        }
-      ]
-    }
-
-    let pingInterval;
-
-    ws.on('open', async () => {
-      ws.send(sendAction)
-      RequestID = await sentRequest(req, res)
-      setHeader()
-      // Start sending pings every 30 seconds to keep the connection alive
-      pingInterval = setInterval(() => {
-        if (ws.readyState === WebSocket.OPEN) {
-          ws.ping(() => {}); // Empty callback, just to send the ping
-        }
-      }, 15000);
-    })
-
-    ws.on('message', async (data) => {
-      try {
-        data = data.toString()
-        console.log("here!!!")
-        console.log(data)
-        let ContentText = JSON.parse(data)?.messages?.[0]
-        let ContentData = JSON.parse(ContentText?.data)
-        const isRequestID = ContentData?.individual_run_request_id
-        if (isRequestID != RequestID || !isRequestID) return
-
-        let output = ""
-
-        if (ContentText?.name === "UPDATE_LAST_MESSAGE") {
-          const MessageArray = ContentData?.payload?.message?.content
-          for (const item of MessageArray) {
-
-            if (item.type === "text") {
-              output = item.text.replace(TextLastContent, "")
-              if (ThinkingStart && !ThinkingEnd) {
-                ThinkingEnd = true
-                output = `${output}\n\n</think>`
-              }
-              TextLastContent = item.text
-            }
-            else if (item.type === "thinking" && MessageArray.length === 1) {
-              output = item.thinking.replace(ThinkingLastContent, "")
-              if (!ThinkingStart) {
-                ThinkingStart = true
-                output = `<think>\n\n${output}`
-              }
-              ThinkingLastContent = item.thinking
-            }
-
-          }
-
-          if (req.body.stream === true) {
-            streamChunk.choices[0].delta.content = output
-            res.write(`data: ${JSON.stringify(streamChunk)}\n\n`)
-          }
-
-        }
-        else if (ContentText?.name === "INDIVIDUAL_RUN_COMPLETE") {
-
-          if (req.body.stream !== true) {
-            output = ThinkingLastContent ? `<think>\n\n${ThinkingLastContent}\n\n</think>\n\n${TextLastContent}` : TextLastContent
-          }
-
-          if (ThinkingLastContent === "" && TextLastContent === "") {
-            output = "该模型在发送请求时遇到错误: \n1. 请检查请求参数,模型支持参数和默认参数可在/v1/models下查看\n2. 参数设置大小是否超过模型限制\n3. 模型当前官网此模型可能负载过高,可以切换别的模型尝试,这属于正常现象\n4. Anthropic系列模型的temperature的取值为0-1,请勿设置超过1的值\n5. 交流与支持群: https://t.me/nodejs_project"
-            streamChunk.choices[0].delta.content = output
-            res.write(`data: ${JSON.stringify(streamChunk)}\n\n`)
-          }
-
-          if (!req.body.stream || req.body.stream !== true) {
-            let responseJson = {
-              "id": MessageID,
-              "object": "chat.completion",
-              "created": Math.floor(Date.now() / 1000),
-              "system_fingerprint": "fp_44709d6fcb",
-              "model": req.body.model,
-              "choices": [
-                {
-                  "index": 0,
-                  "message": {
-                    "role": "assistant",
-                    "content": output
-                  },
-                  "finish_reason": "stop"
-                }
-              ],
-              "usage": {
-                "prompt_tokens": 0,
-                "completion_tokens": 0,
-                "total_tokens": 0
-              }
-            }
-
-            res.json(responseJson)
-            ws.close()
-            return
-          } else {
-            // 流式响应：发送结束标记
-            let finalChunk = {
-              "id": MessageID,
-              "object": "chat.completion.chunk",
-              "system_fingerprint": "fp_44709d6fcb",
-              "created": Math.floor(Date.now() / 1000),
-              "model": req.body.model,
-              "choices": [
-                {
-                  "index": 0,
-                  "delta": {},
-                  "finish_reason": "stop"
-                }
-              ]
-            }
-
-            res.write(`data: ${JSON.stringify(finalChunk)}\n\n`)
-            res.write(`data: [DONE]\n\n`)
-            res.end()
-          }
-          ws.close()
-        }
-
-      } catch (err) {
-        // console.error("处理WebSocket消息出错:", err)
-      }
-    })
-
-    ws.on('error', (err) => {
-      clearInterval(pingInterval); // Stop sending pings
-      // 标准OpenAI错误响应格式
-      res.status(500).json({
-        "error": {
-          "message": err.message,
-          "type": "server_error",
-          "param": null,
-          "code": "server_error"
-        }
-      })
-    })
-
-    const oSeriesModels = ["o4-mini", "o4-mini-high", "o3-mini", "o3-mini-high", "o1", "o3", "o3-2025-04-16", "o4-mini-2025-04-16"];
-    let timeoutDuration = 300 * 1000; // 默认5分钟
-
-    if (oSeriesModels.includes(req.body.model)) {
-      timeoutDuration = 1200 * 1000; // o系列模型20分钟
-    }
-
-    const requestTimeout = setTimeout(() => {
-      clearInterval(pingInterval); // Stop sending pings
-      if (ws.readyState === WebSocket.OPEN) {
-        ws.close()
-        if (!res.headersSent) {
-          // 标准OpenAI超时错误响应格式
-          res.status(504).json({
-            "error": {
-              "message": "请求超时",
-              "type": "timeout",
-              "param": null,
-              "code": "timeout_error"
-            }
-          })
-        }
-      }
-    }, timeoutDuration)
-
-    ws.on('close', () => {
-      clearInterval(pingInterval); // Stop sending pings
-      clearTimeout(requestTimeout); // Clear the main request timeout as well if ws closes first
-    });
-
-  } catch (error) {
-    console.error("错误:", error)
-    // 标准OpenAI通用错误响应格式
-    res.status(500).json({
-      "error": {
-        "message": error.message || "服务器内部错误",
-        "type": "server_error",
-        "param": null,
-        "code": "server_error"
-      }
-    })
-  }
-})
-
-module.exports = router
+const express = require('express')
+const axios = require('axios')
+const WebSocket = require('ws')
+const router = express.Router()
+const { v4: uuidv4 } = require('uuid')
+const { uploadFileBuffer } = require('../lib/upload')
+const verify = require('./verify')
+const modelMap = require('../lib/model-map')
+
+
+async function parseMessages(req, res, next) {
+  const messages = req.body.messages
+  if (!Array.isArray(messages)) {
+    req.processedMessages = []
+    return next()
+  }
+
+  try {
+    const transformedMessages = await Promise.all(messages.map(async (msg) => {
+      // Determine provider to conditionally convert system role for Anthropic
+      const modelName = req.body.model;
+      const modelData = modelMap[modelName]; // modelMap is required at the top
+      const provider = modelData?.provider; // Use optional chaining for safety
+
+      let message = {
+        role: (msg.role === "system" && provider === "anthropic") ? "user" : msg.role,
+        tool_calls: [],
+        template_format: "jinja2"
+      }
+
+      if (Array.isArray(msg.content)) {
+        const contentItems = await Promise.all(msg.content.map(async (item) => {
+          if (item.type === "text") {
+            return {
+              type: "text",
+              text: item.text
+            }
+          }
+          else if (item.type === "image_url") {
+            try {
+              const base64Match = item.image_url.url.match(/^data:image\/\w+;base64,(.+)$/)
+              if (base64Match) {
+                const base64 = base64Match[1]
+                const data = Buffer.from(base64, 'base64')
+                const uploadResult = await uploadFileBuffer(data)
+
+                return {
+                  type: "media",
+                  media: {
+                    "type": "image",
+                    "url": uploadResult.file_url,
+                    "title": `image_${Date.now()}.png`
+                  }
+                }
+              } else {
+                return {
+                  type: "media",
+                  media: {
+                    "type": "image",
+                    "url": item.image_url.url,
+                    "title": "external_image"
+                  }
+                }
+              }
+            } catch (error) {
+              console.error("处理图像时出错:", error)
+              return {
+                type: "text",
+                text: "[图像处理失败]"
+              }
+            }
+          } else {
+            return {
+              type: "text",
+              text: JSON.stringify(item)
+            }
+          }
+        }))
+
+        message.content = contentItems
+      } else {
+        message.content = [
+          {
+            type: "text",
+            text: msg.content || ""
+          }
+        ]
+      }
+
+      return message
+    }))
+
+    req.body.messages = transformedMessages
+    return next()
+  } catch (error) {
+    console.error("处理消息时出错:", error.status)
+    req.body.messages = []
+    return next(error)
+  }
+}
+
+async function getChatID(req, res) {
+  try {
+    const url = 'https://api.promptlayer.com/api/dashboard/v2/workspaces/' + req.account.workspaceId + '/playground_sessions'
+    const headers = { Authorization: "Bearer " + req.account.token }
+    const model_data = modelMap[req.body.model] ? modelMap[req.body.model] : modelMap["claude-3-7-sonnet-20250219"]
+    let data = {
+      "id": uuidv4(),
+      "name": "Not implemented",
+      "prompt_blueprint": {
+        "inference_client_name": null,
+        "metadata": {
+          "model": model_data
+        },
+        "prompt_template": {
+          "type": "chat",
+          "messages": req.body.messages,
+          "tools": req.body.tools || [],
+          "tool_choice": req.body.tool_choice || "none",
+          "input_variables": [],
+          "functions": [],
+          "function_call": null
+        },
+        "provider_base_url_name": null
+      },
+      "input_variables": []
+    }
+
+    for (const item in req.body) {
+      if (item === "messages" || item === "model" || item === "stream") {
+        continue
+      }  else if (model_data.parameters[item]) {
+        model_data.parameters[item] = req.body[item]
+      }
+    }
+    data.prompt_blueprint.metadata.model = model_data
+    console.log(`模型参数: ${data.prompt_blueprint.metadata.model}`)
+
+    const response = await axios.put(url, data, { headers })
+    if (response.data.success) {
+      console.log(`生成会话ID成功: ${response.data.playground_session.id}`)
+      req.chatID = response.data.playground_session.id
+      return response.data.playground_session.id
+    } else {
+      return false
+    }
+  } catch (error) {
+    // console.error("错误:", error.response?.data)
+    res.status(500).json({
+      "error": {
+        "message": error.message || "服务器内部错误",
+        "type": "server_error",
+        "param": null,
+        "code": "server_error"
+      }
+    })
+    return false
+  }
+}
+
+async function sentRequest(req, res) {
+  try {
+    const url = 'https://api.promptlayer.com/api/dashboard/v2/workspaces/' + req.account.workspaceId + '/run_groups'
+    const headers = { Authorization: "Bearer " + req.account.token }
+    const model_data = modelMap[req.body.model] ? modelMap[req.body.model] : modelMap["claude-3-7-sonnet-20250219"];
+    const provider = model_data?.provider; // Get provider
+
+    // Base prompt template structure
+    let prompt_template = {
+      "type": "chat",
+      "messages": req.body.messages,
+      "tools": req.body.tools || [], // Default value
+      "tool_choice": req.body.tool_choice || "none", // Default value
+      "input_variables": [],
+      "functions": [],
+      "function_call": null
+    };
+
+    // Conditionally modify for Mistral/Cohere
+    if (provider === 'mistral' || provider === 'cohere') {
+      prompt_template.tools = null;
+      delete prompt_template.tool_choice; // Remove tool_choice entirely
+      delete prompt_template.function_call;
+    }
+
+    let data = {
+      "id": uuidv4(),
+      "playground_session_id": req.chatID,
+      "shared_prompt_blueprint": {
+        "inference_client_name": null,
+        "metadata": {
+          "model": model_data // Keep original model_data here for metadata
+        },
+        "prompt_template": prompt_template, // Use the adjusted template
+        "provider_base_url_name": null
+      },
+      "individual_run_requests": [
+        {
+          "input_variables": {},
+          "run_group_position": 1
+        }
+      ]
+    };
+
+    console.log(JSON.stringify(data))
+
+    // Update model parameters (this loop remains the same)
+    for (const item in req.body) {
+      if (item === "messages" || item === "model" || item === "stream") {
+        continue
+      } else if (model_data.parameters && model_data.parameters.hasOwnProperty(item)) { // Check if parameters exist and has the property
+        model_data.parameters[item] = req.body[item]
+      }
+    }
+    // Ensure the potentially modified model_data (with updated parameters) is in metadata
+    data.shared_prompt_blueprint.metadata.model = model_data;
+
+    const response = await axios.post(url, data, { headers });
+    if (response.data.success) {
+      return response.data.run_group.individual_run_requests[0].id
+    } else {
+      return false
+    }
+  } catch (error) {
+    // console.error("错误:", error.response?.data)
+    res.status(500).json({
+      "error": {
+        "message": error.message || "服务器内部错误",
+        "type": "server_error",
+        "param": null,
+        "code": "server_error"
+      }
+    })
+  }
+}
+
+// 聊天完成路由
+router.post('/v1/chat/completions', verify, parseMessages, async (req, res) => {
+  // console.log(JSON.stringify(req.body))
+
+  try {
+
+    const setHeader = () => {
+      try {
+        if (req.body.stream === true) {
+          res.setHeader('Content-Type', 'text/event-stream')
+          res.setHeader('Cache-Control', 'no-cache')
+          res.setHeader('Connection', 'keep-alive')
+        } else {
+          res.setHeader('Content-Type', 'application/json')
+        }
+      } catch (error) {
+        // console.error("设置响应头时出错:", error)
+      }
+    }
+
+    const { access_token, clientId } = req.account
+    // 生成会话ID
+    await getChatID(req, res)
+
+    // 发送的数据
+    const sendAction = `{"action":10,"channel":"user:${clientId}","params":{"agent":"react-hooks/2.0.2"}}`
+    // 构建 WebSocket URL
+    const wsUrl = `wss://realtime.ably.io/?access_token=${encodeURIComponent(access_token)}&clientId=${clientId}&format=json&heartbeats=true&v=3&agent=ably-js%2F2.0.2%20browser`
+    // 创建 WebSocket 连接
+    const ws = new WebSocket(wsUrl)
+
+    // 状态详细
+    let ThinkingLastContent = ""
+    let TextLastContent = ""
+    let ThinkingStart = false
+    let ThinkingEnd = false
+    let RequestID = ""
+    let MessageID = "chatcmpl-" + uuidv4()
+    let streamChunk = {
+      "id": MessageID,
+      "object": "chat.completion.chunk",
+      "system_fingerprint": "fp_44709d6fcb",
+      "created": Math.floor(Date.now() / 1000),
+      "model": req.body.model,
+      "choices": [
+        {
+          "index": 0,
+          "delta": {
+            "content": null
+          },
+          "finish_reason": null
+        }
+      ]
+    }
+
+    let pingInterval;
+
+    ws.on('open', async () => {
+      ws.send(sendAction)
+      RequestID = await sentRequest(req, res)
+      setHeader()
+      // Start sending pings every 30 seconds to keep the connection alive
+      pingInterval = setInterval(() => {
+        if (ws.readyState === WebSocket.OPEN) {
+          ws.ping(() => {}); // Empty callback, just to send the ping
+        }
+      }, 15000);
+    })
+
+    ws.on('message', async (data) => {
+      try {
+        data = data.toString()
+        console.log("here!!!")
+        console.log(data)
+        let ContentText = JSON.parse(data)?.messages?.[0]
+        let ContentData = JSON.parse(ContentText?.data)
+        const isRequestID = ContentData?.individual_run_request_id
+        if (isRequestID != RequestID || !isRequestID) return
+
+        let output = ""
+
+        if (ContentText?.name === "UPDATE_LAST_MESSAGE") {
+          const MessageArray = ContentData?.payload?.message?.content
+          for (const item of MessageArray) {
+
+            if (item.type === "text") {
+              output = item.text.replace(TextLastContent, "")
+              if (ThinkingStart && !ThinkingEnd) {
+                ThinkingEnd = true
+                output = `${output}\n\n</think>`
+              }
+              TextLastContent = item.text
+            }
+            else if (item.type === "thinking" && MessageArray.length === 1) {
+              output = item.thinking.replace(ThinkingLastContent, "")
+              if (!ThinkingStart) {
+                ThinkingStart = true
+                output = `<think>\n\n${output}`
+              }
+              ThinkingLastContent = item.thinking
+            }
+
+          }
+
+          if (req.body.stream === true) {
+            streamChunk.choices[0].delta.content = output
+            res.write(`data: ${JSON.stringify(streamChunk)}\n\n`)
+          }
+
+        }
+        else if (ContentText?.name === "INDIVIDUAL_RUN_COMPLETE") {
+
+          if (req.body.stream !== true) {
+            output = ThinkingLastContent ? `<think>\n\n${ThinkingLastContent}\n\n</think>\n\n${TextLastContent}` : TextLastContent
+          }
+
+          if (ThinkingLastContent === "" && TextLastContent === "") {
+            const modelName = req.body.model;
+            const modelData = modelMap[modelName] || modelMap["claude-3-7-sonnet-20250219"]; // Fallback to default if model not found
+            const provider = modelData.provider || "anthropic"; // Default to anthropic if provider not found
+            const providerUpperCase = provider.charAt(0).toUpperCase() + provider.slice(1);
+            output = `${provider}.BadRequestError: Error code: 400 - {'type': 'error', 'error': {'type': 'invalid_request_error', 'message': 'Your credit balance is too low to access the ${providerUpperCase} API. Please go to Plans & Billing to upgrade or purchase credits.'}}`
+            streamChunk.choices[0].delta.content = output
+            res.write(`data: ${JSON.stringify(streamChunk)}\n\n`)
+          }
+
+          if (!req.body.stream || req.body.stream !== true) {
+            let responseJson = {
+              "id": MessageID,
+              "object": "chat.completion",
+              "created": Math.floor(Date.now() / 1000),
+              "system_fingerprint": "fp_44709d6fcb",
+              "model": req.body.model,
+              "choices": [
+                {
+                  "index": 0,
+                  "message": {
+                    "role": "assistant",
+                    "content": output
+                  },
+                  "finish_reason": "stop"
+                }
+              ],
+              "usage": {
+                "prompt_tokens": 0,
+                "completion_tokens": 0,
+                "total_tokens": 0
+              }
+            }
+
+            res.json(responseJson)
+            ws.close()
+            return
+          } else {
+            // 流式响应：发送结束标记
+            let finalChunk = {
+              "id": MessageID,
+              "object": "chat.completion.chunk",
+              "system_fingerprint": "fp_44709d6fcb",
+              "created": Math.floor(Date.now() / 1000),
+              "model": req.body.model,
+              "choices": [
+                {
+                  "index": 0,
+                  "delta": {},
+                  "finish_reason": "stop"
+                }
+              ]
+            }
+
+            res.write(`data: ${JSON.stringify(finalChunk)}\n\n`)
+            res.write(`data: [DONE]\n\n`)
+            res.end()
+          }
+          ws.close()
+        }
+
+      } catch (err) {
+        // console.error("处理WebSocket消息出错:", err)
+      }
+    })
+
+    ws.on('error', (err) => {
+      clearInterval(pingInterval); // Stop sending pings
+      // 标准OpenAI错误响应格式
+      res.status(500).json({
+        "error": {
+          "message": err.message,
+          "type": "server_error",
+          "param": null,
+          "code": "server_error"
+        }
+      })
+    })
+
+    const oSeriesModels = ["o4-mini", "o4-mini-high", "o3-mini", "o3-mini-high", "o1", "o3", "o3-2025-04-16", "o4-mini-2025-04-16"];
+    let timeoutDuration = 300 * 1000; // 默认5分钟
+
+    if (oSeriesModels.includes(req.body.model)) {
+      timeoutDuration = 1200 * 1000; // o系列模型20分钟
+    }
+
+    const requestTimeout = setTimeout(() => {
+      clearInterval(pingInterval); // Stop sending pings
+      if (ws.readyState === WebSocket.OPEN) {
+        ws.close()
+        if (!res.headersSent) {
+          // 标准OpenAI超时错误响应格式
+          res.status(504).json({
+            "error": {
+              "message": "请求超时",
+              "type": "timeout",
+              "param": null,
+              "code": "timeout_error"
+            }
+          })
+        }
+      }
+    }, timeoutDuration)
+
+    ws.on('close', () => {
+      clearInterval(pingInterval); // Stop sending pings
+      clearTimeout(requestTimeout); // Clear the main request timeout as well if ws closes first
+    });
+
+  } catch (error) {
+    console.error("错误:", error)
+    // 标准OpenAI通用错误响应格式
+    res.status(500).json({
+      "error": {
+        "message": error.message || "服务器内部错误",
+        "type": "server_error",
+        "param": null,
+        "code": "server_error"
+      }
+    })
+  }
+})
+
+module.exports = router