chore: 更新版本号至 2.1.9

- 修改了 response_handler.py 中的 _handle_openai_stream_response 和 _handle_openai_normal_response 方法，增加了对推理内容 (reasoning_content) 的支持。
- 更新了 _extract_result 方法的返回值，确保能够提取推理内容。
- 在 gemini_chat_service.py 和 openai_chat_service.py 中，调整了生成配置以包含思考过程的选项。
- 在 vertex_express_chat_service.py 中，增强了对客户端思考配置的处理逻辑，确保优先使用客户端提供的配置。

.env.example

@@ -1,12 +1,19 @@
-# MySQL数据库配置
+# 数据库配置
+DATABASE_TYPE=mysql
+#SQLITE_DATABASE=default_db
 MYSQL_HOST=gemini-balance-mysql
+#MYSQL_SOCKET=/run/mysqld/mysqld.sock
 MYSQL_PORT=3306
 MYSQL_USER=gemini
 MYSQL_PASSWORD=change_me
 MYSQL_DATABASE=default_db
 API_KEYS=["AIzaSyxxxxxxxxxxxxxxxxxxx","AIzaSyxxxxxxxxxxxxxxxxxxx"]
 ALLOWED_TOKENS=["sk-123456"]
-# AUTH_TOKEN=sk-123456
+AUTH_TOKEN=sk-123456
+# For Vertex AI Platform API Keys
+VERTEX_API_KEYS=["AQ.Abxxxxxxxxxxxxxxxxxxx"]
+# For Vertex AI Platform Express API Base URL
+VERTEX_EXPRESS_BASE_URL=https://aiplatform.googleapis.com/v1beta1/publishers/google
 TEST_MODEL=gemini-1.5-flash
 THINKING_MODELS=["gemini-2.5-flash-preview-04-17"]
 THINKING_BUDGET_MAP={"gemini-2.5-flash-preview-04-17": 4000}
@@ -26,6 +33,8 @@ TIME_OUT=300
 # 代理服务器配置 (支持 http 和 socks5)
 # 示例: PROXIES=["http://user:pass@host:port", "socks5://host:port"]
 PROXIES=[]
+# 对同一个API_KEY使用代理列表中固定的IP策略
+PROXIES_USE_CONSISTENCY_HASH_BY_API_KEY=true
 #########################image_generate 相关配置###########################
 PAID_KEY=AIzaSyxxxxxxxxxxxxxxxxxxx
 CREATE_IMAGE_MODEL=imagen-3.0-generate-002
@@ -34,6 +43,7 @@ SMMS_SECRET_TOKEN=XXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
 PICGO_API_KEY=xxxx
 CLOUDFLARE_IMGBED_URL=https://xxxxxxx.pages.dev/upload
 CLOUDFLARE_IMGBED_AUTH_CODE=xxxxxxxxx
+CLOUDFLARE_IMGBED_UPLOAD_FOLDER=
 ##########################################################################
 #########################stream_optimizer 相关配置########################
 STREAM_OPTIMIZER_ENABLED=false
@@ -46,8 +56,27 @@ STREAM_CHUNK_SIZE=5
 ######################### 日志配置 #######################################
 # 日志级别 (debug, info, warning, error, critical)，默认为 info
 LOG_LEVEL=info
+# 是否开启自动删除错误日志
+AUTO_DELETE_ERROR_LOGS_ENABLED=true
+# 自动删除多少天前的错误日志 (1, 7, 30)
+AUTO_DELETE_ERROR_LOGS_DAYS=7
+# 是否开启自动删除请求日志
+AUTO_DELETE_REQUEST_LOGS_ENABLED=false
+# 自动删除多少天前的请求日志 (1, 7, 30)
+AUTO_DELETE_REQUEST_LOGS_DAYS=30
 ##########################################################################
 
+# 假流式配置 (Fake Streaming Configuration)
+# 是否启用假流式输出
+FAKE_STREAM_ENABLED=True
+# 假流式发送空数据的间隔时间（秒）
+FAKE_STREAM_EMPTY_DATA_INTERVAL_SECONDS=5
+
 # 安全设置 (JSON 字符串格式)
 # 注意：这里的示例值可能需要根据实际模型支持情况调整
-SAFETY_SETTINGS='[{"category": "HARM_CATEGORY_HARASSMENT", "threshold": "BLOCK_NONE"}, {"category": "HARM_CATEGORY_HATE_SPEECH", "threshold": "BLOCK_NONE"}, {"category": "HARM_CATEGORY_SEXUALLY_EXPLICIT", "threshold": "BLOCK_NONE"}, {"category": "HARM_CATEGORY_DANGEROUS_CONTENT", "threshold": "BLOCK_NONE"}]'
+SAFETY_SETTINGS=[{"category": "HARM_CATEGORY_HARASSMENT", "threshold": "OFF"}, {"category": "HARM_CATEGORY_HATE_SPEECH", "threshold": "OFF"}, {"category": "HARM_CATEGORY_SEXUALLY_EXPLICIT", "threshold": "OFF"}, {"category": "HARM_CATEGORY_DANGEROUS_CONTENT", "threshold": "OFF"}, {"category": "HARM_CATEGORY_CIVIC_INTEGRITY", "threshold": "BLOCK_NONE"}]
+URL_NORMALIZATION_ENABLED=false
+# tts配置
+TTS_MODEL=gemini-2.5-flash-preview-tts
+TTS_VOICE_NAME=Zephyr
+TTS_SPEED=normal

.github/workflows/release.yml

@@ -3,7 +3,7 @@ name: Publish Release
 on:
   push:
     tags:
-      - 'v*' # 当推送以 "v" 开头的标签时触发（如 v1.0.0, v2.1.0）
+      - "v*" # 当推送以 "v" 开头的标签时触发（如 v1.0.0, v2.1.0）
 
 jobs:
   update-release-draft:
@@ -15,8 +15,17 @@ jobs:
       # Step 1: 检出代码库
       - name: Checkout code
         uses: actions/checkout@v3
+        with:
+          fetch-depth: 0
 
-      # Step 2: 自动生成 Release
+      # Step 2: 自动生成 Release Notes
+      - name: Generate release notes
+        id: changelog
+        uses: mikepenz/release-changelog-builder-action@v4
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+
+      # Step 3: 自动生成 Release
       - name: Create Release
         id: create_release
         uses: actions/create-release@v1
@@ -25,15 +34,16 @@ jobs:
         with:
           tag_name: ${{ github.ref_name }}
           release_name: ${{ github.ref_name }}
+          body: ${{ steps.changelog.outputs.changelog }}
           draft: false
           prerelease: false
-      
-      # Step 3: 可选，构建zip文件
+
+      # Step 4: 可选，构建zip文件
       - name: Create ZIP file
         run: |
           zip -r gemini-balance.zip . -x "*.git*" "*.github*" "*.env*" "logs/*" "tests/*"
 
-      # Step 4: 可选，上传构建文件
+      # Step 5: 可选，上传构建文件
       - name: Upload Release Asset
         uses: actions/upload-release-asset@v1
         env:
@@ -41,5 +51,5 @@ jobs:
         with:
           upload_url: ${{ steps.create_release.outputs.upload_url }}
           asset_path: ./gemini-balance.zip # 替换为你的构建文件路径
-          asset_name: gemini-balance.zip  # 替换为你的文件名
+          asset_name: gemini-balance.zip # 替换为你的文件名
           asset_content_type: application/zip

.gitignore

@@ -257,4 +257,5 @@ $RECYCLE.BIN/
 
 # Custom rules (everything added below won't be overriden by 'Generate .gitignore File' if you use 'Update' option)
 
-tests/
+tests/
+default_db

Dockerfile

@@ -14,6 +14,8 @@ ENV BASE_URL=https://generativelanguage.googleapis.com/v1beta
 ENV TOOLS_CODE_EXECUTION_ENABLED=false
 ENV IMAGE_MODELS='["gemini-2.0-flash-exp"]'
 ENV SEARCH_MODELS='["gemini-2.0-flash-exp","gemini-2.0-pro-exp"]'
+ENV URL_NORMALIZATION_ENABLED=false
+ENV CLOUDFLARE_IMGBED_UPLOAD_FOLDER=""
 
 # Expose port
 EXPOSE 8000

README.md

@@ -1,241 +1,293 @@
-# Gemini Balance - Gemini API 代理和负载均衡器
+[Read this document in Chinese](README_ZH.md)
 
-> ⚠️ 本项目采用 CC BY-NC 4.0（署名-非商业性使用）协议，禁止任何形式的商业倒卖服务，详见 LICENSE 文件。
+# Gemini Balance - Gemini API Proxy and Load Balancer
 
-> 本人从未在各个平台售卖服务，如有遇到售卖此服务者，那一定是倒卖狗，大家切记不要上当受骗。
+<p align="center">
+  <a href="https://trendshift.io/repositories/13692" target="_blank">
+    <img src="https://trendshift.io/api/badge/repositories/13692" alt="snailyp%2Fgemini-balance | Trendshift" style="width: 250px; height: 55px;" width="250" height="55"/>
+  </a>
+</p>
+
+> ⚠️ This project is licensed under the CC BY-NC 4.0 (Attribution-NonCommercial) license. Any form of commercial resale service is prohibited. See the LICENSE file for details.
+
+> I have never sold this service on any platform. If you encounter someone selling this service, they are definitely a reseller. Please be careful not to be deceived.
 
 [![Python](https://img.shields.io/badge/Python-3.9%2B-blue.svg)](https://www.python.org/)
 [![FastAPI](https://img.shields.io/badge/FastAPI-0.100%2B-green.svg)](https://fastapi.tiangolo.com/)
 [![Uvicorn](https://img.shields.io/badge/Uvicorn-running-purple.svg)](https://www.uvicorn.org/)
+[![Telegram Group](https://img.shields.io/badge/Telegram-Group-blue.svg?logo=telegram)](https://t.me/+soaHax5lyI0wZDVl)
 
-## 项目简介
+> Telegram Group: <https://t.me/+soaHax5lyI0wZDVl>
 
-Gemini Balance 是一个基于 Python FastAPI 构建的应用程序，旨在提供 Google Gemini API 的代理和负载均衡功能。它允许您管理多个 Gemini API Key，并通过简单的配置实现 Key 的轮询、认证、模型过滤和状态监控。此外，项目还集成了图像生成和多种图床上传功能，并支持 OpenAI API 格式的代理。
+## Project Introduction
 
-**项目结构:**
+Gemini Balance is an application built with Python FastAPI, designed to provide proxy and load balancing functions for the Google Gemini API. It allows you to manage multiple Gemini API Keys and implement key rotation, authentication, model filtering, and status monitoring through simple configuration. Additionally, the project integrates image generation and multiple image hosting upload functions, and supports proxying in the OpenAI API format.
+
+**Project Structure:**
 
 ```plaintext
 app/
-├── config/       # 配置管理
-├── core/         # 核心应用逻辑 (FastAPI 实例创建, 中间件等)
-├── database/     # 数据库模型和连接
-├── domain/       # 业务领域对象 (可选)
-├── exception/    # 自定义异常
-├── handler/      # 请求处理器 (可选, 或在 router 中处理)
-├── log/          # 日志配置
-├── main.py       # 应用入口
-├── middleware/   # FastAPI 中间件
-├── router/       # API 路由 (Gemini, OpenAI, 状态页等)
-├── scheduler/    # 定时任务 (如 Key 状态检查)
-├── service/      # 业务逻辑服务 (聊天, Key 管理, 统计等)
-├── static/       # 静态文件 (CSS, JS)
-├── templates/    # HTML 模板 (如 Key 状态页)
-├── utils/        # 工具函数
+├── config/       # Configuration management
+├── core/         # Core application logic (FastAPI instance creation, middleware, etc.)
+├── database/     # Database models and connections
+├── domain/       # Business domain objects (optional)
+├── exception/    # Custom exceptions
+├── handler/      # Request handlers (optional, or handled in router)
+├── log/          # Logging configuration
+├── main.py       # Application entry point
+├── middleware/   # FastAPI middleware
+├── router/       # API routes (Gemini, OpenAI, status page, etc.)
+├── scheduler/    # Scheduled tasks (e.g., Key status check)
+├── service/      # Business logic services (chat, Key management, statistics, etc.)
+├── static/       # Static files (CSS, JS)
+├── templates/    # HTML templates (e.g., Key status page)
+├── utils/        # Utility functions
 ```
 
-## ✨ 功能亮点
+## ✨ Feature Highlights
 
-* **多 Key 负载均衡**: 支持配置多个 Gemini API Key (`API_KEYS`)，自动按顺序轮询使用，提高可用性和并发能力。
-* **可视化配置即时生效**: 通过管理后台修改配置后，无需重启服务即可生效，切记要点击保存才会生效。
-![配置面板](files/image4.png)
-* **双协议API 兼容**: 同时支持 Gemini 和 OpenAI 格式的 CHAT API 请求转发。
+* **Multi-Key Load Balancing**: Supports configuring multiple Gemini API Keys (`API_KEYS`) for automatic sequential polling, improving availability and concurrency.
+* **Visual Configuration Takes Effect Immediately**: Configurations modified through the admin backend take effect without restarting the service. Remember to click save for changes to apply.
+    ![Configuration Panel](files/image4.png)
+* **Dual Protocol API Compatibility**: Supports forwarding CHAT API requests in both Gemini and OpenAI formats.
 
-    ```palintext
+    ```plaintext
     openai baseurl `http://localhost:8000(/hf)/v1`
     gemini baseurl `http://localhost:8000(/gemini)/v1beta`
     ```
 
-* **支持图文对话和修改图片**: `IMAGE_MODELS`配置哪个模型可以图文对话和修图的功能，实际调用的时候，用 `配置模型-image`这个模型名对话使用该功能。
-![对话生图](files/image6.png)
-![修改图片](files/image7.png)
-* **支持联网搜索**: 支持联网搜索，`SEARCH_MODELS`配置哪些模型可以联网搜索，实际调用的时候，用 `配置模型-search`这个模型名对话使用该功能
-![联网搜索](files/image8.png)
-* **Key 状态监控**: 提供 `/keys_status` 页面（需要认证），实时查看各 Key 的状态和使用情况。
-![监控面板](files/image.png)
-* **详细的日志记录**: 提供详细的错误日志，方便排查。
-![调用详情](files/image1.png)
-![日志列表](files/image2.png)
-![日志详情](files/image3.png)
-* **支持自定义gemini代理**: 支持自定义gemini代理，比如自行在deno或者cloudflare上搭建gemini代理
-* **openai画图接口兼容**: 将`imagen-3.0-generate-002`模型接口改造成openai画图接口，支持客户端调用。
-* **灵活的添加密钥方式**: 灵活的添加密钥方式，采用正则匹配`gemini_key`,密钥去重
-![添加密钥](files/image5.png)
-* **兼容openai格式embeddings接口**：完美适配openai格式的`embeddings`接口，可用于本地文档向量化。
-* **流式响应优化**: 可选的流式输出优化器 (`STREAM_OPTIMIZER_ENABLED`)，改善长文本流式响应的体验。
-* **失败重试与 Key 管理**: 自动处理 API 请求失败，进行重试 (`MAX_RETRIES`)，并在 Key 失效次数过多时自动禁用 (`MAX_FAILURES`)，定时检查恢复 (`CHECK_INTERVAL_HOURS`)。
-* **Docker 支持**: 支持AMD，ARM架构的docker部署，也可自行构建docker镜像。
-    >镜像地址: docker pull ghcr.io/snailyp/gemini-balance:latest
-* **模型列表自动维护**: 支持openai和gemini模型列表获取，与newapi自动获取模型列表完美兼容，无需手动填写。
-* **支持移除不使用的模型**: 默认提供的模型太多，很多用不上，可以通过`FILTERED_MODELS`过滤掉。
-* **代理支持**: 支持配置 HTTP/SOCKS5 代理服务器 (`PROXIES`)，用于访问 Gemini API，方便在特殊网络环境下使用。支持批量添加代理。
+* **Supports Image-Text Chat and Image Modification**: `IMAGE_MODELS` configures which models can perform image-text chat and image editing. When actually calling, use the `configured_model-image` model name to use this feature.
+    ![Chat with Image Generation](files/image6.png)
+    ![Modify Image](files/image7.png)
+* **Supports Web Search**: Supports web search. `SEARCH_MODELS` configures which models can perform web searches. When actually calling, use the `configured_model-search` model name to use this feature.
+    ![Web Search](files/image8.png)
+* **Key Status Monitoring**: Provides a `/keys_status` page (requires authentication) to view the status and usage of each Key in real-time.
+    ![Monitoring Panel](files/image.png)
+* **Detailed Logging**: Provides detailed error logs for easy troubleshooting.
+    ![Call Details](files/image1.png)
+    ![Log List](files/image2.png)
+    ![Log Details](files/image3.png)
+* **Support for Custom Gemini Proxy**: Supports custom Gemini proxies, such as those built on Deno or Cloudflare.
+* **OpenAI Image Generation API Compatibility**: Adapts the `imagen-3.0-generate-002` model interface to be compatible with the OpenAI image generation API, supporting client calls.
+* **Flexible Key Addition**: Flexible way to add keys using regex matching for `gemini_key`, with key deduplication.
+    ![Add Key](files/image5.png)
+* **OpenAI Format Embeddings API Compatibility**: Perfectly adapts to the OpenAI format `embeddings` interface, usable for local document vectorization.
+* **Streamlined Response Optimization**: Optional stream output optimizer (`STREAM_OPTIMIZER_ENABLED`) to improve the experience of long-text stream responses.
+* **Failure Retry and Key Management**: Automatically handles API request failures, retries (`MAX_RETRIES`), automatically disables Keys after too many failures (`MAX_FAILURES`), and periodically checks for recovery (`CHECK_INTERVAL_HOURS`).
+* **Docker Support**: Supports AMD and ARM architecture Docker deployments. You can also build your own Docker image.
+    > Image address: docker pull ghcr.io/snailyp/gemini-balance:latest
+* **Automatic Model List Maintenance**: Supports fetching OpenAI and Gemini model lists, perfectly compatible with NewAPI's automatic model list fetching, no manual entry required.
+* **Support for Removing Unused Models**: Too many default models are provided, many of which are not used. You can filter them out using `FILTERED_MODELS`.
+* **Proxy Support**: Supports configuring HTTP/SOCKS5 proxy servers (`PROXIES`) for accessing the Gemini API, convenient for use in special network environments. Supports batch adding proxies.
 
-## 🚀 快速开始
+## 🚀 Quick Start
 
-### 自行构建 Docker (推荐)
+### Build Docker Yourself (Recommended)
 
-#### a) dockerfile构建
+#### a) Build with Dockerfile
 
-1. **构建镜像**:
+1. **Build Image**:
 
     ```bash
     docker build -t gemini-balance .
     ```
 
-2. **运行容器**:
+2. **Run Container**:
 
     ```bash
     docker run -d -p 8000:8000 --env-file .env gemini-balance
     ```
 
-    * `-d`: 后台运行。
-    * `-p 8000:8000`: 将容器的 8000 端口映射到主机的 8000 端口。
-    * `--env-file .env`: 使用 `.env` 文件设置环境变量。
+    * `-d`: Run in detached mode.
+    * `-p 8000:8000`: Map port 8000 of the container to port 8000 of the host.
+    * `--env-file .env`: Use the `.env` file to set environment variables.
 
-#### b) 用现有的docker镜像部署
+    > Note: If using an SQLite database, you need to mount a data volume to persist
+>
+    > ```bash
+    > docker run -d -p 8000:8000 --env-file .env -v /path/to/data:/app/data gemini-balance
+    > ```
+>
+    > Where `/path/to/data` is the data storage path on the host, and `/app/data` is the data directory inside the container.
 
-1. **拉取镜像**:
+#### b) Deploy with an Existing Docker Image
 
-   ```bash
-   docker pull ghcr.io/snailyp/gemini-balance:latest
-   ```
+1. **Pull Image**:
 
-2. **运行容器**:
+    ```bash
+    docker pull ghcr.io/snailyp/gemini-balance:latest
+    ```
 
-   ```bash
-   docker run -d -p 8000:8000 --env-file .env ghcr.io/snailyp/gemini-balance:latest
-   ```
+2. **Run Container**:
 
-   * `-d`: 后台运行。
-   * `-p 8000:8000`: 将容器的 8000 端口映射到主机的 8000 端口 (根据需要调整)。
-   * `--env-file .env`: 使用 `.env` 文件设置环境变量 (确保 `.env` 文件存在于执行命令的目录)。
+    ```bash
+    docker run -d -p 8000:8000 --env-file .env ghcr.io/snailyp/gemini-balance:latest
+    ```
 
-### 本地运行 (适用于开发和测试)
+    * `-d`: Run in detached mode.
+    * `-p 8000:8000`: Map port 8000 of the container to port 8000 of the host (adjust as needed).
+    * `--env-file .env`: Use the `.env` file to set environment variables (ensure the `.env` file exists in the directory where the command is executed).
 
-如果您想在本地直接运行源代码进行开发或测试，请按照以下步骤操作：
+    > Note: If using an SQLite database, you need to mount a data volume to persist
+>
+    > ```bash
+    > docker run -d -p 8000:8000 --env-file .env -v /path/to/data:/app/data ghcr.io/snailyp/gemini-balance:latest
+    > ```
+>
+    > Where `/path/to/data` is the data storage path on the host, and `/app/data` is the data directory inside the container.
 
-1. **确保已完成准备工作**:
-    * 克隆仓库到本地。
-    * 安装 Python 3.9 或更高版本。
-    * 在项目根目录下创建并配置好 `.env` 文件 (参考前面的“配置环境变量”部分)。
-    * 安装项目依赖：
+### Run Locally (Suitable for Development and Testing)
+
+If you want to run the source code directly locally for development or testing, follow these steps:
+
+1. **Ensure Prerequisites are Met**:
+    * Clone the repository locally.
+    * Install Python 3.9 or higher.
+    * Create and configure the `.env` file in the project root directory (refer to the "Configure Environment Variables" section above).
+    * Install project dependencies:
 
         ```bash
         pip install -r requirements.txt
         ```
 
-2. **启动应用**:
-    在项目根目录下运行以下命令：
+2. **Start Application**:
+    Run the following command in the project root directory:
 
     ```bash
     uvicorn app.main:app --host 0.0.0.0 --port 8000 --reload
     ```
 
-    * `app.main:app`: 指定 FastAPI 应用实例的位置 (`app` 模块中的 `main.py` 文件里的 `app` 对象)。
-    * `--host 0.0.0.0`: 使应用可以从本地网络中的任何 IP 地址访问。
-    * `--port 8000`: 指定应用监听的端口号 (您可以根据需要修改)。
-    * `--reload`: 启用自动重载功能。当您修改代码时，服务会自动重启，非常适合开发环境 (生产环境请移除此选项)。
+    * `app.main:app`: Specifies the location of the FastAPI application instance (the `app` object in the `main.py` file within the `app` module).
+    * `--host 0.0.0.0`: Makes the application accessible from any IP address on the local network.
+    * `--port 8000`: Specifies the port number the application listens on (you can change this as needed).
+    * `--reload`: Enables automatic reloading. When you modify the code, the service will automatically restart, which is very suitable for development environments (remove this option in production environments).
 
-3. **访问应用**:
-    应用启动后，您可以通过浏览器或 API 工具访问 `http://localhost:8000` (或您指定的主机和端口)。
+3. **Access Application**:
+    After the application starts, you can access `http://localhost:8000` (or the host and port you specified) through a browser or API tool.
 
-### 完整配置项列表
+### Complete Configuration List
 
-| 配置项                       | 说明                                                     | 默认值                             |
-| :--------------------------- | :------------------------------------------------------- | :---------------------------------------------------- |
-| **数据库配置**               |                                                          |                                                       |
-| `MYSQL_HOST`                 | 必填，MySQL 数据库主机地址                               | `localhost`                                           |
-| `MYSQL_PORT`                 | 必填，MySQL 数据库端口                                   | `3306`                                                |
-| `MYSQL_USER`                 | 必填，MySQL 数据库用户名                                 | `your_db_user`                                        |
-| `MYSQL_PASSWORD`             | 必填，MySQL 数据库密码                                   | `your_db_password`                                    |
-| `MYSQL_DATABASE`             | 必填，MySQL 数据库名称                                   | `defaultdb`                                   |
-| **API 相关配置**             |                                                          |                                                       |
-| `API_KEYS`                   | 必填，Gemini API 密钥列表，用于负载均衡                        | `["your-gemini-api-key-1", "your-gemini-api-key-2"]`  |
-| `ALLOWED_TOKENS`             | 必填，允许访问的 Token 列表                                    | `["your-access-token-1", "your-access-token-2"]`      |
-| `AUTH_TOKEN`                 | 可选，超级管理员token，具有所有权限，不填默认使用 ALLOWED_TOKENS 的第一个 | `""`                                                  |
-| `TEST_MODEL`                 | 可选，用于测试密钥是否可用的模型名                             | `gemini-1.5-flash`                                    |
-| `IMAGE_MODELS`               | 可选，支持绘图功能的模型列表                                   | `["gemini-2.0-flash-exp"]`                            |
-| `SEARCH_MODELS`              | 可选，支持搜索功能的模型列表                                   | `["gemini-2.0-flash-exp"]`                            |
-| `FILTERED_MODELS`            | 可选，被禁用的模型列表                                         | `["gemini-1.0-pro-vision-latest", ...]`               |
-| `TOOLS_CODE_EXECUTION_ENABLED` | 可选，是否启用代码执行工具                                     | `false`                                               |
-| `SHOW_SEARCH_LINK`           | 可选，是否在响应中显示搜索结果链接                             | `true`                                                |
-| `SHOW_THINKING_PROCESS`      | 可选，是否显示模型思考过程                                     | `true`                                                |
-| `THINKING_MODELS`            | 可选，支持思考功能的模型列表                                   | `[]`                                                  |
-| `THINKING_BUDGET_MAP`        | 可选，思考功能预算映射 (模型名:预算值)                         | `{}`                                                  |
-| `BASE_URL`                   | 可选，Gemini API 基础 URL，默认无需修改                        | `https://generativelanguage.googleapis.com/v1beta`    |
-| `MAX_FAILURES`               | 可选，允许单个key失败的次数                                    | `3`                                                   |
-| `MAX_RETRIES`                | 可选，API 请求失败时的最大重试次数                             | `3`                                                   |
-| `CHECK_INTERVAL_HOURS`       | 可选，检查禁用 Key 是否恢复的时间间隔 (小时)                   | `1`                                                   |
-| `TIMEZONE`                   | 可选，应用程序使用的时区                                       | `Asia/Shanghai`                                       |
-| `TIME_OUT`                   | 可选，请求超时时间 (秒)                                        | `300`                                                 |
-| `PROXIES`                    | 可选，代理服务器列表 (例如 `http://user:pass@host:port`, `socks5://host:port`) | `[]`                                                  |
-| `LOG_LEVEL`                  | 可选，日志级别，例如 DEBUG, INFO, WARNING, ERROR, CRITICAL     | `INFO`                                                |
-| **图像生成相关**             |                                                          |                                                       |
-| `PAID_KEY`                   | 可选，付费版API Key，用于图片生成等高级功能                    | `your-paid-api-key`                                   |
-| `CREATE_IMAGE_MODEL`         | 可选，图片生成模型                                             | `imagen-3.0-generate-002`                             |
-| `UPLOAD_PROVIDER`            | 可选，图片上传提供商: `smms`, `picgo`, `cloudflare_imgbed`     | `smms`                                                |
-| `SMMS_SECRET_TOKEN`          | 可选，SM.MS图床的API Token                                     | `your-smms-token`                                     |
-| `PICGO_API_KEY`              | 可选，[PicoGo](https://www.picgo.net/)图床的API Key                                      | `your-picogo-apikey`                                  |
-| `CLOUDFLARE_IMGBED_URL`      | 可选，[CloudFlare](https://github.com/MarSeventh/CloudFlare-ImgBed) 图床上传地址                                  | `https://xxxxxxx.pages.dev/upload`                    |
-| `CLOUDFLARE_IMGBED_AUTH_CODE`| 可选，CloudFlare图床的鉴权key                                  | `your-cloudflare-imgber-auth-code`                    |
-| **流式优化器相关**           |                                                          |                                                       |
-| `STREAM_OPTIMIZER_ENABLED`   | 可选，是否启用流式输出优化                                     | `false`                                               |
-| `STREAM_MIN_DELAY`           | 可选，流式输出最小延迟                                         | `0.016`                                               |
-| `STREAM_MAX_DELAY`           | 可选，流式输出最大延迟                                         | `0.024`                                               |
-| `STREAM_SHORT_TEXT_THRESHOLD`| 可选，短文本阈值                                               | `10`                                                  |
-| `STREAM_LONG_TEXT_THRESHOLD` | 可选，长文本阈值                                               | `50`                                                  |
-| `STREAM_CHUNK_SIZE`          | 可选，流式输出块大小                                           | `5`                                                   |
+| Configuration Item             | Description                                                                 | Default Value                                                                                                                                                                                                                            |
+| :----------------------------- | :-------------------------------------------------------------------------- | :--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
+| **Database Configuration**     |                                                                             |                                                                                                                                                                                                                                          |
+| `DATABASE_TYPE`                | Optional, database type, supports `mysql` or `sqlite`                       | `mysql`                                                                                                                                                                                                                                  |
+| `SQLITE_DATABASE`              | Optional, required when using `sqlite`, SQLite database file path           | `default_db`                                                                                                                                                                                                                             |
+| `MYSQL_HOST`                   | Required when using `mysql`, MySQL database host address                    | `localhost`                                                                                                                                                                                                                              |
+| `MYSQL_SOCKET`                 | Optional, MySQL database socket address                                     | `/var/run/mysqld/mysqld.sock`                                                                                                                                                                                                            |
+| `MYSQL_PORT`                   | Required when using `mysql`, MySQL database port                            | `3306`                                                                                                                                                                                                                                   |
+| `MYSQL_USER`                   | Required when using `mysql`, MySQL database username                        | `your_db_user`                                                                                                                                                                                                                           |
+| `MYSQL_PASSWORD`               | Required when using `mysql`, MySQL database password                        | `your_db_password`                                                                                                                                                                                                                       |
+| `MYSQL_DATABASE`               | Required when using `mysql`, MySQL database name                            | `defaultdb`                                                                                                                                                                                                                              |
+| **API Related Configuration**  |                                                                             |                                                                                                                                                                                                                                          |
+| `API_KEYS`                     | Required, list of Gemini API keys for load balancing                        | `["your-gemini-api-key-1", "your-gemini-api-key-2"]`                                                                                                                                                                                     |
+| `ALLOWED_TOKENS`               | Required, list of tokens allowed to access                                  | `["your-access-token-1", "your-access-token-2"]`                                                                                                                                                                                         |
+| `AUTH_TOKEN`                   | Optional, super admin token with all permissions, defaults to the first of `ALLOWED_TOKENS` if not set | `sk-123456`                                                                                                                                                                                                              |
+| `TEST_MODEL`                   | Optional, model name used to test if a key is usable                        | `gemini-1.5-flash`                                                                                                                                                                                                                       |
+| `IMAGE_MODELS`                 | Optional, list of models that support drawing functions                     | `["gemini-2.0-flash-exp"]`                                                                                                                                                                                                               |
+| `SEARCH_MODELS`                | Optional, list of models that support search functions                      | `["gemini-2.0-flash-exp"]`                                                                                                                                                                                                               |
+| `FILTERED_MODELS`              | Optional, list of disabled models                                           | `["gemini-1.0-pro-vision-latest", ...]`                                                                                                                                                                                                  |
+| `TOOLS_CODE_EXECUTION_ENABLED` | Optional, whether to enable the code execution tool                         | `false`                                                                                                                                                                                                                                  |
+| `SHOW_SEARCH_LINK`             | Optional, whether to display search result links in the response            | `true`                                                                                                                                                                                                                                   |
+| `SHOW_THINKING_PROCESS`        | Optional, whether to display the model's thinking process                   | `true`                                                                                                                                                                                                                                   |
+| `THINKING_MODELS`              | Optional, list of models that support thinking functions                    | `[]`                                                                                                                                                                                                                                     |
+| `THINKING_BUDGET_MAP`          | Optional, thinking function budget mapping (model_name:budget_value)        | `{}`                                                                                                                                                                                                                                     |
+| `URL_NORMALIZATION_ENABLED`    | Optional, whether to enable intelligent URL routing mapping                 | `false`                                                                                                                                                                                                                                  |
+| `BASE_URL`                     | Optional, Gemini API base URL, no modification needed by default            | `https://generativelanguage.googleapis.com/v1beta`                                                                                                                                                                                       |
+| `MAX_FAILURES`                 | Optional, number of times a single key is allowed to fail                   | `3`                                                                                                                                                                                                                                      |
+| `MAX_RETRIES`                  | Optional, maximum number of retries for failed API requests                 | `3`                                                                                                                                                                                                                                      |
+| `CHECK_INTERVAL_HOURS`         | Optional, time interval (hours) to check if a disabled Key has recovered    | `1`                                                                                                                                                                                                                                      |
+| `TIMEZONE`                     | Optional, timezone used by the application                                  | `Asia/Shanghai`                                                                                                                                                                                                                          |
+| `TIME_OUT`                     | Optional, request timeout (seconds)                                         | `300`                                                                                                                                                                                                                                    |
+| `PROXIES`                      | Optional, list of proxy servers (e.g., `http://user:pass@host:port`, `socks5://host:port`) | `[]`                                                                                                                                                                                                                                     |
+| `LOG_LEVEL`                    | Optional, log level, e.g., DEBUG, INFO, WARNING, ERROR, CRITICAL          | `INFO`                                                                                                                                                                                                                                   |
+| `AUTO_DELETE_ERROR_LOGS_ENABLED` | Optional, whether to enable automatic deletion of error logs              | `true`                                                                                                                                                                                                                                   |
+| `AUTO_DELETE_ERROR_LOGS_DAYS`  | Optional, automatically delete error logs older than this many days (e.g., 1, 7, 30) | `7`                                                                                                                                                                                                                                      |
+| `AUTO_DELETE_REQUEST_LOGS_ENABLED`| Optional, whether to enable automatic deletion of request logs             | `false`                                                                                                                                                                                                                                  |
+| `AUTO_DELETE_REQUEST_LOGS_DAYS` | Optional, automatically delete request logs older than this many days (e.g., 1, 7, 30) | `30`                                                                                                                                                                                                                                     |
+| `SAFETY_SETTINGS`              | Optional, safety settings (JSON string format), used to configure content safety thresholds. Example values may need adjustment based on actual model support. | `[{"category": "HARM_CATEGORY_HARASSMENT", "threshold": "OFF"}, {"category": "HARM_CATEGORY_HATE_SPEECH", "threshold": "OFF"}, {"category": "HARM_CATEGORY_SEXUALLY_EXPLICIT", "threshold": "OFF"}, {"category": "HARM_CATEGORY_DANGEROUS_CONTENT", "threshold": "OFF"}, {"category": "HARM_CATEGORY_CIVIC_INTEGRITY", "threshold": "BLOCK_NONE"}]` |
+| **TTS Related**                |                                                                             |                                                                                                                                                                                                                                          |
+| `TTS_MODEL`                    | Optional, TTS model name                                                    | `gemini-2.5-flash-preview-tts`                                                                                                                                                                                                           |
+| `TTS_VOICE_NAME`               | Optional, TTS voice name                                                    | `Zephyr`                                                                                                                                                                                                                                 |
+| `TTS_SPEED`                    | Optional, TTS speed                                                         | `normal`                                                                                                                                                                                                                                 |
+| **Image Generation Related**   |                                                                             |                                                                                                                                                                                                                                          |
+| `PAID_KEY`                     | Optional, paid API Key for advanced features like image generation          | `your-paid-api-key`                                                                                                                                                                                                                        |
+| `CREATE_IMAGE_MODEL`           | Optional, image generation model                                            | `imagen-3.0-generate-002`                                                                                                                                                                                                                  |
+| `UPLOAD_PROVIDER`              | Optional, image upload provider: `smms`, `picgo`, `cloudflare_imgbed`       | `smms`                                                                                                                                                                                                                                   |
+| `SMMS_SECRET_TOKEN`            | Optional, API Token for SM.MS image hosting                                 | `your-smms-token`                                                                                                                                                                                                                          |
+| `PICGO_API_KEY`                | Optional, API Key for [PicoGo](https://www.picgo.net/) image hosting        | `your-picogo-apikey`                                                                                                                                                                                                                         |
+| `CLOUDFLARE_IMGBED_URL`        | Optional, [CloudFlare](https://github.com/MarSeventh/CloudFlare-ImgBed) image hosting upload address | `https://xxxxxxx.pages.dev/upload`                                                                                                                                                                                           |
+| `CLOUDFLARE_IMGBED_AUTH_CODE`  | Optional, authentication key for CloudFlare image hosting                   | `your-cloudflare-imgber-auth-code`                                                                                                                                                                                                           |
+| `CLOUDFLARE_IMGBED_UPLOAD_FOLDER` | Optional, upload folder path for CloudFlare image hosting               | `""`                                                                                                                                                                                                                                          |
+| **Stream Optimizer Related**   |                                                                             |                                                                                                                                                                                                                                          |
+| `STREAM_OPTIMIZER_ENABLED`     | Optional, whether to enable stream output optimization                      | `false`                                                                                                                                                                                                                                  |
+| `STREAM_MIN_DELAY`             | Optional, minimum delay for stream output                                   | `0.016`                                                                                                                                                                                                                                  |
+| `STREAM_MAX_DELAY`             | Optional, maximum delay for stream output                                   | `0.024`                                                                                                                                                                                                                                  |
+| `STREAM_SHORT_TEXT_THRESHOLD`  | Optional, short text threshold                                              | `10`                                                                                                                                                                                                                                     |
+| `STREAM_LONG_TEXT_THRESHOLD`   | Optional, long text threshold                                               | `50`                                                                                                                                                                                                                                     |
+| `STREAM_CHUNK_SIZE`            | Optional, stream output chunk size                                          | `5`                                                                                                                                                                                                                                      |
+| **Fake Stream Related**        |                                                                             |                                                                                                                                                                                                                                          |
+| `FAKE_STREAM_ENABLED`          | Optional, whether to enable fake streaming for models or scenarios that don't support streaming | `false`                                                                                                                                                                                                                                  |
+| `FAKE_STREAM_EMPTY_DATA_INTERVAL_SECONDS` | Optional, interval in seconds for sending heartbeat empty data during fake streaming | `5`                                                                                                                                                                                                                                      |
 
-## ⚙️ API 端点
+## ⚙️ API Endpoints
 
-以下是服务提供的主要 API 端点：
+The following are the main API endpoints provided by the service:
 
-### Gemini API 相关 (`(/gemini)/v1beta`)
+### Gemini API Related (`(/gemini)/v1beta`)
 
-* `GET /models`: 列出可用的 Gemini 模型。
-* `POST /models/{model_name}:generateContent`: 使用指定的 Gemini 模型生成内容。
-* `POST /models/{model_name}:streamGenerateContent`: 使用指定的 Gemini 模型流式生成内容。
+* `GET /models`: List available Gemini models.
+* `POST /models/{model_name}:generateContent`: Generate content using the specified Gemini model.
+* `POST /models/{model_name}:streamGenerateContent`: Stream content generation using the specified Gemini model.
 
-### OpenAI API 相关
+### OpenAI API Related
 
-* `GET (/hf)/v1/models`: 列出可用的模型 (底层用的gemini格式)。
-* `POST (/hf)/v1/chat/completions`: 进行聊天补全 (底层用的gemini格式, 支持流式传输)。
-* `POST (/hf)/v1/embeddings`: 创建文本嵌入 (底层用的gemini格式)。
-* `POST (/hf)/v1/images/generations`: 生成图像 (底层用的gemini格式)。
-* `GET /openai/v1/models`: 列出可用的模型 (底层用的openai格式)。
-* `POST /openai/v1/chat/completions`: 进行聊天补全 (底层用的openai格式, 支持流式传输, 可防止截断，速度也快)。
-* `POST /openai/v1/embeddings`: 创建文本嵌入 (底层用的openai格式)。
-* `POST /openai/v1/images/generations`: 生成图像 (底层用的openai格式)。
+* `GET (/hf)/v1/models`: List available models (uses Gemini format underneath).
+* `POST (/hf)/v1/chat/completions`: Perform chat completion (uses Gemini format underneath, supports streaming).
+* `POST (/hf)/v1/embeddings`: Create text embeddings (uses Gemini format underneath).
+* `POST (/hf)/v1/images/generations`: Generate images (uses Gemini format underneath).
+* `GET /openai/v1/models`: List available models (uses OpenAI format underneath).
+* `POST /openai/v1/chat/completions`: Perform chat completion (uses OpenAI format underneath, supports streaming, can prevent truncation, and is faster).
+* `POST /openai/v1/embeddings`: Create text embeddings (uses OpenAI format underneath).
+* `POST /openai/v1/images/generations`: Generate images (uses OpenAI format underneath).
 
-## 🤝 贡献
+## 🤝 Contributing
 
-欢迎提交 Pull Request 或 Issue。
+Pull Requests or Issues are welcome.
 
-## 🎉 特别鸣谢
+## 🎉 Special Thanks
 
-特别鸣谢以下项目和平台为本项目提供图床服务:
+Special thanks to the following projects and platforms for providing image hosting services for this project:
 
 * [PicGo](https://www.picgo.net/)
 * [SM.MS](https://smms.app/)
-* [CloudFlare-ImgBed](https://github.com/MarSeventh/CloudFlare-ImgBed) 开源项目
+* [CloudFlare-ImgBed](https://github.com/MarSeventh/CloudFlare-ImgBed) open source project
 
-## 🙏 感谢贡献者
+## 🙏 Thanks to Contributors
 
-感谢所有为本项目做出贡献的开发者！
+Thanks to all developers who contributed to this project!
 
 [![Contributors](https://contrib.rocks/image?repo=snailyp/gemini-balance)](https://github.com/snailyp/gemini-balance/graphs/contributors)
 
+## Thanks to Our Supporters
+
+A special shout-out to DigitalOcean for providing the rock-solid and dependable cloud infrastructure that keeps this project humming!
+[![DigitalOcean Logo](files/dataocean.svg)](https://m.do.co/c/b249dd7f3b4c)
+
+CDN acceleration and security protection for this project are sponsored by Tencent EdgeOne.
+[![EdgeOne Logo](https://edgeone.ai/media/34fe3a45-492d-4ea4-ae5d-ea1087ca7b4b.png)](https://edgeone.ai/?from=github)
+
 ## ⭐ Star History
 
 [![Star History Chart](https://api.star-history.com/svg?repos=snailyp/gemini-balance&type=Date)](https://star-history.com/#snailyp/gemini-balance&Date)
 
-## 💖 友情项目
+## 💖 Friendly Projects
 
-* **[OneLine](https://github.com/chengtx809/OneLine)** by [chengtx809](https://github.com/chengtx809) - OneLine一线：AI驱动的热点事件时间轴生成工具
+* **[OneLine](https://github.com/chengtx809/OneLine)** by [chengtx809](https://github.com/chengtx809) - OneLine: AI-driven hot event timeline generation tool
 
-## 🎁 项目支持
+## 🎁 Project Support
 
-如果你觉得这个项目对你有帮助，可以考虑通过 [爱发电](https://afdian.com/a/snaily) 支持我。
+If you find this project helpful, consider supporting me via [Afdian](https://afdian.com/a/snaily).
 
-## 许可证
+## License
 
-本项目采用 CC BY-NC 4.0（署名-非商业性使用）协议，禁止任何形式的商业倒卖服务，详见 LICENSE 文件。
+This project is licensed under the CC BY-NC 4.0 (Attribution-NonCommercial) license. Any form of commercial resale service is prohibited. See the LICENSE file for details.

README_ZH.md

@@ -0,0 +1,278 @@
+# Gemini Balance - Gemini API 代理和负载均衡器
+
+<p align="center">
+  <a href="https://trendshift.io/repositories/13692" target="_blank">
+    <img src="https://trendshift.io/api/badge/repositories/13692" alt="snailyp%2Fgemini-balance | Trendshift" style="width: 250px; height: 55px;" width="250" height="55"/>
+  </a>
+</p>
+
+> ⚠️ 本项目采用 CC BY-NC 4.0（署名-非商业性使用）协议，禁止任何形式的商业倒卖服务，详见 LICENSE 文件。
+
+> 本人从未在各个平台售卖服务，如有遇到售卖此服务者，那一定是倒卖狗，大家切记不要上当受骗。
+
+[![Python](https://img.shields.io/badge/Python-3.9%2B-blue.svg)](https://www.python.org/)
+[![FastAPI](https://img.shields.io/badge/FastAPI-0.100%2B-green.svg)](https://fastapi.tiangolo.com/)
+[![Uvicorn](https://img.shields.io/badge/Uvicorn-running-purple.svg)](https://www.uvicorn.org/)
+[![Telegram Group](https://img.shields.io/badge/Telegram-Group-blue.svg?logo=telegram)](https://t.me/+soaHax5lyI0wZDVl)
+> 交流群：https://t.me/+soaHax5lyI0wZDVl
+
+## 项目简介
+
+Gemini Balance 是一个基于 Python FastAPI 构建的应用程序，旨在提供 Google Gemini API 的代理和负载均衡功能。它允许您管理多个 Gemini API Key，并通过简单的配置实现 Key 的轮询、认证、模型过滤和状态监控。此外，项目还集成了图像生成和多种图床上传功能，并支持 OpenAI API 格式的代理。
+
+**项目结构:**
+
+```plaintext
+app/
+├── config/       # 配置管理
+├── core/         # 核心应用逻辑 (FastAPI 实例创建, 中间件等)
+├── database/     # 数据库模型和连接
+├── domain/       # 业务领域对象 (可选)
+├── exception/    # 自定义异常
+├── handler/      # 请求处理器 (可选, 或在 router 中处理)
+├── log/          # 日志配置
+├── main.py       # 应用入口
+├── middleware/   # FastAPI 中间件
+├── router/       # API 路由 (Gemini, OpenAI, 状态页等)
+├── scheduler/    # 定时任务 (如 Key 状态检查)
+├── service/      # 业务逻辑服务 (聊天, Key 管理, 统计等)
+├── static/       # 静态文件 (CSS, JS)
+├── templates/    # HTML 模板 (如 Key 状态页)
+├── utils/        # 工具函数
+```
+
+## ✨ 功能亮点
+
+* **多 Key 负载均衡**: 支持配置多个 Gemini API Key (`API_KEYS`)，自动按顺序轮询使用，提高可用性和并发能力。
+* **可视化配置即时生效**: 通过管理后台修改配置后，无需重启服务即可生效，切记要点击保存才会生效。
+![配置面板](files/image4.png)
+* **双协议API 兼容**: 同时支持 Gemini 和 OpenAI 格式的 CHAT API 请求转发。
+
+    ```palintext
+    openai baseurl `http://localhost:8000(/hf)/v1`
+    gemini baseurl `http://localhost:8000(/gemini)/v1beta`
+    ```
+
+* **支持图文对话和修改图片**: `IMAGE_MODELS`配置哪个模型可以图文对话和修图的功能，实际调用的时候，用 `配置模型-image`这个模型名对话使用该功能。
+![对话生图](files/image6.png)
+![修改图片](files/image7.png)
+* **支持联网搜索**: 支持联网搜索，`SEARCH_MODELS`配置哪些模型可以联网搜索，实际调用的时候，用 `配置模型-search`这个模型名对话使用该功能
+![联网搜索](files/image8.png)
+* **Key 状态监控**: 提供 `/keys_status` 页面（需要认证），实时查看各 Key 的状态和使用情况。
+![监控面板](files/image.png)
+* **详细的日志记录**: 提供详细的错误日志，方便排查。
+![调用详情](files/image1.png)
+![日志列表](files/image2.png)
+![日志详情](files/image3.png)
+* **支持自定义gemini代理**: 支持自定义gemini代理，比如自行在deno或者cloudflare上搭建gemini代理
+* **openai画图接口兼容**: 将`imagen-3.0-generate-002`模型接口改造成openai画图接口，支持客户端调用。
+* **灵活的添加密钥方式**: 灵活的添加密钥方式，采用正则匹配`gemini_key`,密钥去重
+![添加密钥](files/image5.png)
+* **兼容openai格式embeddings接口**：完美适配openai格式的`embeddings`接口，可用于本地文档向量化。
+* **流式响应优化**: 可选的流式输出优化器 (`STREAM_OPTIMIZER_ENABLED`)，改善长文本流式响应的体验。
+* **失败重试与 Key 管理**: 自动处理 API 请求失败，进行重试 (`MAX_RETRIES`)，并在 Key 失效次数过多时自动禁用 (`MAX_FAILURES`)，定时检查恢复 (`CHECK_INTERVAL_HOURS`)。
+* **Docker 支持**: 支持AMD，ARM架构的docker部署，也可自行构建docker镜像。
+    >镜像地址: docker pull ghcr.io/snailyp/gemini-balance:latest
+* **模型列表自动维护**: 支持openai和gemini模型列表获取，与newapi自动获取模型列表完美兼容，无需手动填写。
+* **支持移除不使用的模型**: 默认提供的模型太多，很多用不上，可以通过`FILTERED_MODELS`过滤掉。
+* **代理支持**: 支持配置 HTTP/SOCKS5 代理服务器 (`PROXIES`)，用于访问 Gemini API，方便在特殊网络环境下使用。支持批量添加代理。
+
+## 🚀 快速开始
+
+### 自行构建 Docker (推荐)
+
+#### a) dockerfile构建
+
+1. **构建镜像**:
+
+    ```bash
+    docker build -t gemini-balance .
+    ```
+
+2. **运行容器**:
+
+    ```bash
+    docker run -d -p 8000:8000 --env-file .env gemini-balance
+    ```
+
+    * `-d`: 后台运行。
+    * `-p 8000:8000`: 将容器的 8000 端口映射到主机的 8000 端口。
+    * `--env-file .env`: 使用 `.env` 文件设置环境变量。
+
+    > 注意：如果使用 SQLite 数据库，需要挂载数据卷以持久化数据：
+    > ```bash
+    > docker run -d -p 8000:8000 --env-file .env -v /path/to/data:/app/data gemini-balance
+    > ```
+    > 其中 `/path/to/data` 是主机上的数据存储路径，`/app/data` 是容器内的数据目录。
+
+#### b) 用现有的docker镜像部署
+
+1. **拉取镜像**:
+
+   ```bash
+   docker pull ghcr.io/snailyp/gemini-balance:latest
+   ```
+
+2. **运行容器**:
+
+   ```bash
+   docker run -d -p 8000:8000 --env-file .env ghcr.io/snailyp/gemini-balance:latest
+   ```
+
+   * `-d`: 后台运行。
+   * `-p 8000:8000`: 将容器的 8000 端口映射到主机的 8000 端口 (根据需要调整)。
+   * `--env-file .env`: 使用 `.env` 文件设置环境变量 (确保 `.env` 文件存在于执行命令的目录)。
+
+    > 注意：如果使用 SQLite 数据库，需要挂载数据卷以持久化数据：
+    > ```bash
+    > docker run -d -p 8000:8000 --env-file .env -v /path/to/data:/app/data ghcr.io/snailyp/gemini-balance:latest
+    > ```
+    > 其中 `/path/to/data` 是主机上的数据存储路径，`/app/data` 是容器内的数据目录。
+
+### 本地运行 (适用于开发和测试)
+
+如果您想在本地直接运行源代码进行开发或测试，请按照以下步骤操作：
+
+1. **确保已完成准备工作**:
+    * 克隆仓库到本地。
+    * 安装 Python 3.9 或更高版本。
+    * 在项目根目录下创建并配置好 `.env` 文件 (参考前面的"配置环境变量"部分)。
+    * 安装项目依赖：
+
+        ```bash
+        pip install -r requirements.txt
+        ```
+
+2. **启动应用**:
+    在项目根目录下运行以下命令：
+
+    ```bash
+    uvicorn app.main:app --host 0.0.0.0 --port 8000 --reload
+    ```
+
+    * `app.main:app`: 指定 FastAPI 应用实例的位置 (`app` 模块中的 `main.py` 文件里的 `app` 对象)。
+    * `--host 0.0.0.0`: 使应用可以从本地网络中的任何 IP 地址访问。
+    * `--port 8000`: 指定应用监听的端口号 (您可以根据需要修改)。
+    * `--reload`: 启用自动重载功能。当您修改代码时，服务会自动重启，非常适合开发环境 (生产环境请移除此选项)。
+
+3. **访问应用**:
+    应用启动后，您可以通过浏览器或 API 工具访问 `http://localhost:8000` (或您指定的主机和端口)。
+
+### 完整配置项列表
+
+| 配置项                       | 说明                                                     | 默认值                             |
+| :--------------------------- | :------------------------------------------------------- | :---------------------------------------------------- |
+| **数据库配置**               |                                                          |                                                       |
+| `DATABASE_TYPE`              | 可选，数据库类型，支持 `mysql` 或 `sqlite`               | `mysql`                                              |
+| `SQLITE_DATABASE`            | 可选，当使用 `sqlite` 时必填，SQLite 数据库文件路径 | `default_db`                                         |
+| `MYSQL_HOST`                 | 当使用 `mysql` 时必填，MySQL 数据库主机地址    | `localhost`                                          |
+| `MYSQL_SOCKET`               | 可选，MySQL 数据库 socket 地址                          | `/var/run/mysqld/mysqld.sock`                        |
+| `MYSQL_PORT`                 | 当使用 `mysql` 时必填，MySQL 数据库端口        | `3306`                                               |
+| `MYSQL_USER`                 | 当使用 `mysql` 时必填，MySQL 数据库用户名      | `your_db_user`                                       |
+| `MYSQL_PASSWORD`             | 当使用 `mysql` 时必填，MySQL 数据库密码        | `your_db_password`                                   |
+| `MYSQL_DATABASE`             | 当使用 `mysql` 时必填，MySQL 数据库名称        | `defaultdb`                                          |
+| **API 相关配置**             |                                                          |                                                       |
+| `API_KEYS`                   | 必填，Gemini API 密钥列表，用于负载均衡                        | `["your-gemini-api-key-1", "your-gemini-api-key-2"]`  |
+| `ALLOWED_TOKENS`             | 必填，允许访问的 Token 列表                                    | `["your-access-token-1", "your-access-token-2"]`      |
+| `AUTH_TOKEN`                 | 可选，超级管理员token，具有所有权限，不填默认使用 ALLOWED_TOKENS 的第一个 | `sk-123456`                                                  |
+| `TEST_MODEL`                 | 可选，用于测试密钥是否可用的模型名                             | `gemini-1.5-flash`                                    |
+| `IMAGE_MODELS`               | 可选，支持绘图功能的模型列表                                   | `["gemini-2.0-flash-exp"]`                            |
+| `SEARCH_MODELS`              | 可选，支持搜索功能的模型列表                                   | `["gemini-2.0-flash-exp"]`                            |
+| `FILTERED_MODELS`            | 可选，被禁用的模型列表                                         | `["gemini-1.0-pro-vision-latest", ...]`               |
+| `TOOLS_CODE_EXECUTION_ENABLED` | 可选，是否启用代码执行工具                                     | `false`                                               |
+| `SHOW_SEARCH_LINK`           | 可选，是否在响应中显示搜索结果链接                             | `true`                                                |
+| `SHOW_THINKING_PROCESS`      | 可选，是否显示模型思考过程                                     | `true`                                                |
+| `THINKING_MODELS`            | 可选，支持思考功能的模型列表                                   | `[]`                                                  |
+| `THINKING_BUDGET_MAP`        | 可选，思考功能预算映射 (模型名:预算值)                         | `{}`                                                  |
+| `URL_NORMALIZATION_ENABLED`  | 可选，是否启用智能路由映射功能                                 | `false`                                               |
+| `BASE_URL`                   | 可选，Gemini API 基础 URL，默认无需修改                        | `https://generativelanguage.googleapis.com/v1beta`    |
+| `MAX_FAILURES`               | 可选，允许单个key失败的次数                                    | `3`                                                   |
+| `MAX_RETRIES`                | 可选，API 请求失败时的最大重试次数                             | `3`                                                   |
+| `CHECK_INTERVAL_HOURS`       | 可选，检查禁用 Key 是否恢复的时间间隔 (小时)                   | `1`                                                   |
+| `TIMEZONE`                   | 可选，应用程序使用的时区                                       | `Asia/Shanghai`                                       |
+| `TIME_OUT`                   | 可选，请求超时时间 (秒)                                        | `300`                                                 |
+| `PROXIES`                    | 可选，代理服务器列表 (例如 `http://user:pass@host:port`, `socks5://host:port`) | `[]`                                                  |
+| `LOG_LEVEL`                  | 可选，日志级别，例如 DEBUG, INFO, WARNING, ERROR, CRITICAL     | `INFO`                                                |
+| `AUTO_DELETE_ERROR_LOGS_ENABLED` | 可选，是否开启自动删除错误日志                                 | `true`                                                |
+| `AUTO_DELETE_ERROR_LOGS_DAYS`  | 可选，自动删除多少天前的错误日志 (例如 1, 7, 30)             | `7`                                                   |
+| `AUTO_DELETE_REQUEST_LOGS_ENABLED`| 可选，是否开启自动删除请求日志                               | `false`                                               |
+| `AUTO_DELETE_REQUEST_LOGS_DAYS` | 可选，自动删除多少天前的请求日志 (例如 1, 7, 30)           | `30`                                                  |
+| `SAFETY_SETTINGS`            | 可选，安全设置 (JSON 字符串格式)，用于配置内容安全阈值。示例值可能需要根据实际模型支持情况调整。 | `[{"category": "HARM_CATEGORY_HARASSMENT", "threshold": "OFF"}, {"category": "HARM_CATEGORY_HATE_SPEECH", "threshold": "OFF"}, {"category": "HARM_CATEGORY_SEXUALLY_EXPLICIT", "threshold": "OFF"}, {"category": "HARM_CATEGORY_DANGEROUS_CONTENT", "threshold": "OFF"}, {"category": "HARM_CATEGORY_CIVIC_INTEGRITY", "threshold": "BLOCK_NONE"}]` |
+| **TTS 相关**                 |                                                          |                                                       |
+| `TTS_MODEL`                  | 可选，TTS 模型名称                                           | `gemini-2.5-flash-preview-tts`                        |
+| `TTS_VOICE_NAME`             | 可选，TTS 语音名称                                           | `Zephyr`                                              |
+| `TTS_SPEED`                  | 可选，TTS 语速                                               | `normal`                                              |
+| **图像生成相关**             |                                                          |                                                       |
+| `PAID_KEY`                   | 可选，付费版API Key，用于图片生成等高级功能                    | `your-paid-api-key`                                   |
+| `CREATE_IMAGE_MODEL`         | 可选，图片生成模型                                             | `imagen-3.0-generate-002`                             |
+| `UPLOAD_PROVIDER`            | 可选，图片上传提供商: `smms`, `picgo`, `cloudflare_imgbed`     | `smms`                                                |
+| `SMMS_SECRET_TOKEN`          | 可选，SM.MS图床的API Token                                     | `your-smms-token`                                     |
+| `PICGO_API_KEY`              | 可选，[PicoGo](https://www.picgo.net/)图床的API Key                                      | `your-picogo-apikey`                                  |
+| `CLOUDFLARE_IMGBED_URL`      | 可选，[CloudFlare](https://github.com/MarSeventh/CloudFlare-ImgBed) 图床上传地址                                  | `https://xxxxxxx.pages.dev/upload`                    |
+| `CLOUDFLARE_IMGBED_AUTH_CODE`| 可选，CloudFlare图床的鉴权key                                  | `your-cloudflare-imgber-auth-code`                    |
+| `CLOUDFLARE_IMGBED_UPLOAD_FOLDER`| 可选，CloudFlare图床的上传文件夹路径                        | `""`                                                  |
+| **流式优化器相关**           |                                                          |                                                       |
+| `STREAM_OPTIMIZER_ENABLED`   | 可选，是否启用流式输出优化                                     | `false`                                               |
+| `STREAM_MIN_DELAY`           | 可选，流式输出最小延迟                                         | `0.016`                                               |
+| `STREAM_MAX_DELAY`           | 可选，流式输出最大延迟                                         | `0.024`                                               |
+| `STREAM_SHORT_TEXT_THRESHOLD`| 可选，短文本阈值                                               | `10`                                                  |
+| `STREAM_LONG_TEXT_THRESHOLD` | 可选，长文本阈值                                               | `50`                                                  |
+| `STREAM_CHUNK_SIZE`          | 可选，流式输出块大小                                           | `5`                                                   |
+| **伪流式 (Fake Stream) 相关** |                                                          |                                                       |
+| `FAKE_STREAM_ENABLED`        | 可选，是否启用伪流式传输，用于不支持流式的模型或场景           | `false`                                               |
+| `FAKE_STREAM_EMPTY_DATA_INTERVAL_SECONDS` | 可选，伪流式传输时发送心跳空数据的间隔秒数                   | `5`                                                   |
+
+## ⚙️ API 端点
+
+以下是服务提供的主要 API 端点：
+
+### Gemini API 相关 (`(/gemini)/v1beta`)
+
+* `GET /models`: 列出可用的 Gemini 模型。
+* `POST /models/{model_name}:generateContent`: 使用指定的 Gemini 模型生成内容。
+* `POST /models/{model_name}:streamGenerateContent`: 使用指定的 Gemini 模型流式生成内容。
+
+### OpenAI API 相关
+
+* `GET (/hf)/v1/models`: 列出可用的模型 (底层用的gemini格式)。
+* `POST (/hf)/v1/chat/completions`: 进行聊天补全 (底层用的gemini格式, 支持流式传输)。
+* `POST (/hf)/v1/embeddings`: 创建文本嵌入 (底层用的gemini格式)。
+* `POST (/hf)/v1/images/generations`: 生成图像 (底层用的gemini格式)。
+* `GET /openai/v1/models`: 列出可用的模型 (底层用的openai格式)。
+* `POST /openai/v1/chat/completions`: 进行聊天补全 (底层用的openai格式, 支持流式传输, 可防止截断，速度也快)。
+* `POST /openai/v1/embeddings`: 创建文本嵌入 (底层用的openai格式)。
+* `POST /openai/v1/images/generations`: 生成图像 (底层用的openai格式)。
+
+## 🤝 贡献
+
+欢迎提交 Pull Request 或 Issue。
+
+## 🎉 特别鸣谢
+
+特别鸣谢以下项目和平台为本项目提供图床服务:
+
+* [PicGo](https://www.picgo.net/)
+* [SM.MS](https://smms.app/)
+* [CloudFlare-ImgBed](https://github.com/MarSeventh/CloudFlare-ImgBed) 开源项目
+
+## 🙏 感谢贡献者
+
+感谢所有为本项目做出贡献的开发者！
+
+[![Contributors](https://contrib.rocks/image?repo=snailyp/gemini-balance)](https://github.com/snailyp/gemini-balance/graphs/contributors)
+
+## ⭐ Star History
+
+[![Star History Chart](https://api.star-history.com/svg?repos=snailyp/gemini-balance&type=Date)](https://star-history.com/#snailyp/gemini-balance&Date)
+
+## 💖 友情项目
+
+* **[OneLine](https://github.com/chengtx809/OneLine)** by [chengtx809](https://github.com/chengtx809) - OneLine一线：AI驱动的热点事件时间轴生成工具
+
+## 🎁 项目支持
+
+如果你觉得这个项目对你有帮助，可以考虑通过 [爱发电](https://afdian.com/a/snaily) 支持我。
+
+## 许可证
+
+本项目采用 CC BY-NC 4.0（署名-非商业性使用）协议，禁止任何形式的商业倒卖服务，详见 LICENSE 文件。

VERSION

@@ -1 +1 @@
-2.1.2
+2.1.9

app/config/config.py

@@ -1,26 +1,55 @@
 """
 应用程序配置模块
 """
+
 import datetime
 import json
-from typing import List, Any, Dict, Type
+from typing import Any, Dict, List, Type, get_args, get_origin
 
-from pydantic import ValidationError
+from pydantic import ValidationError, ValidationInfo, field_validator
 from pydantic_settings import BaseSettings
-from sqlalchemy import insert, update, select
+from sqlalchemy import insert, select, update
 
-from app.core.constants import API_VERSION, DEFAULT_CREATE_IMAGE_MODEL, DEFAULT_FILTER_MODELS, DEFAULT_MODEL, DEFAULT_SAFETY_SETTINGS, DEFAULT_STREAM_CHUNK_SIZE, DEFAULT_STREAM_LONG_TEXT_THRESHOLD, DEFAULT_STREAM_MAX_DELAY, DEFAULT_STREAM_MIN_DELAY, DEFAULT_STREAM_SHORT_TEXT_THRESHOLD, DEFAULT_TIMEOUT, MAX_RETRIES
+from app.core.constants import (
+    API_VERSION,
+    DEFAULT_CREATE_IMAGE_MODEL,
+    DEFAULT_FILTER_MODELS,
+    DEFAULT_MODEL,
+    DEFAULT_SAFETY_SETTINGS,
+    DEFAULT_STREAM_CHUNK_SIZE,
+    DEFAULT_STREAM_LONG_TEXT_THRESHOLD,
+    DEFAULT_STREAM_MAX_DELAY,
+    DEFAULT_STREAM_MIN_DELAY,
+    DEFAULT_STREAM_SHORT_TEXT_THRESHOLD,
+    DEFAULT_TIMEOUT,
+    MAX_RETRIES,
+)
 from app.log.logger import Logger
 
 
 class Settings(BaseSettings):
     # 数据库配置
-    MYSQL_HOST: str
-    MYSQL_PORT: int
-    MYSQL_USER: str
-    MYSQL_PASSWORD: str
-    MYSQL_DATABASE: str
-    
+    DATABASE_TYPE: str = "mysql"  # sqlite 或 mysql
+    SQLITE_DATABASE: str = "default_db"
+    MYSQL_HOST: str = ""
+    MYSQL_PORT: int = 3306
+    MYSQL_USER: str = ""
+    MYSQL_PASSWORD: str = ""
+    MYSQL_DATABASE: str = ""
+    MYSQL_SOCKET: str = ""
+
+    # 验证 MySQL 配置
+    @field_validator(
+        "MYSQL_HOST", "MYSQL_PORT", "MYSQL_USER", "MYSQL_PASSWORD", "MYSQL_DATABASE"
+    )
+    def validate_mysql_config(cls, v: Any, info: ValidationInfo) -> Any:
+        if info.data.get("DATABASE_TYPE") == "mysql":
+            if v is None or v == "":
+                raise ValueError(
+                    "MySQL configuration is required when DATABASE_TYPE is 'mysql'"
+                )
+        return v
+
     # API相关配置
     API_KEYS: List[str]
     ALLOWED_TOKENS: List[str]
@@ -30,7 +59,16 @@ class Settings(BaseSettings):
     TEST_MODEL: str = DEFAULT_MODEL
     TIME_OUT: int = DEFAULT_TIMEOUT
     MAX_RETRIES: int = MAX_RETRIES
-    PROXIES: List[str] = [] # 新增：代理服务器列表
+    PROXIES: List[str] = []
+    PROXIES_USE_CONSISTENCY_HASH_BY_API_KEY: bool = True  # 是否使用一致性哈希来选择代理
+    VERTEX_API_KEYS: List[str] = []
+    VERTEX_EXPRESS_BASE_URL: str = "https://aiplatform.googleapis.com/v1beta1/publishers/google"
+
+    # 智能路由配置
+    URL_NORMALIZATION_ENABLED: bool = False  # 是否启用智能路由映射功能
+
+    # 自定义 Headers
+    CUSTOM_HEADERS: Dict[str, str] = {}
 
     # 模型相关配置
     SEARCH_MODELS: List[str] = ["gemini-2.0-flash-exp"]
@@ -39,9 +77,14 @@ class Settings(BaseSettings):
     TOOLS_CODE_EXECUTION_ENABLED: bool = False
     SHOW_SEARCH_LINK: bool = True
     SHOW_THINKING_PROCESS: bool = True
-    THINKING_MODELS: List[str] = [] # 新增：用于思考过程的模型列表
-    THINKING_BUDGET_MAP: Dict[str, float] = {} # 新增：模型对应的预算映射
-    
+    THINKING_MODELS: List[str] = []
+    THINKING_BUDGET_MAP: Dict[str, float] = {}
+
+    # TTS相关配置
+    TTS_MODEL: str = "gemini-2.5-flash-preview-tts"
+    TTS_VOICE_NAME: str = "Zephyr"
+    TTS_SPEED: str = "normal"
+
     # 图像生成相关配置
     PAID_KEY: str = ""
     CREATE_IMAGE_MODEL: str = DEFAULT_CREATE_IMAGE_MODEL
@@ -50,7 +93,8 @@ class Settings(BaseSettings):
     PICGO_API_KEY: str = ""
     CLOUDFLARE_IMGBED_URL: str = ""
     CLOUDFLARE_IMGBED_AUTH_CODE: str = ""
-    
+    CLOUDFLARE_IMGBED_UPLOAD_FOLDER: str = ""
+
     # 流式输出优化器配置
     STREAM_OPTIMIZER_ENABLED: bool = False
     STREAM_MIN_DELAY: float = DEFAULT_STREAM_MIN_DELAY
@@ -59,17 +103,26 @@ class Settings(BaseSettings):
     STREAM_LONG_TEXT_THRESHOLD: int = DEFAULT_STREAM_LONG_TEXT_THRESHOLD
     STREAM_CHUNK_SIZE: int = DEFAULT_STREAM_CHUNK_SIZE
 
+    # 假流式配置 (Fake Streaming Configuration)
+    FAKE_STREAM_ENABLED: bool = False  # 是否启用假流式输出
+    FAKE_STREAM_EMPTY_DATA_INTERVAL_SECONDS: int = 5  # 假流式发送空数据的间隔时间（秒）
+
     # 调度器配置
-    CHECK_INTERVAL_HOURS: int = 1 # 默认检查间隔为1小时
-    TIMEZONE: str = "Asia/Shanghai" # 默认时区
-    
-    # github 
+    CHECK_INTERVAL_HOURS: int = 1  # 默认检查间隔为1小时
+    TIMEZONE: str = "Asia/Shanghai"  # 默认时区
+
+    # github
     GITHUB_REPO_OWNER: str = "snailyp"
     GITHUB_REPO_NAME: str = "gemini-balance"
 
     # 日志配置
-    LOG_LEVEL: str = "INFO" # 默认日志级别
-    SAFETY_SETTINGS: List[Dict[str, str]] = DEFAULT_SAFETY_SETTINGS # 新增：安全设置
+    LOG_LEVEL: str = "INFO"
+    AUTO_DELETE_ERROR_LOGS_ENABLED: bool = True
+    AUTO_DELETE_ERROR_LOGS_DAYS: int = 7
+    AUTO_DELETE_REQUEST_LOGS_ENABLED: bool = False
+    AUTO_DELETE_REQUEST_LOGS_DAYS: int = 30
+    SAFETY_SETTINGS: List[Dict[str, str]] = DEFAULT_SAFETY_SETTINGS
+
 
     def __init__(self, **kwargs):
         super().__init__(**kwargs)
@@ -77,80 +130,120 @@ class Settings(BaseSettings):
         if not self.AUTH_TOKEN and self.ALLOWED_TOKENS:
             self.AUTH_TOKEN = self.ALLOWED_TOKENS[0]
 
+
 # 创建全局配置实例
 settings = Settings()
 
+
 def _parse_db_value(key: str, db_value: str, target_type: Type) -> Any:
     """尝试将数据库字符串值解析为目标 Python 类型"""
-    from app.log.logger import get_config_logger # 函数内导入
-    logger = get_config_logger() # 函数内初始化
+    from app.log.logger import get_config_logger
+
+    logger = get_config_logger()
     try:
-        # 处理 List[str]
-        if target_type == List[str]:
-            try:
-                parsed = json.loads(db_value)
-                if isinstance(parsed, list):
-                    return [str(item) for item in parsed]
-            except json.JSONDecodeError:
-                return [item.strip() for item in db_value.split(',') if item.strip()]
-            logger.warning(f"Could not parse '{db_value}' as List[str] for key '{key}', falling back to comma split or empty list.")
-            return [item.strip() for item in db_value.split(',') if item.strip()]
-        # 处理 Dict[str, float]
-        elif target_type == Dict[str, float]:
-            parsed_dict = {}
-            try:
-                # First attempt: standard JSON parsing
-                parsed = json.loads(db_value)
-                if isinstance(parsed, dict):
-                    parsed_dict = {str(k): float(v) for k, v in parsed.items()}
-                else:
-                     logger.warning(f"Parsed DB value for key '{key}' is not a dictionary type. Value: {db_value}")
-            except (json.JSONDecodeError, ValueError, TypeError) as e1:
-                # Second attempt: try replacing single quotes if JSONDecodeError occurred
-                if isinstance(e1, json.JSONDecodeError) and "'" in db_value:
-                    logger.warning(f"Failed initial JSON parse for key '{key}'. Attempting to replace single quotes. Error: {e1}")
-                    try:
-                        corrected_db_value = db_value.replace("'", '"')
-                        parsed = json.loads(corrected_db_value)
-                        if isinstance(parsed, dict):
-                             parsed_dict = {str(k): float(v) for k, v in parsed.items()}
+        origin_type = get_origin(target_type)
+        args = get_args(target_type)
+
+        # 处理 List 类型
+        if origin_type is list:
+            # 处理 List[str]
+            if args and args[0] == str:
+                try:
+                    parsed = json.loads(db_value)
+                    if isinstance(parsed, list):
+                        return [str(item) for item in parsed]
+                except json.JSONDecodeError:
+                    return [item.strip() for item in db_value.split(",") if item.strip()]
+                logger.warning(
+                    f"Could not parse '{db_value}' as List[str] for key '{key}', falling back to comma split or empty list."
+                )
+                return [item.strip() for item in db_value.split(",") if item.strip()]
+            # 处理 List[Dict[str, str]]
+            elif args and get_origin(args[0]) is dict:
+                try:
+                    parsed = json.loads(db_value)
+                    if isinstance(parsed, list):
+                        valid = all(
+                            isinstance(item, dict)
+                            and all(isinstance(k, str) for k in item.keys())
+                            and all(isinstance(v, str) for v in item.values())
+                            for item in parsed
+                        )
+                        if valid:
+                            return parsed
                         else:
-                            logger.warning(f"Parsed DB value (after quote replacement) for key '{key}' is not a dictionary type. Value: {corrected_db_value}")
-                    except (json.JSONDecodeError, ValueError, TypeError) as e2:
-                         logger.error(f"Could not parse '{db_value}' as Dict[str, float] for key '{key}' even after replacing quotes: {e2}. Returning empty dict.")
-                else:
-                    # Log other errors (ValueError, TypeError) or JSON errors without single quotes
-                    logger.error(f"Could not parse '{db_value}' as Dict[str, float] for key '{key}': {e1}. Returning empty dict.")
-            return parsed_dict # Return the parsed dict or an empty one if all attempts fail
-        # 处理 List[Dict[str, str]]
-        elif target_type == List[Dict[str, str]]:
-            try:
-                parsed = json.loads(db_value)
-                if isinstance(parsed, list):
-                    # 验证列表中的每个元素是否为字典，并且键和值都是字符串
-                    valid = all(
-                        isinstance(item, dict) and
-                        all(isinstance(k, str) for k in item.keys()) and
-                        all(isinstance(v, str) for v in item.values())
-                        for item in parsed
-                    )
-                    if valid:
-                        return parsed
+                            logger.warning(
+                                f"Invalid structure in List[Dict[str, str]] for key '{key}'. Value: {db_value}"
+                            )
+                            return []
                     else:
-                        logger.warning(f"Invalid structure in List[Dict[str, str]] for key '{key}'. Value: {db_value}")
-                        return [] # 或者返回默认值？这里返回空列表
-                else:
-                     logger.warning(f"Parsed DB value for key '{key}' is not a list type. Value: {db_value}")
-                     return []
-            except json.JSONDecodeError:
-                logger.error(f"Could not parse '{db_value}' as JSON for List[Dict[str, str]] for key '{key}'. Returning empty list.")
-                return []
-            except Exception as e:
-                logger.error(f"Error parsing List[Dict[str, str]] for key '{key}': {e}. Value: {db_value}. Returning empty list.")
-                return []
+                        logger.warning(
+                            f"Parsed DB value for key '{key}' is not a list type. Value: {db_value}"
+                        )
+                        return []
+                except json.JSONDecodeError:
+                    logger.error(
+                        f"Could not parse '{db_value}' as JSON for List[Dict[str, str]] for key '{key}'. Returning empty list."
+                    )
+                    return []
+                except Exception as e:
+                    logger.error(
+                        f"Error parsing List[Dict[str, str]] for key '{key}': {e}. Value: {db_value}. Returning empty list."
+                    )
+                    return []
+        # 处理 Dict 类型
+        elif origin_type is dict:
+            # 处理 Dict[str, str]
+            if args and args == (str, str):
+                parsed_dict = {}
+                try:
+                    parsed = json.loads(db_value)
+                    if isinstance(parsed, dict):
+                        parsed_dict = {str(k): str(v) for k, v in parsed.items()}
+                    else:
+                        logger.warning(
+                            f"Parsed DB value for key '{key}' is not a dictionary type. Value: {db_value}"
+                        )
+                except json.JSONDecodeError:
+                     logger.error(f"Could not parse '{db_value}' as Dict[str, str] for key '{key}'. Returning empty dict.")
+                return parsed_dict
+            # 处理 Dict[str, float]
+            elif args and args == (str, float):
+                parsed_dict = {}
+                try:
+                    parsed = json.loads(db_value)
+                    if isinstance(parsed, dict):
+                        parsed_dict = {str(k): float(v) for k, v in parsed.items()}
+                    else:
+                        logger.warning(
+                            f"Parsed DB value for key '{key}' is not a dictionary type. Value: {db_value}"
+                        )
+                except (json.JSONDecodeError, ValueError, TypeError) as e1:
+                    if isinstance(e1, json.JSONDecodeError) and "'" in db_value:
+                        logger.warning(
+                            f"Failed initial JSON parse for key '{key}'. Attempting to replace single quotes. Error: {e1}"
+                        )
+                        try:
+                            corrected_db_value = db_value.replace("'", '"')
+                            parsed = json.loads(corrected_db_value)
+                            if isinstance(parsed, dict):
+                                parsed_dict = {str(k): float(v) for k, v in parsed.items()}
+                            else:
+                                logger.warning(
+                                    f"Parsed DB value (after quote replacement) for key '{key}' is not a dictionary type. Value: {corrected_db_value}"
+                                )
+                        except (json.JSONDecodeError, ValueError, TypeError) as e2:
+                            logger.error(
+                                f"Could not parse '{db_value}' as Dict[str, float] for key '{key}' even after replacing quotes: {e2}. Returning empty dict."
+                            )
+                    else:
+                        logger.error(
+                            f"Could not parse '{db_value}' as Dict[str, float] for key '{key}': {e1}. Returning empty dict."
+                        )
+                return parsed_dict
         # 处理 bool
         elif target_type == bool:
-            return db_value.lower() in ('true', '1', 'yes', 'on')
+            return db_value.lower() in ("true", "1", "yes", "on")
         # 处理 int
         elif target_type == int:
             return int(db_value)
@@ -161,8 +254,11 @@ def _parse_db_value(key: str, db_value: str, target_type: Type) -> Any:
         else:
             return db_value
     except (ValueError, TypeError, json.JSONDecodeError) as e:
-        logger.warning(f"Failed to parse db_value '{db_value}' for key '{key}' as type {target_type}: {e}. Using original string value.")
-        return db_value # 解析失败则返回原始字符串
+        logger.warning(
+            f"Failed to parse db_value '{db_value}' for key '{key}' as type {target_type}: {e}. Using original string value."
+        )
+        return db_value  # 解析失败则返回原始字符串
+
 
 async def sync_initial_settings():
     """
@@ -171,8 +267,9 @@ async def sync_initial_settings():
     2. 将数据库设置合并到内存 settings (数据库优先)。
     3. 将最终的内存 settings 同步回数据库。
     """
-    from app.log.logger import get_config_logger # 函数内导入
-    logger = get_config_logger() # 函数内初始化
+    from app.log.logger import get_config_logger
+
+    logger = get_config_logger()
     # 延迟导入以避免循环依赖和确保数据库连接已初始化
     from app.database.connection import database
     from app.database.models import Settings as SettingsModel
@@ -185,7 +282,9 @@ async def sync_initial_settings():
             await database.connect()
             logger.info("Database connection established for initial sync.")
         except Exception as e:
-            logger.error(f"Failed to connect to database for initial settings sync: {e}. Skipping sync.")
+            logger.error(
+                f"Failed to connect to database for initial settings sync: {e}. Skipping sync."
+            )
             return
 
     try:
@@ -194,18 +293,30 @@ async def sync_initial_settings():
         try:
             query = select(SettingsModel.key, SettingsModel.value)
             results = await database.fetch_all(query)
-            db_settings_raw = [{"key": row["key"], "value": row["value"]} for row in results]
+            db_settings_raw = [
+                {"key": row["key"], "value": row["value"]} for row in results
+            ]
             logger.info(f"Fetched {len(db_settings_raw)} settings from database.")
         except Exception as e:
-            logger.error(f"Failed to fetch settings from database: {e}. Proceeding with environment/dotenv settings.")
+            logger.error(
+                f"Failed to fetch settings from database: {e}. Proceeding with environment/dotenv settings."
+            )
             # 即使数据库读取失败，也要继续执行，确保基于 env/dotenv 的配置能同步到数据库
 
-        db_settings_map: Dict[str, str] = {s['key']: s['value'] for s in db_settings_raw}
+        db_settings_map: Dict[str, str] = {
+            s["key"]: s["value"] for s in db_settings_raw
+        }
 
         # 2. 将数据库设置合并到内存 settings (数据库优先)
         updated_in_memory = False
 
         for key, db_value in db_settings_map.items():
+            if key == "DATABASE_TYPE":
+                logger.debug(
+                    f"Skipping update of '{key}' in memory from database. "
+                    "This setting is controlled by environment/dotenv."
+                )
+                continue
             if hasattr(settings, key):
                 target_type = Settings.__annotations__.get(key)
                 if target_type:
@@ -218,35 +329,46 @@ async def sync_initial_settings():
                         if parsed_db_value != memory_value:
                             # 检查类型是否匹配，以防解析函数返回了不兼容的类型
                             type_match = False
-                            if target_type == List[str] and isinstance(parsed_db_value, list):
-                                type_match = True
-                            elif target_type == Dict[str, float] and isinstance(parsed_db_value, dict):
-                                type_match = True
-                            elif target_type not in (List[str], Dict[str, float]) and isinstance(parsed_db_value, target_type):
+                            origin_type = get_origin(target_type)
+                            if origin_type:  # It's a generic type
+                                if isinstance(parsed_db_value, origin_type):
+                                    type_match = True
+                            # It's a non-generic type, or a specific generic we want to handle
+                            elif isinstance(parsed_db_value, target_type):
                                 type_match = True
 
                             if type_match:
                                 setattr(settings, key, parsed_db_value)
-                                logger.debug(f"Updated setting '{key}' in memory from database value ({target_type}).")
+                                logger.debug(
+                                    f"Updated setting '{key}' in memory from database value ({target_type})."
+                                )
                                 updated_in_memory = True
                             else:
-                                logger.warning(f"Parsed DB value type mismatch for key '{key}'. Expected {target_type}, got {type(parsed_db_value)}. Skipping update.")
+                                logger.warning(
+                                    f"Parsed DB value type mismatch for key '{key}'. Expected {target_type}, got {type(parsed_db_value)}. Skipping update."
+                                )
 
                     except Exception as e:
-                        logger.error(f"Error processing database setting for key '{key}': {e}")
+                        logger.error(
+                            f"Error processing database setting for key '{key}': {e}"
+                        )
             else:
-                 logger.warning(f"Database setting '{key}' not found in Settings model definition. Ignoring.")
-
+                logger.warning(
+                    f"Database setting '{key}' not found in Settings model definition. Ignoring."
+                )
 
         # 如果内存中有更新，重新验证 Pydantic 模型（可选但推荐）
         if updated_in_memory:
             try:
                 # 重新加载以确保类型转换和验证
                 settings = Settings(**settings.model_dump())
-                logger.info("Settings object re-validated after merging database values.")
+                logger.info(
+                    "Settings object re-validated after merging database values."
+                )
             except ValidationError as e:
-                 logger.error(f"Validation error after merging database settings: {e}. Settings might be inconsistent.")
-
+                logger.error(
+                    f"Validation error after merging database settings: {e}. Settings might be inconsistent."
+                )
 
         # 3. 将最终的内存 settings 同步回数据库
         final_memory_settings = settings.model_dump()
@@ -257,21 +379,30 @@ async def sync_initial_settings():
         existing_db_keys = set(db_settings_map.keys())
 
         for key, value in final_memory_settings.items():
+            if key == "DATABASE_TYPE":
+                logger.debug(
+                    f"Skipping synchronization of '{key}' to database. "
+                    "This setting is controlled by environment/dotenv."
+                )
+                continue
+
             # 序列化值为字符串或 JSON 字符串
-            if isinstance(value, (list, dict)): # 处理列表和字典
-                db_value = json.dumps(value, ensure_ascii=False) # 使用 ensure_ascii=False 以支持非 ASCII 字符
+            if isinstance(value, (list, dict)):
+                db_value = json.dumps(
+                    value, ensure_ascii=False
+                )
             elif isinstance(value, bool):
                 db_value = str(value).lower()
-            elif value is None: # 处理 None 值
-                db_value = "" # 或者根据需要设为 NULL 或其他标记
+            elif value is None:
+                db_value = ""
             else:
                 db_value = str(value)
 
             data = {
-                'key': key,
-                'value': db_value,
-                'description': f"{key} configuration setting", # 默认描述
-                'updated_at': now
+                "key": key,
+                "value": db_value,
+                "description": f"{key} configuration setting",
+                "updated_at": now,
             }
 
             if key in existing_db_keys:
@@ -280,7 +411,7 @@ async def sync_initial_settings():
                     settings_to_update.append(data)
             else:
                 # 如果键不在数据库中，则插入
-                data['created_at'] = now
+                data["created_at"] = now
                 settings_to_insert.append(data)
 
         # 在事务中执行批量插入和更新
@@ -289,48 +420,78 @@ async def sync_initial_settings():
                 async with database.transaction():
                     if settings_to_insert:
                         # 获取现有描述以避免覆盖
-                        query_existing = select(SettingsModel.key, SettingsModel.description).where(SettingsModel.key.in_([s['key'] for s in settings_to_insert]))
-                        existing_desc = {row['key']: row['description'] for row in await database.fetch_all(query_existing)}
+                        query_existing = select(
+                            SettingsModel.key, SettingsModel.description
+                        ).where(
+                            SettingsModel.key.in_(
+                                [s["key"] for s in settings_to_insert]
+                            )
+                        )
+                        existing_desc = {
+                            row["key"]: row["description"]
+                            for row in await database.fetch_all(query_existing)
+                        }
                         for item in settings_to_insert:
-                            item['description'] = existing_desc.get(item['key'], item['description'])
+                            item["description"] = existing_desc.get(
+                                item["key"], item["description"]
+                            )
 
                         query_insert = insert(SettingsModel).values(settings_to_insert)
                         await database.execute(query=query_insert)
-                        logger.info(f"Synced (inserted) {len(settings_to_insert)} settings to database.")
+                        logger.info(
+                            f"Synced (inserted) {len(settings_to_insert)} settings to database."
+                        )
 
                     if settings_to_update:
                         # 获取现有描述以避免覆盖
-                        query_existing = select(SettingsModel.key, SettingsModel.description).where(SettingsModel.key.in_([s['key'] for s in settings_to_update]))
-                        existing_desc = {row['key']: row['description'] for row in await database.fetch_all(query_existing)}
+                        query_existing = select(
+                            SettingsModel.key, SettingsModel.description
+                        ).where(
+                            SettingsModel.key.in_(
+                                [s["key"] for s in settings_to_update]
+                            )
+                        )
+                        existing_desc = {
+                            row["key"]: row["description"]
+                            for row in await database.fetch_all(query_existing)
+                        }
 
                         for setting_data in settings_to_update:
-                            setting_data['description'] = existing_desc.get(setting_data['key'], setting_data['description'])
+                            setting_data["description"] = existing_desc.get(
+                                setting_data["key"], setting_data["description"]
+                            )
                             query_update = (
                                 update(SettingsModel)
-                                .where(SettingsModel.key == setting_data['key'])
+                                .where(SettingsModel.key == setting_data["key"])
                                 .values(
-                                    value=setting_data['value'],
-                                    description=setting_data['description'],
-                                    updated_at=setting_data['updated_at']
+                                    value=setting_data["value"],
+                                    description=setting_data["description"],
+                                    updated_at=setting_data["updated_at"],
                                 )
                             )
                             await database.execute(query=query_update)
-                        logger.info(f"Synced (updated) {len(settings_to_update)} settings to database.")
+                        logger.info(
+                            f"Synced (updated) {len(settings_to_update)} settings to database."
+                        )
             except Exception as e:
-                logger.error(f"Failed to sync settings to database during startup: {str(e)}")
+                logger.error(
+                    f"Failed to sync settings to database during startup: {str(e)}"
+                )
         else:
-            logger.info("No setting changes detected between memory and database during initial sync.")
+            logger.info(
+                "No setting changes detected between memory and database during initial sync."
+            )
 
         # 刷新日志等级
         Logger.update_log_levels(final_memory_settings.get("LOG_LEVEL"))
-        
+
     except Exception as e:
         logger.error(f"An unexpected error occurred during initial settings sync: {e}")
     finally:
         if database.is_connected:
-             try:
-                 pass
-             except Exception as e:
-                 logger.error(f"Error disconnecting database after initial sync: {e}")
+            try:
+                pass
+            except Exception as e:
+                logger.error(f"Error disconnecting database after initial sync: {e}")
 
     logger.info("Initial settings synchronization finished.")

app/core/application.py

@@ -1,35 +1,32 @@
 from contextlib import asynccontextmanager
-from pathlib import Path # Add pathlib import
+from pathlib import Path
+
 from fastapi import FastAPI
 from fastapi.staticfiles import StaticFiles
 from fastapi.templating import Jinja2Templates
 
 from app.config.config import settings, sync_initial_settings
+from app.database.connection import connect_to_db, disconnect_from_db
+from app.database.initialization import initialize_database
+from app.exception.exceptions import setup_exception_handlers
 from app.log.logger import get_application_logger
 from app.middleware.middleware import setup_middlewares
-from app.exception.exceptions import setup_exception_handlers
 from app.router.routes import setup_routers
+from app.scheduler.scheduled_tasks import start_scheduler, stop_scheduler
 from app.service.key.key_manager import get_key_manager_instance
-from app.database.connection import connect_to_db, disconnect_from_db
-from app.utils.helpers import get_current_version # Import from helpers
-from app.database.initialization import initialize_database
-from app.scheduler.key_checker import start_scheduler, stop_scheduler
 from app.service.update.update_service import check_for_updates
+from app.utils.helpers import get_current_version
 
 logger = get_application_logger()
 
-# Define project paths using pathlib
-# Assuming this file is at app/core/application.py
 PROJECT_ROOT = Path(__file__).resolve().parent.parent.parent
-# VERSION_FILE_PATH = PROJECT_ROOT / "VERSION" # Removed: Defined in helpers.py
 STATIC_DIR = PROJECT_ROOT / "app" / "static"
 TEMPLATES_DIR = PROJECT_ROOT / "app" / "templates"
 
-# Removed _get_current_version function definition, moved to helpers.py
-
 # 初始化模板引擎，并添加全局变量
 templates = Jinja2Templates(directory="app/templates")
 
+
 # 定义一个函数来更新模板全局变量
 def update_template_globals(app: FastAPI, update_info: dict):
     # Jinja2Templates 实例没有直接更新全局变量的方法
@@ -40,114 +37,105 @@ def update_template_globals(app: FastAPI, update_info: dict):
 
 
 # --- Helper functions for lifespan ---
-
 async def _setup_database_and_config(app_settings):
     """Initializes database, syncs settings, and initializes KeyManager."""
     initialize_database()
     logger.info("Database initialized successfully")
     await connect_to_db()
     await sync_initial_settings()
-    # Initialize KeyManager using potentially updated settings
-    await get_key_manager_instance(app_settings.API_KEYS)
+    await get_key_manager_instance(app_settings.API_KEYS, app_settings.VERTEX_API_KEYS)
     logger.info("Database, config sync, and KeyManager initialized successfully")
 
+
 async def _shutdown_database():
     """Disconnects from the database."""
     await disconnect_from_db()
 
+
 def _start_scheduler():
     """Starts the background scheduler."""
     try:
         start_scheduler()
         logger.info("Scheduler started successfully.")
     except Exception as e:
-         logger.error(f"Failed to start scheduler: {e}")
+        logger.error(f"Failed to start scheduler: {e}")
+
 
 def _stop_scheduler():
     """Stops the background scheduler."""
     stop_scheduler()
 
+
 async def _perform_update_check(app: FastAPI):
     """Checks for updates and stores the info in app.state."""
     update_available, latest_version, error_message = await check_for_updates()
-    current_version = get_current_version() # Use imported function
+    current_version = get_current_version()
     update_info = {
         "update_available": update_available,
         "latest_version": latest_version,
         "error_message": error_message,
-        "current_version": current_version
+        "current_version": current_version,
     }
-    # Ensure app.state exists and store update info
     if not hasattr(app, "state"):
         from starlette.datastructures import State
+
         app.state = State()
     app.state.update_info = update_info
     logger.info(f"Update check completed. Info: {update_info}")
 
-# --- Application Lifespan ---
 
 @asynccontextmanager
 async def lifespan(app: FastAPI):
     """
     Manages the application startup and shutdown events.
-    
+
     Args:
         app: FastAPI应用实例
     """
-    # Startup events
     logger.info("Application starting up...")
     try:
-        # Setup database, config, and KeyManager
-        await _setup_database_and_config(settings) # Pass settings object
-
-        # Perform update check after core components are ready
-        # await _perform_update_check(app) # Removed: Version check moved to frontend API call
-
-        # Start the scheduler
+        await _setup_database_and_config(settings)
+        await _perform_update_check(app)
         _start_scheduler()
 
     except Exception as e:
-        logger.critical(f"Critical error during application startup: {str(e)}", exc_info=True)
-        # Depending on the severity, you might want to prevent the app from fully starting
-        # For now, we log critically and let it yield, potentially in a broken state.
-        # Consider adding more robust error handling here if startup failures should halt the app.
+        logger.critical(
+            f"Critical error during application startup: {str(e)}", exc_info=True
+        )
 
-    yield  # Application runs
+    yield
 
-    # Shutdown events
     logger.info("Application shutting down...")
     _stop_scheduler()
     await _shutdown_database()
 
+
 def create_app() -> FastAPI:
     """
     创建并配置FastAPI应用程序实例
-    
+
     Returns:
         FastAPI: 配置好的FastAPI应用程序实例
     """
-    # Removed: initialize_app() call
 
     # 创建FastAPI应用
-    # Read version from file for consistency
-    current_version = get_current_version() # Use imported function
+    current_version = get_current_version()
     app = FastAPI(
         title="Gemini Balance API",
         description="Gemini API代理服务，支持负载均衡和密钥管理",
         version=current_version,
-        lifespan=lifespan
+        lifespan=lifespan,
     )
 
-    # Initialize app.state early to ensure it exists before lifespan potentially uses it
     if not hasattr(app, "state"):
         from starlette.datastructures import State
+
         app.state = State()
-    # Set a default/initial state for update_info
     app.state.update_info = {
         "update_available": False,
         "latest_version": None,
         "error_message": "Initializing...",
-        "current_version": current_version # Use version read earlier
+        "current_version": current_version,
     }
 
     # 配置静态文件
@@ -155,11 +143,11 @@ def create_app() -> FastAPI:
 
     # 配置中间件
     setup_middlewares(app)
-    
+
     # 配置异常处理器
     setup_exception_handlers(app)
-    
+
     # 配置路由
     setup_routers(app)
-    
+
     return app

app/core/constants.py

@@ -76,4 +76,15 @@ DEFAULT_SAFETY_SETTINGS = [
         {"category": "HARM_CATEGORY_SEXUALLY_EXPLICIT", "threshold": "OFF"},
         {"category": "HARM_CATEGORY_DANGEROUS_CONTENT", "threshold": "OFF"},
         {"category": "HARM_CATEGORY_CIVIC_INTEGRITY", "threshold": "BLOCK_NONE"},
-    ]
+    ]
+
+TTS_VOICE_NAMES = [
+    "Zephyr", "Puck", "Charon", "Kore",
+    "Fenrir", "Leda", "Orus", "Aoede",
+    "Callirhoe", "Autonoe", "Enceladus", "Iapetus",
+    "Umbriel", "Algieba", "Despina", "Erinome",
+    "Algenib", "Rasalgethi", "Laomedeia", "Achernar",
+    "Alnilam", "Schedar", "Gacrux", "Pulcherrima",
+    "Achird", "Zubenelgenubi", "Vindemiatrix", "Sadachbia",
+    "Sadaltager", "Sulafat"
+]

app/database/connection.py

@@ -1,9 +1,10 @@
 """
 数据库连接池模块
 """
+from pathlib import Path
+from urllib.parse import quote_plus
 from databases import Database
 from sqlalchemy import create_engine, MetaData
-# from sqlalchemy.orm import sessionmaker # 不再需要
 from sqlalchemy.ext.declarative import declarative_base
 
 from app.config.config import settings
@@ -12,7 +13,19 @@ from app.log.logger import get_database_logger
 logger = get_database_logger()
 
 # 数据库URL
-DATABASE_URL = f"mysql+pymysql://{settings.MYSQL_USER}:{settings.MYSQL_PASSWORD}@{settings.MYSQL_HOST}:{settings.MYSQL_PORT}/{settings.MYSQL_DATABASE}"
+if settings.DATABASE_TYPE == "sqlite":
+    # 确保 data 目录存在
+    data_dir = Path("data")
+    data_dir.mkdir(exist_ok=True)
+    db_path = data_dir / settings.SQLITE_DATABASE
+    DATABASE_URL = f"sqlite:///{db_path}"
+elif settings.DATABASE_TYPE == "mysql":
+    if settings.MYSQL_SOCKET:
+        DATABASE_URL = f"mysql+pymysql://{settings.MYSQL_USER}:{quote_plus(settings.MYSQL_PASSWORD)}@/{settings.MYSQL_DATABASE}?unix_socket={settings.MYSQL_SOCKET}"
+    else:
+        DATABASE_URL = f"mysql+pymysql://{settings.MYSQL_USER}:{quote_plus(settings.MYSQL_PASSWORD)}@{settings.MYSQL_HOST}:{settings.MYSQL_PORT}/{settings.MYSQL_DATABASE}"
+else:
+    raise ValueError("Unsupported database type. Please set DATABASE_TYPE to 'sqlite' or 'mysql'.")
 
 # 创建数据库引擎
 # pool_pre_ping=True: 在从连接池获取连接前执行简单的 "ping" 测试，确保连接有效
@@ -24,24 +37,24 @@ metadata = MetaData()
 # 创建基类
 Base = declarative_base(metadata=metadata)
 
-# 创建数据库连接池，并配置连接池参数
+# 创建数据库连接池，并配置连接池参数，在sqlite中不使用连接池
 # min_size/max_size: 连接池的最小/最大连接数
 # pool_recycle=3600: 连接在池中允许存在的最大秒数（生命周期）。
 #                    设置为 3600 秒（1小时），确保在 MySQL 默认的 wait_timeout (通常8小时) 或其他网络超时之前回收连接。
 #                    如果遇到连接失效问题，可以尝试调低此值，使其小于实际的 wait_timeout 或网络超时时间。
 # databases 库会自动处理连接失效后的重连尝试。
-database = Database(DATABASE_URL, min_size=5, max_size=20, pool_recycle=1800) # Reduced recycle time to 30 mins
+if settings.DATABASE_TYPE == "sqlite":
+    database = Database(DATABASE_URL)
+else:
+    database = Database(DATABASE_URL, min_size=5, max_size=20, pool_recycle=1800)
 
-# 移除了 SessionLocal 和 get_db 函数
-
-# --- Async connection functions for lifespan/async routes ---
 async def connect_to_db():
     """
     连接到数据库
     """
     try:
         await database.connect()
-        logger.info("Connected to database")
+        logger.info(f"Connected to {settings.DATABASE_TYPE}")
     except Exception as e:
         logger.error(f"Failed to connect to database: {str(e)}")
         raise
@@ -53,6 +66,6 @@ async def disconnect_from_db():
     """
     try:
         await database.disconnect()
-        logger.info("Disconnected from database")
+        logger.info(f"Disconnected from {settings.DATABASE_TYPE}")
     except Exception as e:
         logger.error(f"Failed to disconnect from database: {str(e)}")

app/database/models.py

@@ -2,7 +2,7 @@
 数据库模型模块
 """
 import datetime
-from sqlalchemy import Column, Integer, String, Text, DateTime, JSON, Boolean # 添加 Boolean
+from sqlalchemy import Column, Integer, String, Text, DateTime, JSON, Boolean
 
 from app.database.connection import Base
 
@@ -42,17 +42,18 @@ class ErrorLog(Base):
     def __repr__(self):
         return f"<ErrorLog(id='{self.id}', gemini_key='{self.gemini_key}')>"
 
-# 新增 RequestLog 模型
+
 class RequestLog(Base):
     """
     API 请求日志表
     """
+
     __tablename__ = "t_request_log"
 
     id = Column(Integer, primary_key=True, autoincrement=True)
     request_time = Column(DateTime, default=datetime.datetime.now, comment="请求时间")
     model_name = Column(String(100), nullable=True, comment="模型名称")
-    api_key = Column(String(100), nullable=True, comment="使用的API密钥") # 考虑安全性，后续可优化
+    api_key = Column(String(100), nullable=True, comment="使用的API密钥")
     is_success = Column(Boolean, nullable=False, comment="请求是否成功")
     status_code = Column(Integer, nullable=True, comment="API响应状态码")
     latency_ms = Column(Integer, nullable=True, comment="请求耗时(毫秒)")

app/database/services.py

@@ -71,7 +71,7 @@ async def update_setting(key: str, value: str, description: Optional[str] = None
                 .values(
                     value=value,
                     description=description if description else setting["description"],
-                    updated_at=datetime.now() # Use datetime.now()
+                    updated_at=datetime.now()
                 )
             )
             await database.execute(query)
@@ -85,8 +85,8 @@ async def update_setting(key: str, value: str, description: Optional[str] = None
                     key=key,
                     value=value,
                     description=description,
-                    created_at=datetime.now(), # Use datetime.now()
-                    updated_at=datetime.now() # Use datetime.now()
+                    created_at=datetime.now(),
+                    updated_at=datetime.now()
                 )
             )
             await database.execute(query)
@@ -158,8 +158,8 @@ async def get_error_logs(
     error_code_search: Optional[str] = None,
     start_date: Optional[datetime] = None,
     end_date: Optional[datetime] = None,
-    sort_by: str = 'id',        # 新增排序字段
-    sort_order: str = 'desc'   # 新增排序顺序 ('asc' or 'desc')
+    sort_by: str = 'id',
+    sort_order: str = 'desc'
 ) -> List[Dict[str, Any]]:
     """
     获取错误日志，支持搜索、日期过滤和排序
@@ -189,7 +189,6 @@ async def get_error_logs(
             ErrorLog.request_time
         )
         
-        # Apply filters
         if key_search:
             query = query.where(ErrorLog.gemini_key.ilike(f"%{key_search}%"))
         if error_search:
@@ -200,41 +199,33 @@ async def get_error_logs(
         if start_date:
             query = query.where(ErrorLog.request_time >= start_date)
         if end_date:
-            # Use the datetime object directly for comparison
             query = query.where(ErrorLog.request_time < end_date)
         if error_code_search:
             try:
-                # Attempt to convert search string to integer for exact match
                 error_code_int = int(error_code_search)
                 query = query.where(ErrorLog.error_code == error_code_int)
             except ValueError:
-                # If conversion fails, log a warning and potentially skip this filter
-                # or handle as needed (e.g., return no results for invalid code format)
                 logger.warning(f"Invalid format for error_code_search: '{error_code_search}'. Expected an integer. Skipping error code filter.")
-                # Optionally, force no results if the format is invalid:
-                # query = query.where(False) # This ensures no rows are returned
 
-        # 添加排序逻辑
-        sort_column = getattr(ErrorLog, sort_by, ErrorLog.id) # 获取排序字段，默认为 id
+        sort_column = getattr(ErrorLog, sort_by, ErrorLog.id)
         if sort_order.lower() == 'asc':
             query = query.order_by(asc(sort_column))
         else:
             query = query.order_by(desc(sort_column))
 
-        # Apply limit and offset
         query = query.limit(limit).offset(offset)
 
         result = await database.fetch_all(query)
         return [dict(row) for row in result]
     except Exception as e:
-        logger.exception(f"Failed to get error logs with filters: {str(e)}") # Use exception for stack trace
+        logger.exception(f"Failed to get error logs with filters: {str(e)}")
         raise
 
 
 async def get_error_logs_count(
     key_search: Optional[str] = None,
     error_search: Optional[str] = None,
-    error_code_search: Optional[str] = None, # Added error code search
+    error_code_search: Optional[str] = None,
     start_date: Optional[datetime] = None,
     end_date: Optional[datetime] = None
 ) -> int:
@@ -254,7 +245,6 @@ async def get_error_logs_count(
     try:
         query = select(func.count()).select_from(ErrorLog)
 
-        # Apply the same filters as get_error_logs
         if key_search:
             query = query.where(ErrorLog.gemini_key.ilike(f"%{key_search}%"))
         if error_search:
@@ -265,23 +255,19 @@ async def get_error_logs_count(
         if start_date:
             query = query.where(ErrorLog.request_time >= start_date)
         if end_date:
-            # Use the datetime object directly for comparison
             query = query.where(ErrorLog.request_time < end_date)
         if error_code_search:
             try:
-                # Attempt to convert search string to integer for exact match
                 error_code_int = int(error_code_search)
                 query = query.where(ErrorLog.error_code == error_code_int)
             except ValueError:
-                # If conversion fails, log a warning and potentially skip this filter
                 logger.warning(f"Invalid format for error_code_search in count: '{error_code_search}'. Expected an integer. Skipping error code filter.")
-                # Optionally, force count to 0 if the format is invalid:
-                # return 0 # Or query = query.where(False) before fetching
+
 
         count_result = await database.fetch_one(query)
         return count_result[0] if count_result else 0
     except Exception as e:
-        logger.exception(f"Failed to count error logs with filters: {str(e)}") # Use exception for stack trace
+        logger.exception(f"Failed to count error logs with filters: {str(e)}")
         raise
 
 
@@ -307,7 +293,7 @@ async def get_error_log_details(log_id: int) -> Optional[Dict[str, Any]]:
                 try:
                     log_dict['request_msg'] = json.dumps(log_dict['request_msg'], ensure_ascii=False, indent=2)
                 except TypeError:
-                    log_dict['request_msg'] = str(log_dict['request_msg']) # Fallback to string
+                    log_dict['request_msg'] = str(log_dict['request_msg'])
             return log_dict
         else:
             return None
@@ -315,7 +301,6 @@ async def get_error_log_details(log_id: int) -> Optional[Dict[str, Any]]:
         logger.exception(f"Failed to get error log details for ID {log_id}: {str(e)}")
         raise
 
-# --- 异步删除函数 (使用 databases 库) ---
 
 async def delete_error_logs_by_ids(log_ids: List[int]) -> int:
     """
@@ -345,7 +330,7 @@ async def delete_error_logs_by_ids(log_ids: List[int]) -> int:
     except Exception as e:
         # 数据库连接或执行错误
         logger.error(f"Error during bulk deletion of error logs {log_ids}: {e}", exc_info=True)
-        raise # Re-raise the exception for the router to handle
+        raise
 
 async def delete_error_log_by_id(log_id: int) -> bool:
     """
@@ -364,7 +349,7 @@ async def delete_error_log_by_id(log_id: int) -> bool:
 
         if not exists:
             logger.warning(f"Attempted to delete non-existent error log with ID: {log_id}")
-            return False # 或者可以抛出 404 异常，由路由处理
+            return False
 
         # 执行删除
         delete_query = delete(ErrorLog).where(ErrorLog.id == log_id)
@@ -373,10 +358,36 @@ async def delete_error_log_by_id(log_id: int) -> bool:
         return True
     except Exception as e:
         logger.error(f"Error deleting error log with ID {log_id}: {e}", exc_info=True)
-        raise # Re-raise the exception for the router to handle
-
-# --- RequestLog Services (保持异步) ---
-
+        raise
+ 
+ 
+async def delete_all_error_logs() -> int:
+    """
+    删除所有错误日志条目。
+ 
+    Returns:
+        int: 被删除的错误日志数量。
+    """
+    try:
+        # 1. 获取删除前的总数
+        count_query = select(func.count()).select_from(ErrorLog)
+        total_to_delete = await database.fetch_val(count_query)
+ 
+        if total_to_delete == 0:
+            logger.info("No error logs found to delete.")
+            return 0
+ 
+        # 2. 执行删除操作
+        delete_query = delete(ErrorLog)
+        await database.execute(delete_query)
+        
+        logger.info(f"Successfully deleted all {total_to_delete} error logs.")
+        return total_to_delete
+    except Exception as e:
+        logger.error(f"Failed to delete all error logs: {str(e)}", exc_info=True)
+        raise
+ 
+ 
 # 新增函数：添加请求日志
 async def add_request_log(
     model_name: Optional[str],
@@ -412,7 +423,6 @@ async def add_request_log(
             latency_ms=latency_ms
         )
         await database.execute(query)
-        # logger.debug(f"Added request log: key={api_key[:4]}..., success={is_success}, model={model_name}") # Use debug level
         return True
     except Exception as e:
         logger.error(f"Failed to add request log: {str(e)}")

app/domain/gemini_models.py

@@ -40,15 +40,16 @@ class GenerationConfig(BaseModel):
     frequencyPenalty: Optional[float] = None
     responseLogprobs: Optional[bool] = None
     logprobs: Optional[int] = None
+    thinkingConfig: Optional[Dict[str, Any]] = None
 
 
 class SystemInstruction(BaseModel):
-    role: str = "system"
-    parts: List[Dict[str, Any]] | Dict[str, Any]
+    role: Optional[str] = "system"
+    parts: Union[List[Dict[str, Any]], Dict[str, Any]]
 
 
 class GeminiContent(BaseModel):
-    role: str
+    role: Optional[str] = None
     parts: List[Dict[str, Any]]
 
 

app/domain/image_models.py

@@ -1,23 +1,20 @@
+from typing import Union
+
+
 class ImageMetadata:
-    def __init__(self, width: int, height: int, filename: str, size: int, url: str, delete_url: str | None = None):
+    def __init__(self, width: int, height: int, filename: str, size: int, url: str, delete_url: Union[str, None] = None):
         self.width = width
         self.height = height
         self.filename = filename
         self.size = size
         self.url = url
         self.delete_url = delete_url
-    
-    
 class UploadResponse:
     def __init__(self, success: bool, code: str, message: str, data: ImageMetadata):
         self.success = success
         self.code = code
         self.message = message
         self.data = data
-    
-    
 class ImageUploader:
     def upload(self, file: bytes, filename: str) -> UploadResponse:
         raise NotImplementedError
-    
-    

app/domain/openai_models.py

@@ -32,4 +32,11 @@ class ImageGenerationRequest(BaseModel):
     size: Optional[str] = "1024x1024"
     quality: Optional[str] = None
     style: Optional[str] = None
-    response_format: Optional[str] = "b64_json"
+    response_format: Optional[str] = "url"
+
+
+class TTSRequest(BaseModel):
+    model: str = "gemini-2.5-flash-preview-tts"
+    input: str
+    voice: str = "Kore"
+    response_format: Optional[str] = "wav"

app/handler/message_converter.py

@@ -128,12 +128,7 @@ class OpenAIMessageConverter(MessageConverter):
             raise ValueError(f"Unsupported media format: {format}")
 
         try:
-            # Decode Base64 to check size
-            # Be careful with memory usage for very large files
-            # Consider streaming decoding or checking length heuristic first if memory is a concern
-            decoded_data = base64.b64decode(
-                data, validate=True
-            )  # Use validate=True for stricter check
+            decoded_data = base64.b64decode(data, validate=True)
             if len(decoded_data) > max_size:
                 logger.error(
                     f"Media data size ({len(decoded_data)} bytes) exceeds limit ({max_size} bytes)."
@@ -141,7 +136,6 @@ class OpenAIMessageConverter(MessageConverter):
                 raise ValueError(
                     f"Media data size exceeds limit of {max_size // 1024 // 1024}MB"
                 )
-            # No need to return decoded_data, just the original base64 if valid
             return data
         except base64.binascii.Error as e:
             logger.error(f"Invalid Base64 data provided: {e}")
@@ -163,7 +157,6 @@ class OpenAIMessageConverter(MessageConverter):
             if "content" in msg and isinstance(msg["content"], list):
                 for content_item in msg["content"]:
                     if not isinstance(content_item, dict):
-                        # Skip non-dict items if any unexpected format appears
                         logger.warning(
                             f"Skipping unexpected content item format: {type(content_item)}"
                         )
@@ -184,13 +177,11 @@ class OpenAIMessageConverter(MessageConverter):
                             logger.error(
                                 f"Failed to convert image URL {content_item['image_url']['url']}: {e}"
                             )
-                            # Decide how to handle: skip part, add error text, etc.
                             parts.append(
                                 {
                                     "text": f"[Error processing image: {content_item['image_url']['url']}]"
                                 }
                             )
-                    # --- Add handling for input_audio ---
                     elif content_type == "input_audio" and content_item.get(
                         "input_audio"
                     ):
@@ -205,7 +196,6 @@ class OpenAIMessageConverter(MessageConverter):
                             continue
 
                         try:
-                            # Validate size and format
                             validated_data = self._validate_media_data(
                                 audio_format,
                                 audio_data,

app/handler/response_handler.py

@@ -29,7 +29,7 @@ class GeminiResponseHandler(ResponseHandler):
         self.thinking_status = False
 
     def handle_response(
-        self, response: Dict[str, Any], model: str, stream: bool = False
+        self, response: Dict[str, Any], model: str, stream: bool = False, usage_metadata: Optional[Dict[str, Any]] = None
     ) -> Dict[str, Any]:
         if stream:
             return _handle_gemini_stream_response(response, model, stream)
@@ -37,31 +37,33 @@ class GeminiResponseHandler(ResponseHandler):
 
 
 def _handle_openai_stream_response(
-    response: Dict[str, Any], model: str, finish_reason: str
+    response: Dict[str, Any], model: str, finish_reason: str, usage_metadata: Optional[Dict[str, Any]]
 ) -> Dict[str, Any]:
-    text, tool_calls = _extract_result(
+    text, reasoning_content, tool_calls, _ = _extract_result(
         response, model, stream=True, gemini_format=False
     )
-    if not text and not tool_calls:
+    if not text and not tool_calls and not reasoning_content:
         delta = {}
     else:
-        delta = {"content": text, "role": "assistant"}
+        delta = {"content": text, "reasoning_content": reasoning_content, "role": "assistant"}
         if tool_calls:
             delta["tool_calls"] = tool_calls
-
-    return {
+    template_chunk = {
         "id": f"chatcmpl-{uuid.uuid4()}",
         "object": "chat.completion.chunk",
         "created": int(time.time()),
         "model": model,
         "choices": [{"index": 0, "delta": delta, "finish_reason": finish_reason}],
     }
+    if usage_metadata:
+        template_chunk["usage"] = {"prompt_tokens": usage_metadata.get("promptTokenCount", 0), "completion_tokens": usage_metadata.get("candidatesTokenCount",0), "total_tokens": usage_metadata.get("totalTokenCount", 0)}
+    return template_chunk
 
 
 def _handle_openai_normal_response(
-    response: Dict[str, Any], model: str, finish_reason: str
+    response: Dict[str, Any], model: str, finish_reason: str, usage_metadata: Optional[Dict[str, Any]]
 ) -> Dict[str, Any]:
-    text, tool_calls = _extract_result(
+    text, reasoning_content, tool_calls, _ = _extract_result(
         response, model, stream=False, gemini_format=False
     )
     return {
@@ -75,12 +77,13 @@ def _handle_openai_normal_response(
                 "message": {
                     "role": "assistant",
                     "content": text,
+                    "reasoning_content": reasoning_content,
                     "tool_calls": tool_calls,
                 },
                 "finish_reason": finish_reason,
             }
         ],
-        "usage": {"prompt_tokens": 0, "completion_tokens": 0, "total_tokens": 0},
+        "usage": {"prompt_tokens": usage_metadata.get("promptTokenCount", 0), "completion_tokens": usage_metadata.get("candidatesTokenCount",0), "total_tokens": usage_metadata.get("totalTokenCount", 0)},
     }
 
 
@@ -98,10 +101,11 @@ class OpenAIResponseHandler(ResponseHandler):
         model: str,
         stream: bool = False,
         finish_reason: str = None,
+        usage_metadata: Optional[Dict[str, Any]] = None,
     ) -> Optional[Dict[str, Any]]:
         if stream:
-            return _handle_openai_stream_response(response, model, finish_reason)
-        return _handle_openai_normal_response(response, model, finish_reason)
+            return _handle_openai_stream_response(response, model, finish_reason, usage_metadata)
+        return _handle_openai_normal_response(response, model, finish_reason, usage_metadata)
 
     def handle_image_chat_response(
         self, image_str: str, model: str, stream=False, finish_reason="stop"
@@ -153,17 +157,22 @@ def _extract_result(
     model: str,
     stream: bool = False,
     gemini_format: bool = False,
-) -> tuple[str, List[Dict[str, Any]]]:
-    text, tool_calls = "", []
+) -> tuple[str, Optional[str], List[Dict[str, Any]], Optional[bool]]:
+    text, reasoning_content, tool_calls, thought = "", "", [], None
     if stream:
         if response.get("candidates"):
             candidate = response["candidates"][0]
             content = candidate.get("content", {})
             parts = content.get("parts", [])
             if not parts:
-                return "", []
+                return "", None, [], None
             if "text" in parts[0]:
                 text = parts[0].get("text")
+                if "thought" in parts[0]:
+                    if not gemini_format and settings.SHOW_THINKING_PROCESS:
+                        reasoning_content = text
+                        text = ""
+                    thought = parts[0].get("thought")
             elif "executableCode" in parts[0]:
                 text = _format_code_block(parts[0]["executableCode"])
             elif "codeExecution" in parts[0]:
@@ -181,30 +190,18 @@ def _extract_result(
     else:
         if response.get("candidates"):
             candidate = response["candidates"][0]
-            if "thinking" in model:
-                if settings.SHOW_THINKING_PROCESS:
-                    if len(candidate["content"]["parts"]) == 2:
-                        text = (
-                            "> thinking\n\n"
-                            + candidate["content"]["parts"][0]["text"]
-                            + "\n\n---\n> output\n\n"
-                            + candidate["content"]["parts"][1]["text"]
-                        )
-                    else:
-                        text = candidate["content"]["parts"][0]["text"]
-                else:
-                    if len(candidate["content"]["parts"]) == 2:
-                        text = candidate["content"]["parts"][1]["text"]
-                    else:
-                        text = candidate["content"]["parts"][0]["text"]
-            else:
-                text = ""
-                if "parts" in candidate["content"]:
-                    for part in candidate["content"]["parts"]:
-                        if "text" in part:
+            text, reasoning_content = "", ""
+            if "parts" in candidate["content"]:
+                for part in candidate["content"]["parts"]:
+                    if "text" in part:
+                        if "thought" in part and settings.SHOW_THINKING_PROCESS:
+                            reasoning_content += part["text"]
+                        else:
                             text += part["text"]
-                        elif "inlineData" in part:
-                            text += _extract_image_data(part)
+                        if "thought" in part and thought is None:
+                            thought = part.get("thought")
+                    elif "inlineData" in part:
+                        text += _extract_image_data(part)
 
             text = _add_search_link_text(model, candidate, text)
             tool_calls = _extract_tool_calls(
@@ -212,7 +209,7 @@ def _extract_result(
             )
         else:
             text = "暂无返回"
-    return text, tool_calls
+    return text, reasoning_content, tool_calls, thought
 
 
 def _extract_image_data(part: dict) -> str:
@@ -230,6 +227,7 @@ def _extract_image_data(part: dict) -> str:
             provider=settings.UPLOAD_PROVIDER,
             base_url=settings.CLOUDFLARE_IMGBED_URL,
             auth_code=settings.CLOUDFLARE_IMGBED_AUTH_CODE,
+            upload_folder=settings.CLOUDFLARE_IMGBED_UPLOAD_FOLDER,
         )
     current_date = time.strftime("%Y/%m/%d")
     filename = f"{current_date}/{uuid.uuid4().hex[:8]}.png"
@@ -285,13 +283,16 @@ def _extract_tool_calls(
 def _handle_gemini_stream_response(
     response: Dict[str, Any], model: str, stream: bool
 ) -> Dict[str, Any]:
-    text, tool_calls = _extract_result(
+    text, reasoning_content, tool_calls, thought = _extract_result(
         response, model, stream=stream, gemini_format=True
     )
     if tool_calls:
         content = {"parts": tool_calls, "role": "model"}
     else:
-        content = {"parts": [{"text": text}], "role": "model"}
+        part = {"text": text}
+        if thought is not None:
+            part["thought"] = thought
+        content = {"parts": [part], "role": "model"}
     response["candidates"][0]["content"] = content
     return response
 
@@ -299,13 +300,18 @@ def _handle_gemini_stream_response(
 def _handle_gemini_normal_response(
     response: Dict[str, Any], model: str, stream: bool
 ) -> Dict[str, Any]:
-    text, tool_calls = _extract_result(
+    text, reasoning_content, tool_calls, thought = _extract_result(
         response, model, stream=stream, gemini_format=True
     )
+    parts = []
     if tool_calls:
-        content = {"parts": tool_calls, "role": "model"}
+        parts = tool_calls
     else:
-        content = {"parts": [{"text": text}], "role": "model"}
+        if thought is not None:
+            parts.append({"text": reasoning_content,"thought": thought})
+        part = {"text": text}
+        parts.append(part)
+    content = {"parts": parts, "role": "model"}
     response["candidates"][0]["content"] = content
     return response
 

app/handler/retry_handler.py

@@ -2,7 +2,7 @@
 from functools import wraps
 from typing import Callable, TypeVar
 
-from app.core.constants import MAX_RETRIES
+from app.config.config import settings
 from app.log.logger import get_retry_logger
 
 T = TypeVar("T")
@@ -12,8 +12,7 @@ logger = get_retry_logger()
 class RetryHandler:
     """重试处理装饰器"""
 
-    def __init__(self, max_retries: int = MAX_RETRIES, key_arg: str = "api_key"):
-        self.max_retries = max_retries
+    def __init__(self, key_arg: str = "api_key"):
         self.key_arg = key_arg
 
     def __call__(self, func: Callable[..., T]) -> Callable[..., T]:
@@ -21,14 +20,14 @@ class RetryHandler:
         async def wrapper(*args, **kwargs) -> T:
             last_exception = None
 
-            for attempt in range(self.max_retries):
+            for attempt in range(settings.MAX_RETRIES):
                 retries = attempt + 1
                 try:
                     return await func(*args, **kwargs)
                 except Exception as e:
                     last_exception = e
                     logger.warning(
-                        f"API call failed with error: {str(e)}. Attempt {retries} of {self.max_retries}"
+                        f"API call failed with error: {str(e)}. Attempt {retries} of {settings.MAX_RETRIES}"
                     )
 
                     # 从函数参数中获取 key_manager

app/log/logger.py

@@ -217,4 +217,17 @@ def get_api_client_logger():
 
 
 def get_openai_compatible_logger():
-    return Logger.setup_logger("openai_compatible")
+    return Logger.setup_logger("openai_compatible")
+
+
+def get_error_log_logger():
+    return Logger.setup_logger("error_log")
+
+
+def get_request_log_logger():
+    return Logger.setup_logger("request_log")
+
+
+def get_vertex_express_logger():
+    return Logger.setup_logger("vertex_express")
+

app/main.py

@@ -1,4 +1,8 @@
 import uvicorn
+from dotenv import load_dotenv
+
+# 在导入应用程序配置之前加载 .env 文件到环境变量
+load_dotenv()
 
 from app.core.application import create_app
 from app.log.logger import get_main_logger

app/middleware/middleware.py

@@ -8,6 +8,7 @@ from fastapi.responses import RedirectResponse
 from starlette.middleware.base import BaseHTTPMiddleware
 
 # from app.middleware.request_logging_middleware import RequestLoggingMiddleware
+from app.middleware.smart_routing_middleware import SmartRoutingMiddleware
 from app.core.constants import API_VERSION
 from app.core.security import verify_auth_token
 from app.log.logger import get_middleware_logger
@@ -32,6 +33,7 @@ class AuthMiddleware(BaseHTTPMiddleware):
             and not request.url.path.startswith("/hf")
             and not request.url.path.startswith("/openai")
             and not request.url.path.startswith("/api/version/check")
+            and not request.url.path.startswith("/vertex-express")
         ):
 
             auth_token = request.cookies.get("auth_token")
@@ -51,6 +53,9 @@ def setup_middlewares(app: FastAPI) -> None:
     Args:
         app: FastAPI应用程序实例
     """
+    # 添加智能路由中间件（必须在认证中间件之前）
+    app.add_middleware(SmartRoutingMiddleware)
+
     # 添加认证中间件
     app.add_middleware(AuthMiddleware)
 
@@ -60,7 +65,7 @@ def setup_middlewares(app: FastAPI) -> None:
     # 配置CORS中间件
     app.add_middleware(
         CORSMiddleware,
-        allow_origins=["*"],  # 生产环境建议配置具体的域名
+        allow_origins=["*"],
         allow_credentials=True,
         allow_methods=[
             "GET",
@@ -68,8 +73,8 @@ def setup_middlewares(app: FastAPI) -> None:
             "PUT",
             "DELETE",
             "OPTIONS",
-        ],  # 明确指定允许的HTTP方法
-        allow_headers=["*"],  # 生产环境建议配置具体的请求头
-        expose_headers=["*"],  # 允许前端访问的响应头
-        max_age=600,  # 预检请求缓存时间(秒)
+        ],
+        allow_headers=["*"],
+        expose_headers=["*"],
+        max_age=600,
     )

app/middleware/smart_routing_middleware.py

@@ -0,0 +1,210 @@
+from fastapi import Request
+from starlette.middleware.base import BaseHTTPMiddleware
+from app.config.config import settings
+from app.log.logger import get_main_logger
+import re
+
+logger = get_main_logger()
+
+class SmartRoutingMiddleware(BaseHTTPMiddleware):
+    def __init__(self, app):
+        super().__init__(app)
+        # 简化的路由规则 - 直接根据检测结果路由
+        pass
+
+    async def dispatch(self, request: Request, call_next):
+        if not settings.URL_NORMALIZATION_ENABLED:
+            return await call_next(request)
+        logger.debug(f"request: {request}")
+        original_path = str(request.url.path)
+        method = request.method
+        
+        # 尝试修复URL
+        fixed_path, fix_info = self.fix_request_url(original_path, method, request)
+
+        if fixed_path != original_path:
+            logger.info(f"URL fixed: {method} {original_path} → {fixed_path}")
+            if fix_info:
+                logger.debug(f"Fix details: {fix_info}")
+
+            # 重写请求路径
+            request.scope["path"] = fixed_path
+            request.scope["raw_path"] = fixed_path.encode()
+        
+        return await call_next(request)
+
+    def fix_request_url(self, path: str, method: str, request: Request) -> tuple:
+        """简化的URL修复逻辑"""
+
+        # 首先检查是否已经是正确的格式，如果是则不处理
+        if self.is_already_correct_format(path):
+            return path, None
+
+        # 1. 最高优先级：包含generateContent → Gemini格式
+        if "generatecontent" in path.lower() or "v1beta/models" in path.lower():
+            return self.fix_gemini_by_operation(path, method, request)
+
+        # 2. 第二优先级：包含/openai/ → OpenAI格式
+        if "/openai/" in path.lower():
+            return self.fix_openai_by_operation(path, method)
+
+        # 3. 第三优先级：包含/v1/ → v1格式
+        if "/v1/" in path.lower():
+            return self.fix_v1_by_operation(path, method)
+
+        # 4. 第四优先级：包含/chat/completions → chat功能
+        if "/chat/completions" in path.lower():
+            return "/v1/chat/completions", {"type": "v1_chat"}
+
+        # 5. 默认：原样传递
+        return path, None
+
+    def is_already_correct_format(self, path: str) -> bool:
+        """检查是否已经是正确的API格式"""
+        # 检查是否已经是正确的端点格式
+        correct_patterns = [
+            r"^/v1beta/models/[^/:]+:(generate|streamGenerate)Content$",  # Gemini原生
+            r"^/gemini/v1beta/models/[^/:]+:(generate|streamGenerate)Content$",  # Gemini带前缀
+            r"^/v1beta/models$",  # Gemini模型列表
+            r"^/gemini/v1beta/models$",  # Gemini带前缀的模型列表
+            r"^/v1/(chat/completions|models|embeddings|images/generations|audio/speech)$",  # v1格式
+            r"^/openai/v1/(chat/completions|models|embeddings|images/generations|audio/speech)$",  # OpenAI格式
+            r"^/hf/v1/(chat/completions|models|embeddings|images/generations|audio/speech)$",  # HF格式
+            r"^/vertex-express/v1beta/models/[^/:]+:(generate|streamGenerate)Content$",  # Vertex Express Gemini格式
+            r"^/vertex-express/v1beta/models$",  # Vertex Express模型列表
+            r"^/vertex-express/v1/(chat/completions|models|embeddings|images/generations)$",  # Vertex Express OpenAI格式
+        ]
+
+        for pattern in correct_patterns:
+            if re.match(pattern, path):
+                return True
+
+        return False
+
+    def fix_gemini_by_operation(
+        self, path: str, method: str, request: Request
+    ) -> tuple:
+        """根据Gemini操作修复，考虑端点偏好"""
+        if method == "GET":
+            return "/v1beta/models", {
+                "role": "gemini_models",
+            }
+
+        # 提取模型名称
+        try:
+            model_name = self.extract_model_name(path, request)
+        except ValueError:
+            # 无法提取模型名称，返回原路径不做处理
+            return path, None
+
+        # 检测是否为流式请求
+        is_stream = self.detect_stream_request(path, request)
+
+        # 检查是否有vertex-express偏好
+        if "/vertex-express/" in path.lower():
+            if is_stream:
+                target_url = (
+                    f"/vertex-express/v1beta/models/{model_name}:streamGenerateContent"
+                )
+            else:
+                target_url = (
+                    f"/vertex-express/v1beta/models/{model_name}:generateContent"
+                )
+
+            fix_info = {
+                "rule": (
+                    "vertex_express_generate"
+                    if not is_stream
+                    else "vertex_express_stream"
+                ),
+                "preference": "vertex_express_format",
+                "is_stream": is_stream,
+                "model": model_name,
+            }
+        else:
+            # 标准Gemini端点
+            if is_stream:
+                target_url = f"/v1beta/models/{model_name}:streamGenerateContent"
+            else:
+                target_url = f"/v1beta/models/{model_name}:generateContent"
+
+            fix_info = {
+                "rule": "gemini_generate" if not is_stream else "gemini_stream",
+                "preference": "gemini_format",
+                "is_stream": is_stream,
+                "model": model_name,
+            }
+
+        return target_url, fix_info
+
+    def fix_openai_by_operation(self, path: str, method: str) -> tuple:
+        """根据操作类型修复OpenAI格式"""
+        if method == "POST":
+            if "chat" in path.lower() or "completion" in path.lower():
+                return "/openai/v1/chat/completions", {"type": "openai_chat"}
+            elif "embedding" in path.lower():
+                return "/openai/v1/embeddings", {"type": "openai_embeddings"}
+            elif "image" in path.lower():
+                return "/openai/v1/images/generations", {"type": "openai_images"}
+            elif "audio" in path.lower():
+                return "/openai/v1/audio/speech", {"type": "openai_audio"}
+        elif method == "GET":
+            if "model" in path.lower():
+                return "/openai/v1/models", {"type": "openai_models"}
+
+        return path, None
+
+    def fix_v1_by_operation(self, path: str, method: str) -> tuple:
+        """根据操作类型修复v1格式"""
+        if method == "POST":
+            if "chat" in path.lower() or "completion" in path.lower():
+                return "/v1/chat/completions", {"type": "v1_chat"}
+            elif "embedding" in path.lower():
+                return "/v1/embeddings", {"type": "v1_embeddings"}
+            elif "image" in path.lower():
+                return "/v1/images/generations", {"type": "v1_images"}
+            elif "audio" in path.lower():
+                return "/v1/audio/speech", {"type": "v1_audio"}
+        elif method == "GET":
+            if "model" in path.lower():
+                return "/v1/models", {"type": "v1_models"}
+
+        return path, None
+
+    def detect_stream_request(self, path: str, request: Request) -> bool:
+        """检测是否为流式请求"""
+        # 1. 路径中包含stream关键词
+        if "stream" in path.lower():
+            return True
+
+        # 2. 查询参数
+        if request.query_params.get("stream") == "true":
+            return True
+
+        return False
+
+    def extract_model_name(self, path: str, request: Request) -> str:
+        """从请求中提取模型名称，用于构建Gemini API URL"""
+        # 1. 从请求体中提取
+        try:
+            if hasattr(request, "_body") and request._body:
+                import json
+
+                body = json.loads(request._body.decode())
+                if "model" in body and body["model"]:
+                    return body["model"]
+        except Exception:
+            pass
+
+        # 2. 从查询参数中提取
+        model_param = request.query_params.get("model")
+        if model_param:
+            return model_param
+
+        # 3. 从路径中提取（用于已包含模型名称的路径）
+        match = re.search(r"/models/([^/:]+)", path, re.IGNORECASE)
+        if match:
+            return match.group(1)
+
+        # 4. 如果无法提取模型名称，抛出异常
+        raise ValueError("Unable to extract model name from request")

app/router/config_routes.py

@@ -1,15 +1,17 @@
 """
 配置路由模块
 """
-from typing import Any, Dict
+
+from typing import Any, Dict, List
+
 from fastapi import APIRouter, HTTPException, Request
 from fastapi.responses import RedirectResponse
+from pydantic import BaseModel, Field
 
 from app.core.security import verify_auth_token
-from app.log.logger import get_config_routes_logger, Logger # 导入 Logger 类
+from app.log.logger import Logger, get_config_routes_logger
 from app.service.config.config_service import ConfigService
 
-# 创建路由
 router = APIRouter(prefix="/api/config", tags=["config"])
 
 logger = get_config_routes_logger()
@@ -34,10 +36,10 @@ async def update_config(config_data: Dict[str, Any], request: Request):
         result = await ConfigService.update_config(config_data)
         # 配置更新成功后，立即更新所有 logger 的级别
         Logger.update_log_levels(config_data["LOG_LEVEL"])
-        logger.info("Log levels updated after configuration change.") # 添加日志记录
+        logger.info("Log levels updated after configuration change.")
         return result
     except Exception as e:
-        logger.error(f"Error updating config or log levels: {e}", exc_info=True) # 记录详细错误
+        logger.error(f"Error updating config or log levels: {e}", exc_info=True)
         raise HTTPException(status_code=400, detail=str(e))
 
 
@@ -51,3 +53,81 @@ async def reset_config(request: Request):
         return await ConfigService.reset_config()
     except Exception as e:
         raise HTTPException(status_code=400, detail=str(e))
+
+
+class DeleteKeysRequest(BaseModel):
+    keys: List[str] = Field(..., description="List of API keys to delete")
+
+
+@router.delete("/keys/{key_to_delete}", response_model=Dict[str, Any])
+async def delete_single_key(key_to_delete: str, request: Request):
+    auth_token = request.cookies.get("auth_token")
+    if not auth_token or not verify_auth_token(auth_token):
+        logger.warning(f"Unauthorized attempt to delete key: {key_to_delete}")
+        return RedirectResponse(url="/", status_code=302)
+    try:
+        logger.info(f"Attempting to delete key: {key_to_delete}")
+        result = await ConfigService.delete_key(key_to_delete)
+        if not result.get("success"):
+            raise HTTPException(
+                status_code=(
+                    404 if "not found" in result.get("message", "").lower() else 400
+                ),
+                detail=result.get("message"),
+            )
+        return result
+    except HTTPException as e:
+        raise e
+    except Exception as e:
+        logger.error(f"Error deleting key '{key_to_delete}': {e}", exc_info=True)
+        raise HTTPException(status_code=500, detail=f"Error deleting key: {str(e)}")
+
+
+@router.post("/keys/delete-selected", response_model=Dict[str, Any])
+async def delete_selected_keys_route(
+    delete_request: DeleteKeysRequest, request: Request
+):
+    auth_token = request.cookies.get("auth_token")
+    if not auth_token or not verify_auth_token(auth_token):
+        logger.warning("Unauthorized attempt to bulk delete keys")
+        return RedirectResponse(url="/", status_code=302)
+
+    if not delete_request.keys:
+        logger.warning("Attempt to bulk delete keys with an empty list.")
+        raise HTTPException(status_code=400, detail="No keys provided for deletion.")
+
+    try:
+        logger.info(f"Attempting to bulk delete {len(delete_request.keys)} keys.")
+        result = await ConfigService.delete_selected_keys(delete_request.keys)
+        if not result.get("success") and result.get("deleted_count", 0) == 0:
+            raise HTTPException(
+                status_code=400, detail=result.get("message", "Failed to delete keys.")
+            )
+        return result
+    except HTTPException as e:
+        raise e
+    except Exception as e:
+        logger.error(f"Error bulk deleting keys: {e}", exc_info=True)
+        raise HTTPException(
+            status_code=500, detail=f"Error bulk deleting keys: {str(e)}"
+        )
+
+
+@router.get("/ui/models")
+async def get_ui_models(request: Request):
+    auth_token_cookie = request.cookies.get("auth_token")
+    if not auth_token_cookie or not verify_auth_token(auth_token_cookie):
+        logger.warning("Unauthorized access attempt to /api/config/ui/models")
+        raise HTTPException(status_code=403, detail="Not authenticated")
+
+    try:
+        models = await ConfigService.fetch_ui_models()
+        return models
+    except HTTPException as e:
+        raise e
+    except Exception as e:
+        logger.error(f"Unexpected error in /ui/models endpoint: {e}", exc_info=True)
+        raise HTTPException(
+            status_code=500,
+            detail=f"An unexpected error occurred while fetching UI models: {str(e)}",
+        )

app/router/error_log_routes.py

@@ -1,55 +1,69 @@
 """
 日志路由模块
 """
-from typing import List, Optional, Dict
+
 from datetime import datetime
+from typing import Dict, List, Optional
+
+from fastapi import (
+    APIRouter,
+    Body,
+    HTTPException,
+    Path,
+    Query,
+    Request,
+    Response,
+    status,
+)
 from pydantic import BaseModel
-from fastapi import APIRouter, HTTPException, Request, Query, Path, Body, Response, status
 
 from app.core.security import verify_auth_token
 from app.log.logger import get_log_routes_logger
-# 假设这些服务函数已更新或添加
-from app.database.services import (
-    get_error_logs,
-    get_error_logs_count,
-    get_error_log_details,
-    delete_error_logs_by_ids, # 新增导入
-    delete_error_log_by_id    # 新增导入
-)
-# Removed get_db import comment as it's fully removed now
+from app.service.error_log import error_log_service
 
-# 创建路由
 router = APIRouter(prefix="/api/logs", tags=["logs"])
 
 logger = get_log_routes_logger()
 
 
-# Define a response model that includes the total count for pagination
-# 用于列表响应的模型，假设 get_error_logs 返回包含 error_code 的字典
 class ErrorLogListItem(BaseModel):
     id: int
     gemini_key: Optional[str] = None
     error_type: Optional[str] = None
-    error_code: Optional[int] = None # 列表显示错误码 (应为整数)
+    error_code: Optional[int] = None
     model_name: Optional[str] = None
     request_time: Optional[datetime] = None
 
+
 class ErrorLogListResponse(BaseModel):
-    logs: List[ErrorLogListItem] # 使用定义的模型列表
+    logs: List[ErrorLogListItem]
     total: int
 
+
 @router.get("/errors", response_model=ErrorLogListResponse)
 async def get_error_logs_api(
     request: Request,
     limit: int = Query(10, ge=1, le=1000),
     offset: int = Query(0, ge=0),
-    key_search: Optional[str] = Query(None, description="Search term for Gemini key (partial match)"),
-    error_search: Optional[str] = Query(None, description="Search term for error type or log message"), # 数据库查询需处理
-    error_code_search: Optional[str] = Query(None, description="Search term for error code"), # Added error code search parameter
-    start_date: Optional[datetime] = Query(None, description="Start datetime for filtering"),
-    end_date: Optional[datetime] = Query(None, description="End datetime for filtering"),
-    sort_by: str = Query('id', description="Field to sort by (e.g., 'id', 'request_time')"), # 新增排序参数
-    sort_order: str = Query('desc', description="Sort order ('asc' or 'desc')") # 新增排序参数
+    key_search: Optional[str] = Query(
+        None, description="Search term for Gemini key (partial match)"
+    ),
+    error_search: Optional[str] = Query(
+        None, description="Search term for error type or log message"
+    ),
+    error_code_search: Optional[str] = Query(
+        None, description="Search term for error code"
+    ),
+    start_date: Optional[datetime] = Query(
+        None, description="Start datetime for filtering"
+    ),
+    end_date: Optional[datetime] = Query(
+        None, description="End datetime for filtering"
+    ),
+    sort_by: str = Query(
+        "id", description="Field to sort by (e.g., 'id', 'request_time')"
+    ),
+    sort_order: str = Query("desc", description="Sort order ('asc' or 'desc')"),
 ):
     """
     获取错误日志列表 (返回错误码)，支持过滤和排序
@@ -72,49 +86,42 @@ async def get_error_logs_api(
     auth_token = request.cookies.get("auth_token")
     if not auth_token or not verify_auth_token(auth_token):
         logger.warning("Unauthorized access attempt to error logs list")
-        # API 返回 401 更合适
         raise HTTPException(status_code=401, detail="Not authenticated")
-    
+
     try:
-        # 假设 get_error_logs 现在返回包含 error_code 的字典列表
-        # 并且可以接受 include_error_code 参数 (如果需要显式指定)
-        logs_data = await get_error_logs(
+        result = await error_log_service.process_get_error_logs(
             limit=limit,
             offset=offset,
             key_search=key_search,
-            error_search=error_search, # 数据库查询需要处理这个
-            error_code_search=error_code_search, # Pass error code search to DB function
+            error_search=error_search,
+            error_code_search=error_code_search,
             start_date=start_date,
             end_date=end_date,
-            sort_by=sort_by,       # 传递排序参数
-            sort_order=sort_order  # 传递排序参数
+            sort_by=sort_by,
+            sort_order=sort_order,
         )
-        # Fetch total count with the same search parameters
-        total_count = await get_error_logs_count(
-            key_search=key_search,
-            error_search=error_search,
-            error_code_search=error_code_search, # Pass error code search to DB count function
-            start_date=start_date,
-            end_date=end_date
-        )
-        # 验证并转换数据以匹配 Pydantic 模型
+        logs_data = result["logs"]
+        total_count = result["total"]
+
         validated_logs = [ErrorLogListItem(**log) for log in logs_data]
         return ErrorLogListResponse(logs=validated_logs, total=total_count)
     except Exception as e:
         logger.exception(f"Failed to get error logs list: {str(e)}")
-        raise HTTPException(status_code=500, detail=f"Failed to get error logs list: {str(e)}")
+        raise HTTPException(
+            status_code=500, detail=f"Failed to get error logs list: {str(e)}"
+        )
 
 
-# 新增：获取错误日志详情的路由
 class ErrorLogDetailResponse(BaseModel):
     id: int
     gemini_key: Optional[str] = None
     error_type: Optional[str] = None
-    error_log: Optional[str] = None # 详情接口返回完整的 error_log
-    request_msg: Optional[str] = None # 详情接口返回 request_msg
+    error_log: Optional[str] = None
+    request_msg: Optional[str] = None
     model_name: Optional[str] = None
     request_time: Optional[datetime] = None
 
+
 @router.get("/errors/{log_id}/details", response_model=ErrorLogDetailResponse)
 async def get_error_log_detail_api(request: Request, log_id: int = Path(..., ge=1)):
     """
@@ -122,31 +129,31 @@ async def get_error_log_detail_api(request: Request, log_id: int = Path(..., ge=
     """
     auth_token = request.cookies.get("auth_token")
     if not auth_token or not verify_auth_token(auth_token):
-        logger.warning(f"Unauthorized access attempt to error log details for ID: {log_id}")
+        logger.warning(
+            f"Unauthorized access attempt to error log details for ID: {log_id}"
+        )
         raise HTTPException(status_code=401, detail="Not authenticated")
 
     try:
-        # 假设存在一个函数 get_error_log_details(log_id) 来获取完整信息
-        log_details = await get_error_log_details(log_id=log_id)
+        log_details = await error_log_service.process_get_error_log_details(
+            log_id=log_id
+        )
         if not log_details:
             raise HTTPException(status_code=404, detail="Error log not found")
 
-        # 假设 get_error_log_details 返回一个字典或兼容 Pydantic 的对象
         return ErrorLogDetailResponse(**log_details)
     except HTTPException as http_exc:
-        # Re-raise HTTPException (like 404)
         raise http_exc
     except Exception as e:
         logger.exception(f"Failed to get error log details for ID {log_id}: {str(e)}")
-        raise HTTPException(status_code=500, detail=f"Failed to get error log details: {str(e)}")
+        raise HTTPException(
+            status_code=500, detail=f"Failed to get error log details: {str(e)}"
+        )
 
 
-# 新增：批量删除错误日志
 @router.delete("/errors", status_code=status.HTTP_204_NO_CONTENT)
 async def delete_error_logs_bulk_api(
-    request: Request,
-    payload: Dict[str, List[int]] = Body(...) # Expects {"ids": [1, 2, 3]}
-    # Ensure db dependency is fully removed
+    request: Request, payload: Dict[str, List[int]] = Body(...)
 ):
     """
     批量删除错误日志 (异步)
@@ -161,23 +168,45 @@ async def delete_error_logs_bulk_api(
         raise HTTPException(status_code=400, detail="No log IDs provided for deletion.")
 
     try:
-        # 调用异步服务函数
-        deleted_count = await delete_error_logs_by_ids(log_ids)
+        deleted_count = await error_log_service.process_delete_error_logs_by_ids(
+            log_ids
+        )
         # 注意：异步函数返回的是尝试删除的数量，可能不是精确值
-        logger.info(f"Attempted bulk deletion for {deleted_count} error logs with IDs: {log_ids}")
+        logger.info(
+            f"Attempted bulk deletion for {deleted_count} error logs with IDs: {log_ids}"
+        )
         return Response(status_code=status.HTTP_204_NO_CONTENT)
     except Exception as e:
         logger.exception(f"Error bulk deleting error logs with IDs {log_ids}: {str(e)}")
-        raise HTTPException(status_code=500, detail="Internal server error during bulk deletion")
+        raise HTTPException(
+            status_code=500, detail="Internal server error during bulk deletion"
+        )
 
 
-# 新增：删除单个错误日志
+@router.delete("/errors/all", status_code=status.HTTP_204_NO_CONTENT)
+async def delete_all_error_logs_api(request: Request):
+    """
+    删除所有错误日志 (异步)
+    """
+    auth_token = request.cookies.get("auth_token")
+    if not auth_token or not verify_auth_token(auth_token):
+        logger.warning("Unauthorized access attempt to delete all error logs")
+        raise HTTPException(status_code=401, detail="Not authenticated")
+ 
+    try:
+        deleted_count = await error_log_service.process_delete_all_error_logs()
+        logger.info(f"Successfully deleted all {deleted_count} error logs.")
+        # No body needed for 204 response
+        return Response(status_code=status.HTTP_204_NO_CONTENT)
+    except Exception as e:
+        logger.exception(f"Error deleting all error logs: {str(e)}")
+        raise HTTPException(
+            status_code=500, detail="Internal server error during deletion of all logs"
+        )
+ 
+ 
 @router.delete("/errors/{log_id}", status_code=status.HTTP_204_NO_CONTENT)
-async def delete_error_log_api(
-    request: Request,
-    log_id: int = Path(..., ge=1)
-    # Ensure db dependency is fully removed
-):
+async def delete_error_log_api(request: Request, log_id: int = Path(..., ge=1)):
     """
     删除单个错误日志 (异步)
     """
@@ -185,17 +214,20 @@ async def delete_error_log_api(
     if not auth_token or not verify_auth_token(auth_token):
         logger.warning(f"Unauthorized access attempt to delete error log ID: {log_id}")
         raise HTTPException(status_code=401, detail="Not authenticated")
-
+ 
     try:
-        # 调用异步服务函数
-        success = await delete_error_log_by_id(log_id)
+        success = await error_log_service.process_delete_error_log_by_id(log_id)
         if not success:
             # 服务层现在在未找到时返回 False，我们在这里转换为 404
-            raise HTTPException(status_code=404, detail=f"Error log with ID {log_id} not found")
+            raise HTTPException(
+                status_code=404, detail=f"Error log with ID {log_id} not found"
+            )
         logger.info(f"Successfully deleted error log with ID: {log_id}")
         return Response(status_code=status.HTTP_204_NO_CONTENT)
     except HTTPException as http_exc:
-        raise http_exc # Re-raise 404 or other HTTP exceptions
+        raise http_exc
     except Exception as e:
         logger.exception(f"Error deleting error log with ID {log_id}: {str(e)}")
-        raise HTTPException(status_code=500, detail="Internal server error during deletion")
+        raise HTTPException(
+            status_code=500, detail="Internal server error during deletion"
+        )

app/router/gemini_routes.py

@@ -5,7 +5,7 @@ import asyncio
 from app.config.config import settings
 from app.log.logger import get_gemini_logger
 from app.core.security import SecurityService
-from app.domain.gemini_models import GeminiContent, GeminiRequest, ResetSelectedKeysRequest, VerifySelectedKeysRequest # 添加导入
+from app.domain.gemini_models import GeminiContent, GeminiRequest, ResetSelectedKeysRequest, VerifySelectedKeysRequest
 from app.service.chat.gemini_chat_service import GeminiChatService
 from app.service.key.key_manager import KeyManager, get_key_manager_instance
 from app.service.model.model_service import ModelService
@@ -13,12 +13,10 @@ from app.handler.retry_handler import RetryHandler
 from app.handler.error_handler import handle_route_errors
 from app.core.constants import API_VERSION
 
-# 路由设置
 router = APIRouter(prefix=f"/gemini/{API_VERSION}")
 router_v1beta = APIRouter(prefix=f"/{API_VERSION}")
 logger = get_gemini_logger()
 
-# 初始化服务
 security_service = SecurityService()
 model_service = ModelService()
 
@@ -52,14 +50,14 @@ async def list_models(
     try:
         api_key = await key_manager.get_first_valid_key()
         if not api_key:
-             raise HTTPException(status_code=503, detail="No valid API keys available to fetch models.")
+            raise HTTPException(status_code=503, detail="No valid API keys available to fetch models.")
         logger.info(f"Using API key: {api_key}")
 
-        models_data =await model_service.get_gemini_models(api_key)
+        models_data = await model_service.get_gemini_models(api_key)
         if not models_data or "models" not in models_data:
-             raise HTTPException(status_code=500, detail="Failed to fetch base models list.")
+            raise HTTPException(status_code=500, detail="Failed to fetch base models list.")
 
-        models_json = deepcopy(models_data) # 操作副本以防修改原始缓存
+        models_json = deepcopy(models_data)
         model_mapping = {x.get("name", "").split("/", maxsplit=1)[-1]: x for x in models_json.get("models", [])}
 
         def add_derived_model(base_name, suffix, display_suffix):
@@ -74,7 +72,6 @@ async def list_models(
             item["description"] = display_name
             models_json["models"].append(item)
 
-        # 添加衍生模型
         if settings.SEARCH_MODELS:
             for name in settings.SEARCH_MODELS:
                 add_derived_model(name, "-search", " For Search")
@@ -88,7 +85,6 @@ async def list_models(
         logger.info("Gemini models list request successful")
         return models_json
     except HTTPException as http_exc:
-        # 重新抛出已知的 HTTP 异常
         raise http_exc
     except Exception as e:
         logger.error(f"Error getting Gemini models list: {str(e)}")
@@ -99,7 +95,7 @@ async def list_models(
 
 @router.post("/models/{model_name}:generateContent")
 @router_v1beta.post("/models/{model_name}:generateContent")
-@RetryHandler(max_retries=settings.MAX_RETRIES, key_arg="api_key")
+@RetryHandler(key_arg="api_key")
 async def generate_content(
     model_name: str,
     request: GeminiRequest,
@@ -128,7 +124,7 @@ async def generate_content(
 
 @router.post("/models/{model_name}:streamGenerateContent")
 @router_v1beta.post("/models/{model_name}:streamGenerateContent")
-@RetryHandler(max_retries=settings.MAX_RETRIES, key_arg="api_key")
+@RetryHandler(key_arg="api_key")
 async def stream_generate_content(
     model_name: str,
     request: GeminiRequest,
@@ -139,7 +135,6 @@ async def stream_generate_content(
 ):
     """处理 Gemini 流式内容生成请求。"""
     operation_name = "gemini_stream_generate_content"
-    # 流式请求的成功/失败日志在流处理中更复杂，这里仅用上下文管理器处理启动错误
     async with handle_route_errors(logger, operation_name, failure_message="Streaming request initiation failed"):
         logger.info(f"Handling Gemini streaming content generation for model: {model_name}")
         logger.debug(f"Request: \n{request.model_dump_json(indent=2)}")
@@ -153,9 +148,38 @@ async def stream_generate_content(
             request=request,
             api_key=api_key
         )
-        # 注意：流本身的错误需要在服务层或流迭代中处理，这里只返回流响应
         return StreamingResponse(response_stream, media_type="text/event-stream")
 
+
+@router.post("/models/{model_name}:countTokens")
+@router_v1beta.post("/models/{model_name}:countTokens")
+@RetryHandler(key_arg="api_key")
+async def count_tokens(
+    model_name: str,
+    request: GeminiRequest,
+    _=Depends(security_service.verify_key_or_goog_api_key),
+    api_key: str = Depends(get_next_working_key),
+    key_manager: KeyManager = Depends(get_key_manager),
+    chat_service: GeminiChatService = Depends(get_chat_service)
+):
+    """处理 Gemini token 计数请求。"""
+    operation_name = "gemini_count_tokens"
+    async with handle_route_errors(logger, operation_name, failure_message="Token counting failed"):
+        logger.info(f"Handling Gemini token count request for model: {model_name}")
+        logger.debug(f"Request: \n{request.model_dump_json(indent=2)}")
+        logger.info(f"Using API key: {api_key}")
+
+        if not await model_service.check_model_support(model_name):
+            raise HTTPException(status_code=400, detail=f"Model {model_name} is not supported")
+
+        response = await chat_service.count_tokens(
+            model=model_name,
+            request=request,
+            api_key=api_key
+        )
+        return response
+
+
 @router.post("/reset-all-fail-counts")
 async def reset_all_key_fail_counts(key_type: str = None, key_manager: KeyManager = Depends(get_key_manager)):
     """批量重置Gemini API密钥的失败计数，可选择性地仅重置有效或无效密钥"""
@@ -203,7 +227,7 @@ async def reset_selected_key_fail_counts(
     """批量重置选定Gemini API密钥的失败计数"""
     logger.info("-" * 50 + "reset_selected_gemini_key_fail_counts" + "-" * 50)
     keys_to_reset = request.keys
-    key_type = request.key_type # 获取类型用于日志记录和响应消息
+    key_type = request.key_type
     logger.info(f"Received reset request for {len(keys_to_reset)} selected {key_type} keys.")
 
     if not keys_to_reset:
@@ -219,38 +243,31 @@ async def reset_selected_key_fail_counts(
                 if result:
                     reset_count += 1
                 else:
-                    # 记录未找到的密钥，但不视为致命错误
                     logger.warning(f"Key not found during selective reset: {key}")
             except Exception as key_error:
-                # 记录单个密钥重置时的错误
                 logger.error(f"Error resetting key {key}: {str(key_error)}")
                 errors.append(f"Key {key}: {str(key_error)}")
 
         if errors:
-             # 如果有错误，报告部分成功或完全失败
              error_message = f"批量重置完成，但出现错误: {'; '.join(errors)}"
-             # 确定最终状态码和成功标志
              final_success = reset_count > 0
-             status_code = 207 if final_success and errors else 500 # 207 Multi-Status if partially successful, 500 if completely failed
+             status_code = 207 if final_success and errors else 500
              return JSONResponse({
                  "success": final_success,
                  "message": error_message,
                  "reset_count": reset_count
              }, status_code=status_code)
 
-        # 完全成功的情况
         return JSONResponse({
             "success": True,
             "message": f"成功重置 {reset_count} 个选定 {key_type} 密钥的失败计数",
             "reset_count": reset_count
         })
     except Exception as e:
-        # 捕获循环外的意外错误
         logger.error(f"Failed to process reset selected key failure counts request: {str(e)}")
         return JSONResponse({"success": False, "message": f"批量重置处理失败: {str(e)}"}, status_code=500)
 
 
-
 @router.post("/reset-fail-count/{api_key}")
 async def reset_key_fail_count(api_key: str, key_manager: KeyManager = Depends(get_key_manager)):
     """重置指定Gemini API密钥的失败计数"""
@@ -266,6 +283,7 @@ async def reset_key_fail_count(api_key: str, key_manager: KeyManager = Depends(g
         logger.error(f"Failed to reset key failure count: {str(e)}")
         return JSONResponse({"success": False, "message": f"重置失败: {str(e)}"}, status_code=500)
 
+
 @router.post("/verify-key/{api_key}")
 async def verify_key(api_key: str, chat_service: GeminiChatService = Depends(get_chat_service), key_manager: KeyManager = Depends(get_key_manager)):
     """验证Gemini API密钥的有效性"""
@@ -273,14 +291,14 @@ async def verify_key(api_key: str, chat_service: GeminiChatService = Depends(get
     logger.info("Verifying API key validity")
     
     try:
-        # 使用generate_content接口测试key的有效性
         gemini_request = GeminiRequest(
             contents=[
                 GeminiContent(
                     role="user",
-                    parts=[{"text": "hi"}]
+                    parts=[{"text": "hi"}],
                 )
-            ]
+            ],
+            generation_config={"temperature": 0.7, "top_p": 1.0, "max_output_tokens": 10}
         )
         
         response = await chat_service.generate_content(
@@ -294,7 +312,6 @@ async def verify_key(api_key: str, chat_service: GeminiChatService = Depends(get
     except Exception as e:
         logger.error(f"Key verification failed: {str(e)}")
         
-        # 验证出现异常时增加失败计数
         async with key_manager.failure_count_lock:
             if api_key in key_manager.key_failure_counts:
                 key_manager.key_failure_counts[api_key] += 1
@@ -318,79 +335,63 @@ async def verify_selected_keys(
         return JSONResponse({"success": False, "message": "没有提供需要验证的密钥"}, status_code=400)
 
     successful_keys = []
-    failed_keys = {} # 存储失败的 key 和错误信息
+    failed_keys = {}
 
     async def _verify_single_key(api_key: str):
         """内部函数，用于验证单个密钥并处理异常"""
-        nonlocal successful_keys, failed_keys # 允许修改外部列表和字典
+        nonlocal successful_keys, failed_keys
         try:
-            # 重用单密钥验证逻辑的核心部分
             gemini_request = GeminiRequest(
-                contents=[GeminiContent(role="user", parts=[{"text": "hi"}])]
+                contents=[GeminiContent(role="user", parts=[{"text": "hi"}])],
+                generation_config={"temperature": 0.7, "top_p": 1.0, "max_output_tokens": 10}
             )
-            # 注意：这里直接调用 chat_service.generate_content，不依赖于 key_manager 获取密钥
             await chat_service.generate_content(
                 settings.TEST_MODEL,
                 gemini_request,
                 api_key
             )
-            # 如果上面没有抛出异常，则认为密钥有效
             successful_keys.append(api_key)
             return api_key, "valid", None
         except Exception as e:
             error_message = str(e)
             logger.warning(f"Key verification failed for {api_key}: {error_message}")
-            # 验证失败时增加失败计数 (使用与 /verify-key 一致的逻辑)
             async with key_manager.failure_count_lock:
                 if api_key in key_manager.key_failure_counts:
                     key_manager.key_failure_counts[api_key] += 1
                     logger.warning(f"Bulk verification exception for key: {api_key}, incrementing failure count")
                 else:
-                     # 如果密钥不在计数中（可能刚添加或从未失败），初始化为1
                      key_manager.key_failure_counts[api_key] = 1
                      logger.warning(f"Bulk verification exception for key: {api_key}, initializing failure count to 1")
-            failed_keys[api_key] = error_message # 记录失败的 key 和错误信息
+            failed_keys[api_key] = error_message
             return api_key, "invalid", error_message
 
-    # 并发执行所有密钥的验证
     tasks = [_verify_single_key(key) for key in keys_to_verify]
-    results = await asyncio.gather(*tasks, return_exceptions=True) # return_exceptions=True 捕获任务本身的异常
+    results = await asyncio.gather(*tasks, return_exceptions=True)
 
-    # 处理并发执行的结果
     for result in results:
         if isinstance(result, Exception):
-            # 捕获 asyncio.gather 可能遇到的异常（例如任务被取消）
             logger.error(f"An unexpected error occurred during bulk verification task: {result}")
-            # 可以选择如何处理这种任务级别的错误，这里我们简单记录
-            # 也可以将其计入 invalid_count 或单独记录
         elif result:
-             # result 可能是 (key, status, error) 或 Exception
              if not isinstance(result, Exception) and result:
                  key, status, error = result
-                 # 失败信息已在 _verify_single_key 中记录到 failed_keys
              elif isinstance(result, Exception):
-                 # 记录任务本身的异常，可以关联到一个特定的 key 如果可能的话
-                 # 这里简化处理，只记录日志
                  logger.error(f"Task execution error during bulk verification: {result}")
 
     valid_count = len(successful_keys)
     invalid_count = len(failed_keys)
     logger.info(f"Bulk verification finished. Valid: {valid_count}, Invalid: {invalid_count}")
 
-    # 根据是否有失败的 key 决定最终消息和状态
     if failed_keys:
         message = f"批量验证完成。成功: {valid_count}, 失败: {invalid_count}。"
-        # 即使有失败也认为是部分成功，返回 200 OK，让前端处理详细结果
         return JSONResponse({
-            "success": True, # 表示请求处理完成，具体结果看内容
+            "success": True,
             "message": message,
             "successful_keys": successful_keys,
             "failed_keys": failed_keys,
-            "valid_count": valid_count, # 保留计数方便前端快速展示
+            "valid_count": valid_count,
             "invalid_count": invalid_count
         })
     else:
-        # 完全成功
         message = f"批量验证成功完成。所有 {valid_count} 个密钥均有效。"
         return JSONResponse({
             "success": True,

app/router/openai_compatiable_routes.py

@@ -18,7 +18,6 @@ from app.service.openai_compatiable.openai_compatiable_service import OpenAIComp
 router = APIRouter()
 logger = get_openai_compatible_logger()
 
-# 初始化服务
 security_service = SecurityService()
 
 async def get_key_manager():
@@ -52,7 +51,7 @@ async def list_models(
 
 
 @router.post("/openai/v1/chat/completions")
-@RetryHandler(max_retries=settings.MAX_RETRIES, key_arg="api_key")
+@RetryHandler(key_arg="api_key")
 async def chat_completion(
     request: ChatRequest,
     _=Depends(security_service.verify_authorization),
@@ -62,29 +61,23 @@ async def chat_completion(
 ):
     """处理聊天补全请求，支持流式响应和特定模型切换。"""
     operation_name = "chat_completion"
-    # 检查是否为图像生成相关的聊天模型，如果是，则使用付费密钥
     is_image_chat = request.model == f"{settings.CREATE_IMAGE_MODEL}-chat"
-    current_api_key = api_key # 保存原始key（可能是普通key）
+    current_api_key = api_key
     if is_image_chat:
-        current_api_key = await key_manager.get_paid_key() # 获取付费密钥
+        current_api_key = await key_manager.get_paid_key()
 
     async with handle_route_errors(logger, operation_name):
         logger.info(f"Handling chat completion request for model: {request.model}")
         logger.debug(f"Request: \n{request.model_dump_json(indent=2)}")
-        logger.info(f"Using API key: {current_api_key}") # 使用 current_api_key
+        logger.info(f"Using API key: {current_api_key}")
 
         if is_image_chat:
-            # 图像生成聊天，调用特定服务，不处理流式
             response = await openai_service.create_image_chat_completion(request, current_api_key)
-            return response # 直接返回结果
+            return response
         else:
-            # 普通聊天补全
             response = await openai_service.create_chat_completion(request, current_api_key)
-            # 处理流式响应
             if request.stream:
-                 # 假设 openai_service.create_chat_completion 在流式时返回异步生成器
                 return StreamingResponse(response, media_type="text/event-stream")
-            # 非流式直接返回结果
             return response
 
 
@@ -98,7 +91,6 @@ async def generate_image(
     operation_name = "generate_image"
     async with handle_route_errors(logger, operation_name):
         logger.info(f"Handling image generation request for prompt: {request.prompt}")
-        # 强制使用配置的模型，确保请求中包含正确的模型信息
         request.model = settings.CREATE_IMAGE_MODEL
         return await openai_service.generate_images(request)
 

app/router/openai_routes.py

@@ -1,4 +1,4 @@
-from fastapi import APIRouter, Depends, HTTPException
+from fastapi import APIRouter, Depends, HTTPException, Response
 from fastapi.responses import StreamingResponse
 
 from app.config.config import settings
@@ -7,24 +7,26 @@ from app.domain.openai_models import (
     ChatRequest,
     EmbeddingRequest,
     ImageGenerationRequest,
+    TTSRequest,
 )
 from app.handler.retry_handler import RetryHandler
-from app.handler.error_handler import handle_route_errors # 导入共享错误处理器
+from app.handler.error_handler import handle_route_errors
 from app.log.logger import get_openai_logger
 from app.service.chat.openai_chat_service import OpenAIChatService
 from app.service.embedding.embedding_service import EmbeddingService
 from app.service.image.image_create_service import ImageCreateService
+from app.service.tts.tts_service import TTSService
 from app.service.key.key_manager import KeyManager, get_key_manager_instance
 from app.service.model.model_service import ModelService
 
 router = APIRouter()
 logger = get_openai_logger()
 
-# 初始化服务
 security_service = SecurityService()
 model_service = ModelService()
 embedding_service = EmbeddingService()
 image_create_service = ImageCreateService()
+tts_service = TTSService()
 
 
 async def get_key_manager():
@@ -42,6 +44,11 @@ async def get_openai_chat_service(key_manager: KeyManager = Depends(get_key_mana
     return OpenAIChatService(settings.BASE_URL, key_manager)
 
 
+async def get_tts_service():
+    """获取TTS服务实例"""
+    return tts_service
+
+
 @router.get("/v1/models")
 @router.get("/hf/v1/models")
 async def list_models(
@@ -59,45 +66,40 @@ async def list_models(
 
 @router.post("/v1/chat/completions")
 @router.post("/hf/v1/chat/completions")
-@RetryHandler(max_retries=settings.MAX_RETRIES, key_arg="api_key")
+@RetryHandler(key_arg="api_key")
 async def chat_completion(
     request: ChatRequest,
     _=Depends(security_service.verify_authorization),
     api_key: str = Depends(get_next_working_key_wrapper),
-    key_manager: KeyManager = Depends(get_key_manager), # 保留 key_manager 用于获取 paid_key
+    key_manager: KeyManager = Depends(get_key_manager),
     chat_service: OpenAIChatService = Depends(get_openai_chat_service),
 ):
     """处理 OpenAI 聊天补全请求，支持流式响应和特定模型切换。"""
     operation_name = "chat_completion"
-    # 检查是否为图像生成相关的聊天模型
     is_image_chat = request.model == f"{settings.CREATE_IMAGE_MODEL}-chat"
-    current_api_key = api_key # 保存原始 key
+    current_api_key = api_key
     if is_image_chat:
-        current_api_key = await key_manager.get_paid_key() # 获取付费密钥
+        current_api_key = await key_manager.get_paid_key()
 
     async with handle_route_errors(logger, operation_name):
         logger.info(f"Handling chat completion request for model: {request.model}")
         logger.debug(f"Request: \n{request.model_dump_json(indent=2)}")
         logger.info(f"Using API key: {current_api_key}")
 
-        # 检查模型支持性应在错误处理块内，以便捕获并记录错误
         if not await model_service.check_model_support(request.model):
-            # 使用 HTTPException，会被 handle_route_errors 捕获并记录
             raise HTTPException(
                 status_code=400, detail=f"Model {request.model} is not supported"
             )
 
         if is_image_chat:
-            # 图像生成聊天
             response = await chat_service.create_image_chat_completion(request, current_api_key)
-            return response # 直接返回，不处理流式
-        else:
-            # 普通聊天补全
-            response = await chat_service.create_chat_completion(request, current_api_key)
-            # 处理流式响应
             if request.stream:
                 return StreamingResponse(response, media_type="text/event-stream")
-            # 非流式直接返回结果
+            return response
+        else:
+            response = await chat_service.create_chat_completion(request, current_api_key)
+            if request.stream:
+                return StreamingResponse(response, media_type="text/event-stream")
             return response
 
 
@@ -111,8 +113,6 @@ async def generate_image(
     operation_name = "generate_image"
     async with handle_route_errors(logger, operation_name):
         logger.info(f"Handling image generation request for prompt: {request.prompt}")
-        # 注意：这里假设 image_create_service.generate_images 是同步函数
-        # 如果它是异步的，需要 await
         response = image_create_service.generate_images(request)
         return response
 
@@ -155,3 +155,21 @@ async def get_keys_list(
             },
             "total": len(keys_status["valid_keys"]) + len(keys_status["invalid_keys"]),
         }
+
+
+@router.post("/v1/audio/speech")
+@router.post("/hf/v1/audio/speech")
+async def text_to_speech(
+    request: TTSRequest,
+    _=Depends(security_service.verify_authorization),
+    api_key: str = Depends(get_next_working_key_wrapper),
+    tts_service: TTSService = Depends(get_tts_service),
+):
+    """处理 OpenAI TTS 请求。"""
+    operation_name = "text_to_speech"
+    async with handle_route_errors(logger, operation_name):
+        logger.info(f"Handling TTS request for model: {request.model}")
+        logger.debug(f"Request: \n{request.model_dump_json(indent=2)}")
+        logger.info(f"Using API key: {api_key}")
+        audio_data = await tts_service.create_tts(request, api_key)
+        return Response(content=audio_data, media_type="audio/wav")

app/router/routes.py

@@ -8,13 +8,12 @@ from fastapi.templating import Jinja2Templates
 
 from app.core.security import verify_auth_token
 from app.log.logger import get_routes_logger
-from app.router import error_log_routes, gemini_routes, openai_routes, config_routes, scheduler_routes, stats_routes, version_routes, openai_compatiable_routes
+from app.router import error_log_routes, gemini_routes, openai_routes, config_routes, scheduler_routes, stats_routes, version_routes, openai_compatiable_routes, vertex_express_routes
 from app.service.key.key_manager import get_key_manager_instance
 from app.service.stats.stats_service import StatsService
 
 logger = get_routes_logger()
 
-# 配置Jinja2模板
 templates = Jinja2Templates(directory="app/templates")
 
 
@@ -25,7 +24,6 @@ def setup_routers(app: FastAPI) -> None:
     Args:
         app: FastAPI应用程序实例
     """
-    # 包含API路由
     app.include_router(openai_routes.router)
     app.include_router(gemini_routes.router)
     app.include_router(gemini_routes.router_v1beta)
@@ -35,13 +33,12 @@ def setup_routers(app: FastAPI) -> None:
     app.include_router(stats_routes.router)
     app.include_router(version_routes.router)
     app.include_router(openai_compatiable_routes.router)
+    app.include_router(vertex_express_routes.router)
 
-    # 添加页面路由
     setup_page_routes(app)
 
-    # 添加健康检查路由
     setup_health_routes(app)
-    setup_api_stats_routes(app) # Add API stats routes
+    setup_api_stats_routes(app)
 
 
 def setup_page_routes(app: FastAPI) -> None:
@@ -106,16 +103,14 @@ def setup_page_routes(app: FastAPI) -> None:
                     "request": request,
                     "valid_keys": keys_status["valid_keys"],
                     "invalid_keys": keys_status["invalid_keys"],
-                    "total_keys": total_keys, # Renamed for clarity
-                    "valid_key_count": valid_key_count, # Added count
-                    "invalid_key_count": invalid_key_count, # Added count
-                    "api_stats": api_stats, # <-- Pass stats to template
+                    "total_keys": total_keys,
+                    "valid_key_count": valid_key_count,
+                    "invalid_key_count": invalid_key_count,
+                    "api_stats": api_stats,
                 },
             )
         except Exception as e:
             logger.error(f"Error retrieving keys status or API stats: {str(e)}")
-            # Optionally, render template with error or default stats
-            # For now, re-raise to show error page
             raise
             
     @app.get("/config", response_class=HTMLResponse)
@@ -175,16 +170,13 @@ def setup_api_stats_routes(app: FastAPI) -> None:
     async def api_stats_details(request: Request, period: str):
         """获取指定时间段内的 API 调用详情"""
         try:
-            # 验证认证
             auth_token = request.cookies.get("auth_token")
             if not auth_token or not verify_auth_token(auth_token):
                 logger.warning("Unauthorized access attempt to API stats details")
-                # Returning JSON error instead of redirect for API endpoint
                 return {"error": "Unauthorized"}, 401
 
             logger.info(f"Fetching API call details for period: {period}")
-            # Use the service instance here as well
-            stats_service = StatsService() # Create an instance
+            stats_service = StatsService()
             details = await stats_service.get_api_call_details(period)
             return details
         except ValueError as e:

app/router/scheduler_routes.py

@@ -2,22 +2,20 @@
 定时任务控制路由模块
 """
 
-from fastapi import APIRouter, Request, HTTPException, status # 移除 Depends, 添加 Request
+from fastapi import APIRouter, Request, HTTPException, status
 from fastapi.responses import JSONResponse
 
-from app.core.security import verify_auth_token # 导入 verify_auth_token
-from app.scheduler.key_checker import start_scheduler, stop_scheduler
-from app.log.logger import get_scheduler_routes # 使用路由日志记录器
+from app.core.security import verify_auth_token
+from app.scheduler.scheduled_tasks import start_scheduler, stop_scheduler
+from app.log.logger import get_scheduler_routes
 
 logger = get_scheduler_routes()
 
 router = APIRouter(
     prefix="/api/scheduler",
     tags=["Scheduler"]
-    # 移除全局依赖
 )
 
-# 认证检查的辅助函数
 async def verify_token(request: Request):
     auth_token = request.cookies.get("auth_token")
     if not auth_token or not verify_auth_token(auth_token):
@@ -29,14 +27,12 @@ async def verify_token(request: Request):
         )
 
 @router.post("/start", summary="启动定时任务")
-async def start_scheduler_endpoint(request: Request): # 添加 request 参数
+async def start_scheduler_endpoint(request: Request):
     """Start the background scheduler task"""
-    """
-    await verify_token(request) # 在函数开始处进行认证检查
-    """
+    await verify_token(request)
     try:
         logger.info("Received request to start scheduler.")
-        start_scheduler() # 调用 key_checker 中的函数
+        start_scheduler()
         return JSONResponse(content={"message": "Scheduler started successfully."}, status_code=status.HTTP_200_OK)
     except Exception as e:
         logger.error(f"Error starting scheduler: {str(e)}", exc_info=True)
@@ -46,14 +42,12 @@ async def start_scheduler_endpoint(request: Request): # 添加 request 参数
         )
 
 @router.post("/stop", summary="停止定时任务")
-async def stop_scheduler_endpoint(request: Request): # 添加 request 参数
+async def stop_scheduler_endpoint(request: Request):
     """Stop the background scheduler task"""
-    """
-    await verify_token(request) # 在函数开始处进行认证检查
-    """
+    await verify_token(request)
     try:
         logger.info("Received request to stop scheduler.")
-        stop_scheduler() # 调用 key_checker 中的函数
+        stop_scheduler()
         return JSONResponse(content={"message": "Scheduler stopped successfully."}, status_code=status.HTTP_200_OK)
     except Exception as e:
         logger.error(f"Error stopping scheduler: {str(e)}", exc_info=True)

app/router/stats_routes.py

@@ -45,9 +45,6 @@ async def get_key_usage_details(key: str):
     try:
         usage_details = await stats_service.get_key_usage_details_last_24h(key)
         if usage_details is None:
-            # Handle case where key might be valid but has no recent usage,
-            # or if the service layer explicitly returns None for other reasons.
-            # Returning an empty dict is usually fine for the frontend.
             return {}
         return usage_details
     except Exception as e:

app/router/version_routes.py

@@ -21,10 +21,9 @@ async def get_version_info():
     检查当前应用程序版本与最新的 GitHub release 版本。
     """
     try:
-        current_version = get_current_version() # Use imported function
+        current_version = get_current_version()
         update_available, latest_version, error_message = await check_for_updates()
 
-        # Log the result for debugging
         logger.info(f"Version check API result: current={current_version}, latest={latest_version}, available={update_available}, error='{error_message}'")
 
         return VersionInfo(

app/router/vertex_express_routes.py

@@ -0,0 +1,146 @@
+from fastapi import APIRouter, Depends, HTTPException
+from fastapi.responses import StreamingResponse
+from copy import deepcopy
+from app.config.config import settings
+from app.log.logger import get_vertex_express_logger
+from app.core.security import SecurityService
+from app.domain.gemini_models import GeminiRequest
+from app.service.chat.vertex_express_chat_service import GeminiChatService
+from app.service.key.key_manager import KeyManager, get_key_manager_instance
+from app.service.model.model_service import ModelService
+from app.handler.retry_handler import RetryHandler
+from app.handler.error_handler import handle_route_errors
+from app.core.constants import API_VERSION
+
+router = APIRouter(prefix=f"/vertex-express/{API_VERSION}")
+logger = get_vertex_express_logger()
+
+security_service = SecurityService()
+model_service = ModelService()
+
+
+async def get_key_manager():
+    """获取密钥管理器实例"""
+    return await get_key_manager_instance()
+
+
+async def get_next_working_key(key_manager: KeyManager = Depends(get_key_manager)):
+    """获取下一个可用的API密钥"""
+    return await key_manager.get_next_working_vertex_key()
+
+
+async def get_chat_service(key_manager: KeyManager = Depends(get_key_manager)):
+    """获取Gemini聊天服务实例"""
+    return GeminiChatService(settings.VERTEX_EXPRESS_BASE_URL, key_manager)
+
+
+@router.get("/models")
+async def list_models(
+    _=Depends(security_service.verify_key_or_goog_api_key),
+    key_manager: KeyManager = Depends(get_key_manager)
+):
+    """获取可用的 Gemini 模型列表，并根据配置添加衍生模型（搜索、图像、非思考）。"""
+    operation_name = "list_gemini_models"
+    logger.info("-" * 50 + operation_name + "-" * 50)
+    logger.info("Handling Gemini models list request")
+
+    try:
+        api_key = await key_manager.get_first_valid_key()
+        if not api_key:
+            raise HTTPException(status_code=503, detail="No valid API keys available to fetch models.")
+        logger.info(f"Using API key: {api_key}")
+
+        models_data = await model_service.get_gemini_models(api_key)
+        if not models_data or "models" not in models_data:
+            raise HTTPException(status_code=500, detail="Failed to fetch base models list.")
+
+        models_json = deepcopy(models_data)
+        model_mapping = {x.get("name", "").split("/", maxsplit=1)[-1]: x for x in models_json.get("models", [])}
+
+        def add_derived_model(base_name, suffix, display_suffix):
+            model = model_mapping.get(base_name)
+            if not model:
+                logger.warning(f"Base model '{base_name}' not found for derived model '{suffix}'.")
+                return
+            item = deepcopy(model)
+            item["name"] = f"models/{base_name}{suffix}"
+            display_name = f'{item.get("displayName", base_name)}{display_suffix}'
+            item["displayName"] = display_name
+            item["description"] = display_name
+            models_json["models"].append(item)
+
+        if settings.SEARCH_MODELS:
+            for name in settings.SEARCH_MODELS:
+                add_derived_model(name, "-search", " For Search")
+        if settings.IMAGE_MODELS:
+            for name in settings.IMAGE_MODELS:
+                 add_derived_model(name, "-image", " For Image")
+        if settings.THINKING_MODELS:
+            for name in settings.THINKING_MODELS:
+                add_derived_model(name, "-non-thinking", " Non Thinking")
+
+        logger.info("Gemini models list request successful")
+        return models_json
+    except HTTPException as http_exc:
+        raise http_exc
+    except Exception as e:
+        logger.error(f"Error getting Gemini models list: {str(e)}")
+        raise HTTPException(
+            status_code=500, detail="Internal server error while fetching Gemini models list"
+        ) from e
+
+
+@router.post("/models/{model_name}:generateContent")
+@RetryHandler(key_arg="api_key")
+async def generate_content(
+    model_name: str,
+    request: GeminiRequest,
+    _=Depends(security_service.verify_key_or_goog_api_key),
+    api_key: str = Depends(get_next_working_key),
+    key_manager: KeyManager = Depends(get_key_manager),
+    chat_service: GeminiChatService = Depends(get_chat_service)
+):
+    """处理 Gemini 非流式内容生成请求。"""
+    operation_name = "gemini_generate_content"
+    async with handle_route_errors(logger, operation_name, failure_message="Content generation failed"):
+        logger.info(f"Handling Gemini content generation request for model: {model_name}")
+        logger.debug(f"Request: \n{request.model_dump_json(indent=2)}")
+        logger.info(f"Using API key: {api_key}")
+
+        if not await model_service.check_model_support(model_name):
+            raise HTTPException(status_code=400, detail=f"Model {model_name} is not supported")
+
+        response = await chat_service.generate_content(
+            model=model_name,
+            request=request,
+            api_key=api_key
+        )
+        return response
+
+
+@router.post("/models/{model_name}:streamGenerateContent")
+@RetryHandler(key_arg="api_key")
+async def stream_generate_content(
+    model_name: str,
+    request: GeminiRequest,
+    _=Depends(security_service.verify_key_or_goog_api_key),
+    api_key: str = Depends(get_next_working_key),
+    key_manager: KeyManager = Depends(get_key_manager),
+    chat_service: GeminiChatService = Depends(get_chat_service)
+):
+    """处理 Gemini 流式内容生成请求。"""
+    operation_name = "gemini_stream_generate_content"
+    async with handle_route_errors(logger, operation_name, failure_message="Streaming request initiation failed"):
+        logger.info(f"Handling Gemini streaming content generation for model: {model_name}")
+        logger.debug(f"Request: \n{request.model_dump_json(indent=2)}")
+        logger.info(f"Using API key: {api_key}")
+
+        if not await model_service.check_model_support(model_name):
+            raise HTTPException(status_code=400, detail=f"Model {model_name} is not supported")
+
+        response_stream = chat_service.stream_generate_content(
+            model=model_name,
+            request=request,
+            api_key=api_key
+        )
+        return StreamingResponse(response_stream, media_type="text/event-stream")

app/scheduler/key_checker.py

@@ -1,102 +0,0 @@
-from apscheduler.schedulers.asyncio import AsyncIOScheduler
-from app.service.key.key_manager import get_key_manager_instance
-from app.service.chat.gemini_chat_service import GeminiChatService
-from app.domain.gemini_models import GeminiRequest, GeminiContent
-from app.config.config import settings
-from app.log.logger import Logger # 导入 Logger 类
-
-logger = Logger.setup_logger("scheduler") # 使用 Logger.setup_logger
-
-async def check_failed_keys():
-    """
-    定时检查失败次数大于0的API密钥，并尝试验证它们。
-    如果验证成功，重置失败计数；如果失败，增加失败计数。
-    """
-    logger.info("Starting scheduled check for failed API keys...")
-    try:
-        key_manager = await get_key_manager_instance()
-        # 确保 KeyManager 已经初始化
-        if not key_manager or not hasattr(key_manager, 'key_failure_counts'):
-            logger.warning("KeyManager instance not available or not initialized. Skipping check.")
-            return
-
-        # 创建 GeminiChatService 实例用于验证
-        # 注意：这里直接创建实例，而不是通过依赖注入，因为这是后台任务
-        chat_service = GeminiChatService(settings.BASE_URL, key_manager)
-
-        # 获取需要检查的 key 列表 (失败次数 > 0)
-        keys_to_check = []
-        async with key_manager.failure_count_lock: # 访问共享数据需要加锁
-            # 复制一份以避免在迭代时修改字典
-            failure_counts_copy = key_manager.key_failure_counts.copy()
-            keys_to_check = [key for key, count in failure_counts_copy.items() if count > 0] # 检查所有失败次数大于0的key
-
-        if not keys_to_check:
-            logger.info("No keys with failure count > 0 found. Skipping verification.")
-            return
-
-        logger.info(f"Found {len(keys_to_check)} keys with failure count > 0 to verify.")
-
-        for key in keys_to_check:
-            # 隐藏部分 key 用于日志记录
-            log_key = f"{key[:4]}...{key[-4:]}" if len(key) > 8 else key
-            logger.info(f"Verifying key: {log_key}...")
-            try:
-                # 构造测试请求
-                gemini_request = GeminiRequest(
-                    contents=[
-                        GeminiContent(
-                            role="user",
-                            parts=[{"text": "hi"}] # 使用简单的文本进行验证
-                        )
-                    ]
-                )
-                # 调用 generate_content 进行验证
-                await chat_service.generate_content(
-                    settings.TEST_MODEL, # 使用配置中定义的测试模型
-                    gemini_request,
-                    key
-                )
-                # 如果没有抛出异常，说明 key 有效
-                logger.info(f"Key {log_key} verification successful. Resetting failure count.")
-                await key_manager.reset_key_failure_count(key)
-            except Exception as e:
-                # 验证失败，增加失败计数
-                logger.warning(f"Key {log_key} verification failed: {str(e)}. Incrementing failure count.")
-                # 直接操作计数器，需要加锁
-                async with key_manager.failure_count_lock:
-                    # 再次检查 key 是否存在且失败次数未达上限
-                    if key in key_manager.key_failure_counts and key_manager.key_failure_counts[key] < key_manager.MAX_FAILURES:
-                        key_manager.key_failure_counts[key] += 1
-                        logger.info(f"Failure count for key {log_key} incremented to {key_manager.key_failure_counts[key]}.")
-                    elif key in key_manager.key_failure_counts:
-                         logger.warning(f"Key {log_key} reached MAX_FAILURES ({key_manager.MAX_FAILURES}). Not incrementing further.")
-
-
-    except Exception as e:
-        logger.error(f"An error occurred during the scheduled key check: {str(e)}", exc_info=True)
-
-def setup_scheduler():
-    """设置并启动 APScheduler"""
-    scheduler = AsyncIOScheduler(timezone=str(settings.TIMEZONE)) # 从配置读取时区
-    # 添加定时任务，例如每小时执行一次 (可以调整)
-    scheduler.add_job(check_failed_keys, 'interval', hours=settings.CHECK_INTERVAL_HOURS)
-    scheduler.start()
-    logger.info(f"Scheduler started. Key check job scheduled to run every {settings.CHECK_INTERVAL_HOURS} hour(s).")
-    return scheduler
-
-# 可以在这里添加一个全局的 scheduler 实例，以便在应用关闭时优雅地停止
-scheduler_instance = None
-
-def start_scheduler():
-    global scheduler_instance
-    if scheduler_instance is None or not scheduler_instance.running:
-        logger.info("Starting scheduler...")
-        scheduler_instance = setup_scheduler()
-    logger.info("Scheduler is already running.")
-
-def stop_scheduler():
-    global scheduler_instance
-    if scheduler_instance and scheduler_instance.running:
-        scheduler_instance.shutdown()
-        logger.info("Scheduler stopped.")

app/scheduler/scheduled_tasks.py

@@ -0,0 +1,159 @@
+
+from apscheduler.schedulers.asyncio import AsyncIOScheduler
+
+from app.config.config import settings
+from app.domain.gemini_models import GeminiContent, GeminiRequest
+from app.log.logger import Logger
+from app.service.chat.gemini_chat_service import GeminiChatService
+from app.service.error_log.error_log_service import delete_old_error_logs
+from app.service.key.key_manager import get_key_manager_instance
+from app.service.request_log.request_log_service import delete_old_request_logs_task
+
+logger = Logger.setup_logger("scheduler")
+
+
+async def check_failed_keys():
+    """
+    定时检查失败次数大于0的API密钥，并尝试验证它们。
+    如果验证成功，重置失败计数；如果失败，增加失败计数。
+    """
+    logger.info("Starting scheduled check for failed API keys...")
+    try:
+        key_manager = await get_key_manager_instance()
+        # 确保 KeyManager 已经初始化
+        if not key_manager or not hasattr(key_manager, "key_failure_counts"):
+            logger.warning(
+                "KeyManager instance not available or not initialized. Skipping check."
+            )
+            return
+
+        # 创建 GeminiChatService 实例用于验证
+        # 注意：这里直接创建实例，而不是通过依赖注入，因为这是后台任务
+        chat_service = GeminiChatService(settings.BASE_URL, key_manager)
+
+        # 获取需要检查的 key 列表 (失败次数 > 0)
+        keys_to_check = []
+        async with key_manager.failure_count_lock:  # 访问共享数据需要加锁
+            # 复制一份以避免在迭代时修改字典
+            failure_counts_copy = key_manager.key_failure_counts.copy()
+            keys_to_check = [
+                key for key, count in failure_counts_copy.items() if count > 0
+            ]  # 检查所有失败次数大于0的key
+
+        if not keys_to_check:
+            logger.info("No keys with failure count > 0 found. Skipping verification.")
+            return
+
+        logger.info(
+            f"Found {len(keys_to_check)} keys with failure count > 0 to verify."
+        )
+
+        for key in keys_to_check:
+            # 隐藏部分 key 用于日志记录
+            log_key = f"{key[:4]}...{key[-4:]}" if len(key) > 8 else key
+            logger.info(f"Verifying key: {log_key}...")
+            try:
+                # 构造测试请求
+                gemini_request = GeminiRequest(
+                    contents=[
+                        GeminiContent(
+                            role="user",
+                            parts=[{"text": "hi"}],
+                        )
+                    ]
+                )
+                await chat_service.generate_content(
+                    settings.TEST_MODEL, gemini_request, key
+                )
+                logger.info(
+                    f"Key {log_key} verification successful. Resetting failure count."
+                )
+                await key_manager.reset_key_failure_count(key)
+            except Exception as e:
+                logger.warning(
+                    f"Key {log_key} verification failed: {str(e)}. Incrementing failure count."
+                )
+                # 直接操作计数器，需要加锁
+                async with key_manager.failure_count_lock:
+                    # 再次检查 key 是否存在且失败次数未达上限
+                    if (
+                        key in key_manager.key_failure_counts
+                        and key_manager.key_failure_counts[key]
+                        < key_manager.MAX_FAILURES
+                    ):
+                        key_manager.key_failure_counts[key] += 1
+                        logger.info(
+                            f"Failure count for key {log_key} incremented to {key_manager.key_failure_counts[key]}."
+                        )
+                    elif key in key_manager.key_failure_counts:
+                        logger.warning(
+                            f"Key {log_key} reached MAX_FAILURES ({key_manager.MAX_FAILURES}). Not incrementing further."
+                        )
+
+    except Exception as e:
+        logger.error(
+            f"An error occurred during the scheduled key check: {str(e)}", exc_info=True
+        )
+
+
+def setup_scheduler():
+    """设置并启动 APScheduler"""
+    scheduler = AsyncIOScheduler(timezone=str(settings.TIMEZONE))  # 从配置读取时区
+    # 添加检查失败密钥的定时任务
+    scheduler.add_job(
+        check_failed_keys,
+        "interval",
+        hours=settings.CHECK_INTERVAL_HOURS,
+        id="check_failed_keys_job",
+        name="Check Failed API Keys",
+    )
+    logger.info(
+        f"Key check job scheduled to run every {settings.CHECK_INTERVAL_HOURS} hour(s)."
+    )
+
+    # 新增：添加自动删除错误日志的定时任务，每天凌晨3点执行
+    scheduler.add_job(
+        delete_old_error_logs,
+        "cron",
+        hour=3,
+        minute=0,
+        id="delete_old_error_logs_job",
+        name="Delete Old Error Logs",
+    )
+    logger.info("Auto-delete error logs job scheduled to run daily at 3:00 AM.")
+
+    # 新增：添加自动删除请求日志的定时任务，每天凌晨3点05分执行
+    scheduler.add_job(
+        delete_old_request_logs_task,
+        "cron",
+        hour=3,
+        minute=5,
+        id="delete_old_request_logs_job",
+        name="Delete Old Request Logs",
+    )
+    logger.info(
+        f"Auto-delete request logs job scheduled to run daily at 3:05 AM, if enabled and AUTO_DELETE_REQUEST_LOGS_DAYS is set to {settings.AUTO_DELETE_REQUEST_LOGS_DAYS} days."
+    )
+
+    scheduler.start()
+    logger.info("Scheduler started with all jobs.")
+    return scheduler
+
+
+# 可以在这里添加一个全局的 scheduler 实例，以便在应用关闭时优雅地停止
+scheduler_instance = None
+
+
+def start_scheduler():
+    global scheduler_instance
+    if scheduler_instance is None or not scheduler_instance.running:
+        logger.info("Starting scheduler...")
+        scheduler_instance = setup_scheduler()
+    logger.info("Scheduler is already running.")
+
+
+def stop_scheduler():
+    global scheduler_instance
+    if scheduler_instance and scheduler_instance.running:
+        scheduler_instance.shutdown()
+        logger.info("Scheduler stopped.")

app/service/chat/gemini_chat_service.py

@@ -2,17 +2,18 @@
 
 import json
 import re
-import datetime # Add datetime import
-import time # Add time import
+import datetime
+import time
 from typing import Any, AsyncGenerator, Dict, List
 from app.config.config import settings
+from app.core.constants import GEMINI_2_FLASH_EXP_SAFETY_SETTINGS
 from app.domain.gemini_models import GeminiRequest
 from app.handler.response_handler import GeminiResponseHandler
 from app.handler.stream_optimizer import gemini_optimizer
 from app.log.logger import get_gemini_logger
 from app.service.client.api_client import GeminiApiClient
 from app.service.key.key_manager import KeyManager
-from app.database.services import add_error_log, add_request_log # Import add_request_log
+from app.database.services import add_error_log, add_request_log
 
 logger = get_gemini_logger()
 
@@ -27,6 +28,33 @@ def _has_image_parts(contents: List[Dict[str, Any]]) -> bool:
     return False
 
 
+def _clean_json_schema_properties(obj: Any) -> Any:
+    """清理JSON Schema中Gemini API不支持的字段"""
+    if not isinstance(obj, dict):
+        return obj
+    
+    # Gemini API不支持的JSON Schema字段
+    unsupported_fields = {
+        "exclusiveMaximum", "exclusiveMinimum", "const", "examples", 
+        "contentEncoding", "contentMediaType", "if", "then", "else",
+        "allOf", "anyOf", "oneOf", "not", "definitions", "$schema",
+        "$id", "$ref", "$comment", "readOnly", "writeOnly"
+    }
+    
+    cleaned = {}
+    for key, value in obj.items():
+        if key in unsupported_fields:
+            continue
+        if isinstance(value, dict):
+            cleaned[key] = _clean_json_schema_properties(value)
+        elif isinstance(value, list):
+            cleaned[key] = [_clean_json_schema_properties(item) for item in value]
+        else:
+            cleaned[key] = value
+    
+    return cleaned
+
+
 def _build_tools(model: str, payload: Dict[str, Any]) -> List[Dict[str, Any]]:
     """构建工具"""
     
@@ -39,7 +67,15 @@ def _build_tools(model: str, payload: Dict[str, Any]) -> List[Dict[str, Any]]:
             for k, v in item.items():
                 if k == "functionDeclarations" and v and isinstance(v, list):
                     functions = record.get("functionDeclarations", [])
-                    functions.extend(v)
+                    # 清理每个函数声明中的不支持字段
+                    cleaned_functions = []
+                    for func in v:
+                        if isinstance(func, dict):
+                            cleaned_func = _clean_json_schema_properties(func)
+                            cleaned_functions.append(cleaned_func)
+                        else:
+                            cleaned_functions.append(func)
+                    functions.extend(cleaned_functions)
                     record["functionDeclarations"] = functions
                 else:
                     record[k] = v
@@ -73,20 +109,28 @@ def _build_tools(model: str, payload: Dict[str, Any]) -> List[Dict[str, Any]]:
 def _get_safety_settings(model: str) -> List[Dict[str, str]]:
     """获取安全设置"""
     if model == "gemini-2.0-flash-exp":
-        return [
-            {"category": "HARM_CATEGORY_HARASSMENT", "threshold": "OFF"},
-            {"category": "HARM_CATEGORY_HATE_SPEECH", "threshold": "OFF"},
-            {"category": "HARM_CATEGORY_SEXUALLY_EXPLICIT", "threshold": "OFF"},
-            {"category": "HARM_CATEGORY_DANGEROUS_CONTENT", "threshold": "OFF"},
-            {"category": "HARM_CATEGORY_CIVIC_INTEGRITY", "threshold": "OFF"},
-        ]
-    return [
-        {"category": "HARM_CATEGORY_HARASSMENT", "threshold": "OFF"},
-        {"category": "HARM_CATEGORY_HATE_SPEECH", "threshold": "OFF"},
-        {"category": "HARM_CATEGORY_SEXUALLY_EXPLICIT", "threshold": "OFF"},
-        {"category": "HARM_CATEGORY_DANGEROUS_CONTENT", "threshold": "OFF"},
-        {"category": "HARM_CATEGORY_CIVIC_INTEGRITY", "threshold": "BLOCK_NONE"},
-    ]
+        return GEMINI_2_FLASH_EXP_SAFETY_SETTINGS
+    return settings.SAFETY_SETTINGS
+
+
+def _filter_empty_parts(contents: List[Dict[str, Any]]) -> List[Dict[str, Any]]:
+    """Filters out contents with empty or invalid parts."""
+    if not contents:
+        return []
+
+    filtered_contents = []
+    for content in contents:
+        if not content or "parts" not in content or not isinstance(content.get("parts"), list):
+            continue
+
+        valid_parts = [part for part in content["parts"] if isinstance(part, dict) and part]
+
+        if valid_parts:
+            new_content = content.copy()
+            new_content["parts"] = valid_parts
+            filtered_contents.append(new_content)
+
+    return filtered_contents
 
 
 def _build_payload(model: str, request: GeminiRequest) -> Dict[str, Any]:
@@ -98,21 +142,41 @@ def _build_payload(model: str, request: GeminiRequest) -> Dict[str, Any]:
             request_dict["generationConfig"].pop("maxOutputTokens")
     
     payload = {
-        "contents": request_dict.get("contents", []),
+        "contents": _filter_empty_parts(request_dict.get("contents", [])),
         "tools": _build_tools(model, request_dict),
         "safetySettings": _get_safety_settings(model),
-        "generationConfig": request_dict.get("generationConfig", {}),
-        "systemInstruction": request_dict.get("systemInstruction", ""),
+        "generationConfig": request_dict.get("generationConfig"),
+        "systemInstruction": request_dict.get("systemInstruction"),
     }
 
+    # 确保 generationConfig 不为 None
+    if payload["generationConfig"] is None:
+        payload["generationConfig"] = {}
+
     if model.endswith("-image") or model.endswith("-image-generation"):
         payload.pop("systemInstruction")
         payload["generationConfig"]["responseModalities"] = ["Text", "Image"]
-        
-    if model.endswith("-non-thinking"):
-        payload["generationConfig"]["thinkingConfig"] = {"thinkingBudget": 0} 
-    if model in settings.THINKING_BUDGET_MAP:
-        payload["generationConfig"]["thinkingConfig"] = {"thinkingBudget": settings.THINKING_BUDGET_MAP.get(model,1000)}
+    
+    # 处理思考配置：优先使用客户端提供的配置，否则使用默认配置
+    client_thinking_config = None
+    if request.generationConfig and request.generationConfig.thinkingConfig:
+        client_thinking_config = request.generationConfig.thinkingConfig
+    
+    if client_thinking_config is not None:
+        # 客户端提供了思考配置，直接使用
+        payload["generationConfig"]["thinkingConfig"] = client_thinking_config
+    else:
+        # 客户端没有提供思考配置，使用默认配置    
+        if model.endswith("-non-thinking"):
+            payload["generationConfig"]["thinkingConfig"] = {"thinkingBudget": 0} 
+        elif model in settings.THINKING_BUDGET_MAP:
+            if settings.SHOW_THINKING_PROCESS:
+                payload["generationConfig"]["thinkingConfig"] = {
+                    "thinkingBudget": settings.THINKING_BUDGET_MAP.get(model,1000),
+                    "includeThoughts": True
+                }
+            else:
+                payload["generationConfig"]["thinkingConfig"] = {"thinkingBudget": settings.THINKING_BUDGET_MAP.get(model,1000)}
 
     return payload
 
@@ -142,7 +206,7 @@ class GeminiChatService:
         self, original_response: Dict[str, Any], text: str
     ) -> Dict[str, Any]:
         """创建包含指定文本的响应"""
-        response_copy = json.loads(json.dumps(original_response))  # 深拷贝
+        response_copy = json.loads(json.dumps(original_response))
         if response_copy.get("candidates") and response_copy["candidates"][0].get(
             "content", {}
         ).get("parts"):
@@ -155,7 +219,7 @@ class GeminiChatService:
         """生成内容"""
         payload = _build_payload(model, request)
         start_time = time.perf_counter()
-        request_datetime = datetime.datetime.now() # Record request time
+        request_datetime = datetime.datetime.now()
         is_success = False
         status_code = None
         response = None
@@ -163,20 +227,18 @@ class GeminiChatService:
         try:
             response = await self.api_client.generate_content(payload, model, api_key)
             is_success = True
-            status_code = 200 # Assume 200 on success
+            status_code = 200
             return self.response_handler.handle_response(response, model, stream=False)
         except Exception as e:
             is_success = False
             error_log_msg = str(e)
             logger.error(f"Normal API call failed with error: {error_log_msg}")
-            # Try to parse status code from exception
             match = re.search(r"status code (\d+)", error_log_msg)
             if match:
                 status_code = int(match.group(1))
             else:
-                status_code = 500 # Default to 500 if parsing fails
+                status_code = 500
 
-            # Log error to error log table
             await add_error_log(
                 gemini_key=api_key,
                 model_name=model,
@@ -185,11 +247,58 @@ class GeminiChatService:
                 error_code=status_code,
                 request_msg=payload
             )
-            raise e # Re-throw exception for upstream handling
+            raise e
+        finally:
+            end_time = time.perf_counter()
+            latency_ms = int((end_time - start_time) * 1000)
+            await add_request_log(
+                model_name=model,
+                api_key=api_key,
+                is_success=is_success,
+                status_code=status_code,
+                latency_ms=latency_ms,
+                request_time=request_datetime
+            )
+
+    async def count_tokens(
+        self, model: str, request: GeminiRequest, api_key: str
+    ) -> Dict[str, Any]:
+        """计算token数量"""
+        # countTokens API只需要contents
+        payload = {"contents": _filter_empty_parts(request.model_dump().get("contents", []))}
+        start_time = time.perf_counter()
+        request_datetime = datetime.datetime.now()
+        is_success = False
+        status_code = None
+        response = None
+
+        try:
+            response = await self.api_client.count_tokens(payload, model, api_key)
+            is_success = True
+            status_code = 200
+            return response
+        except Exception as e:
+            is_success = False
+            error_log_msg = str(e)
+            logger.error(f"Count tokens API call failed with error: {error_log_msg}")
+            match = re.search(r"status code (\d+)", error_log_msg)
+            if match:
+                status_code = int(match.group(1))
+            else:
+                status_code = 500
+
+            await add_error_log(
+                gemini_key=api_key,
+                model_name=model,
+                error_type="gemini-count-tokens",
+                error_log=error_log_msg,
+                error_code=status_code,
+                request_msg=payload
+            )
+            raise e
         finally:
             end_time = time.perf_counter()
             latency_ms = int((end_time - start_time) * 1000)
-            # Log request to request log table
             await add_request_log(
                 model_name=model,
                 api_key=api_key,
@@ -214,7 +323,7 @@ class GeminiChatService:
             request_datetime = datetime.datetime.now()
             start_time = time.perf_counter()
             current_attempt_key = api_key
-            final_api_key = current_attempt_key # Update final key used
+            final_api_key = current_attempt_key
             try:
                 async for line in self.api_client.stream_generate_content(
                     payload, model, current_attempt_key
@@ -251,16 +360,14 @@ class GeminiChatService:
                 logger.warning(
                     f"Streaming API call failed with error: {error_log_msg}. Attempt {retries} of {max_retries}"
                 )
-                # Parse error code for logging
                 match = re.search(r"status code (\d+)", error_log_msg)
                 if match:
                     status_code = int(match.group(1))
                 else:
                     status_code = 500
 
-                # Log error to error log table
                 await add_error_log(
-                    gemini_key=current_attempt_key, # Log key used for this failed attempt
+                    gemini_key=current_attempt_key,
                     model_name=model,
                     error_type="gemini-chat-stream",
                     error_log=error_log_msg,
@@ -268,28 +375,26 @@ class GeminiChatService:
                     request_msg=payload
                 )
 
-                # Attempt to switch API Key
                 api_key = await self.key_manager.handle_api_failure(current_attempt_key, retries)
                 if api_key:
                     logger.info(f"Switched to new API key: {api_key}")
-                else: # No more keys or retries exceeded by handle_api_failure logic
-                        logger.error(f"No valid API key available after {retries} retries.")
-                        break # Exit loop if no key available
+                else:
+                    logger.error(f"No valid API key available after {retries} retries.")
+                    break
 
                 if retries >= max_retries:
                     logger.error(
                         f"Max retries ({max_retries}) reached for streaming."
                     )
-                    break # Exit loop after max retries
+                    break
             finally:
-                # Log the final outcome of the streaming request
                 end_time = time.perf_counter()
                 latency_ms = int((end_time - start_time) * 1000)
                 await add_request_log(
                     model_name=model,
-                    api_key=final_api_key, # Log the last key used
-                    is_success=is_success, # Log the final success status
-                    status_code=status_code, # Log the last known status code
-                    latency_ms=latency_ms, # Log total time including retries
+                    api_key=final_api_key,
+                    is_success=is_success,
+                    status_code=status_code,
+                    latency_ms=latency_ms,
                     request_time=request_datetime
                 )

app/service/chat/openai_chat_service.py

@@ -1,5 +1,6 @@
 # app/services/chat_service.py
 
+import asyncio
 import datetime
 import json
 import re
@@ -8,6 +9,7 @@ from copy import deepcopy
 from typing import Any, AsyncGenerator, Dict, List, Optional, Union
 
 from app.config.config import settings
+from app.core.constants import GEMINI_2_FLASH_EXP_SAFETY_SETTINGS
 from app.database.services import (
     add_error_log,
     add_request_log,
@@ -24,16 +26,43 @@ from app.service.key.key_manager import KeyManager
 logger = get_openai_logger()
 
 
-def _has_media_parts(contents: List[Dict[str, Any]]) -> bool:
-    """判断消息是否包含图片、音频或视频部分 (inline_data)"""
-    for content in contents:
-        if content and "parts" in content and isinstance(content["parts"], list):
-            for part in content["parts"]:
-                if isinstance(part, dict) and "inline_data" in part:
+def _has_media_parts(messages: List[Dict[str, Any]]) -> bool:
+    """判断消息是否包含多媒体部分"""
+    for message in messages:
+        if "parts" in message:
+            for part in message["parts"]:
+                if "image_url" in part or "inline_data" in part:
                     return True
     return False
 
 
+def _clean_json_schema_properties(obj: Any) -> Any:
+    """清理JSON Schema中Gemini API不支持的字段"""
+    if not isinstance(obj, dict):
+        return obj
+    
+    # Gemini API不支持的JSON Schema字段
+    unsupported_fields = {
+        "exclusiveMaximum", "exclusiveMinimum", "const", "examples", 
+        "contentEncoding", "contentMediaType", "if", "then", "else",
+        "allOf", "anyOf", "oneOf", "not", "definitions", "$schema",
+        "$id", "$ref", "$comment", "readOnly", "writeOnly"
+    }
+    
+    cleaned = {}
+    for key, value in obj.items():
+        if key in unsupported_fields:
+            continue
+        if isinstance(value, dict):
+            cleaned[key] = _clean_json_schema_properties(value)
+        elif isinstance(value, list):
+            cleaned[key] = [_clean_json_schema_properties(item) for item in value]
+        else:
+            cleaned[key] = value
+    
+    return cleaned
+
+
 def _build_tools(
     request: ChatRequest, messages: List[Dict[str, Any]]
 ) -> List[Dict[str, Any]]:
@@ -49,7 +78,7 @@ def _build_tools(
             or model.endswith("-image")
             or model.endswith("-image-generation")
         )
-        and not _has_media_parts(messages)  # Use the updated check
+        and not _has_media_parts(messages)
     ):
         tool["codeExecution"] = {}
         logger.debug("Code execution tool enabled.")
@@ -74,6 +103,8 @@ def _build_tools(
                 ):
                     function.pop("parameters", None)
 
+                # 清理函数中的不支持字段
+                function = _clean_json_schema_properties(function)
                 function_declarations.append(function)
 
         if function_declarations:
@@ -81,8 +112,13 @@ def _build_tools(
             names, functions = set(), []
             for fc in function_declarations:
                 if fc.get("name") not in names:
-                    names.add(fc.get("name"))
-                    functions.append(fc)
+                    if fc.get("name")=="googleSearch":
+                        # cherry开启内置搜索时，添加googleSearch工具
+                        tool["googleSearch"] = {}
+                    else:
+                        # 其他函数，添加到functionDeclarations中
+                        names.add(fc.get("name"))
+                        functions.append(fc)
 
             tool["functionDeclarations"] = functions
 
@@ -102,7 +138,7 @@ def _get_safety_settings(model: str) -> List[Dict[str, str]]:
     #     and "gemini-2.0-pro-exp" not in model
     # ):
     if model == "gemini-2.0-flash-exp":
-        return settings.GEMINI_2_FLASH_EXP_SAFETY_SETTINGS
+        return GEMINI_2_FLASH_EXP_SAFETY_SETTINGS
     return settings.SAFETY_SETTINGS
 
 
@@ -130,9 +166,13 @@ def _build_payload(
     if request.model.endswith("-non-thinking"):
         payload["generationConfig"]["thinkingConfig"] = {"thinkingBudget": 0}
     if request.model in settings.THINKING_BUDGET_MAP:
-        payload["generationConfig"]["thinkingConfig"] = {
-            "thinkingBudget": settings.THINKING_BUDGET_MAP.get(request.model, 1000)
-        }
+        if settings.SHOW_THINKING_PROCESS:
+            payload["generationConfig"]["thinkingConfig"] = {
+                "thinkingBudget": settings.THINKING_BUDGET_MAP.get(request.model, 1000),
+                "includeThoughts": True
+            }
+        else:
+            payload["generationConfig"]["thinkingConfig"] = {"thinkingBudget": settings.THINKING_BUDGET_MAP.get(request.model, 1000)}
 
     if (
         instruction
@@ -171,7 +211,7 @@ class OpenAIChatService:
         self, original_chunk: Dict[str, Any], text: str
     ) -> Dict[str, Any]:
         """创建包含指定文本的OpenAI响应块"""
-        chunk_copy = json.loads(json.dumps(original_chunk))  # 深拷贝
+        chunk_copy = json.loads(json.dumps(original_chunk))
         if chunk_copy.get("choices") and "delta" in chunk_copy["choices"][0]:
             chunk_copy["choices"][0]["delta"]["content"] = text
         return chunk_copy
@@ -182,10 +222,8 @@ class OpenAIChatService:
         api_key: str,
     ) -> Union[Dict[str, Any], AsyncGenerator[str, None]]:
         """创建聊天完成"""
-        # 转换消息格式
         messages, instruction = self.message_converter.convert(request.messages)
 
-        # 构建请求payload
         payload = _build_payload(request, messages, instruction)
 
         if request.stream:
@@ -203,16 +241,20 @@ class OpenAIChatService:
         response = None
         try:
             response = await self.api_client.generate_content(payload, model, api_key)
+            usage_metadata = response.get("usageMetadata", {})
             is_success = True
             status_code = 200
             return self.response_handler.handle_response(
-                response, model, stream=False, finish_reason="stop"
+                response,
+                model,
+                stream=False,
+                finish_reason="stop",
+                usage_metadata=usage_metadata,
             )
         except Exception as e:
             is_success = False
             error_log_msg = str(e)
             logger.error(f"Normal API call failed with error: {error_log_msg}")
-            # Try to parse status code from exception
             match = re.search(r"status code (\d+)", error_log_msg)
             if match:
                 status_code = int(match.group(1))
@@ -240,10 +282,119 @@ class OpenAIChatService:
                 request_time=request_datetime,
             )
 
+    async def _fake_stream_logic_impl(
+        self, model: str, payload: Dict[str, Any], api_key: str
+    ) -> AsyncGenerator[str, None]:
+        """处理伪流式 (fake stream) 的核心逻辑"""
+        logger.info(
+            f"Fake streaming enabled for model: {model}. Calling non-streaming endpoint."
+        )
+        keep_sending_empty_data = True
+
+        async def send_empty_data_locally() -> AsyncGenerator[str, None]:
+            """定期发送空数据以保持连接"""
+            while keep_sending_empty_data:
+                await asyncio.sleep(settings.FAKE_STREAM_EMPTY_DATA_INTERVAL_SECONDS)
+                if keep_sending_empty_data:
+                    empty_chunk = self.response_handler.handle_response({}, model, stream=True, finish_reason='stop', usage_metadata=None)
+                    yield f"data: {json.dumps(empty_chunk)}\n\n"
+                    logger.debug("Sent empty data chunk for fake stream heartbeat.")
+
+        empty_data_generator = send_empty_data_locally()
+        api_response_task = asyncio.create_task(
+            self.api_client.generate_content(payload, model, api_key)
+        )
+
+        try:
+            while not api_response_task.done():
+                try:
+                    next_empty_chunk = await asyncio.wait_for(
+                        empty_data_generator.__anext__(), timeout=0.1
+                    )
+                    yield next_empty_chunk
+                except asyncio.TimeoutError:
+                    pass
+                except (
+                    StopAsyncIteration
+                ):
+                    break
+
+            response = await api_response_task
+        finally:
+            keep_sending_empty_data = False
+
+        if response and response.get("candidates"):
+            response = self.response_handler.handle_response(response, model, stream=True, finish_reason='stop', usage_metadata=response.get("usageMetadata", {}))
+            yield f"data: {json.dumps(response)}\n\n"
+            logger.info(f"Sent full response content for fake stream: {model}")
+        else:
+            error_message = "Failed to get response from model"
+            if (
+                response and isinstance(response, dict) and response.get("error")
+            ):
+                error_details = response.get("error")
+                if isinstance(error_details, dict):
+                    error_message = error_details.get("message", error_message)
+
+            logger.error(
+                f"No candidates or error in response for fake stream model {model}: {response}"
+            )
+            error_chunk = self.response_handler.handle_response({}, model, stream=True, finish_reason='stop', usage_metadata=None)
+            yield f"data: {json.dumps(error_chunk)}\n\n"
+
+    async def _real_stream_logic_impl(
+        self, model: str, payload: Dict[str, Any], api_key: str
+    ) -> AsyncGenerator[str, None]:
+        """处理真实流式 (real stream) 的核心逻辑"""
+        tool_call_flag = False
+        usage_metadata = None
+        async for line in self.api_client.stream_generate_content(
+            payload, model, api_key
+        ):
+            if line.startswith("data:"):
+                chunk_str = line[6:]
+                if not chunk_str or chunk_str.isspace():
+                    logger.debug(
+                        f"Received empty data line for model {model}, skipping."
+                    )
+                    continue
+                try:
+                    chunk = json.loads(chunk_str)
+                    usage_metadata = chunk.get("usageMetadata", {})
+                except json.JSONDecodeError:
+                    logger.error(
+                        f"Failed to decode JSON from stream for model {model}: {chunk_str}"
+                    )
+                    continue
+                openai_chunk = self.response_handler.handle_response(
+                    chunk, model, stream=True, finish_reason=None, usage_metadata=usage_metadata
+                )
+                if openai_chunk:
+                    text = self._extract_text_from_openai_chunk(openai_chunk)
+                    if text and settings.STREAM_OPTIMIZER_ENABLED:
+                        async for (
+                            optimized_chunk_data
+                        ) in openai_optimizer.optimize_stream_output(
+                            text,
+                            lambda t: self._create_char_openai_chunk(openai_chunk, t),
+                            lambda c: f"data: {json.dumps(c)}\n\n",
+                        ):
+                            yield optimized_chunk_data
+                    else:
+                        if openai_chunk.get("choices") and openai_chunk["choices"][0].get("delta", {}).get("tool_calls"):
+                            tool_call_flag = True
+
+                        yield f"data: {json.dumps(openai_chunk)}\n\n"
+
+        if tool_call_flag:
+            yield f"data: {json.dumps(self.response_handler.handle_response({}, model, stream=True, finish_reason='tool_calls', usage_metadata=usage_metadata))}\n\n"
+        else:
+            yield f"data: {json.dumps(self.response_handler.handle_response({}, model, stream=True, finish_reason='stop', usage_metadata=usage_metadata))}\n\n"
+
     async def _handle_stream_completion(
         self, model: str, payload: Dict[str, Any], api_key: str
     ) -> AsyncGenerator[str, None]:
-        """处理流式聊天完成，添加重试逻辑"""
+        """处理流式聊天完成，添加重试逻辑和假流式支持"""
         retries = 0
         max_retries = settings.MAX_RETRIES
         is_success = False
@@ -253,63 +404,53 @@ class OpenAIChatService:
         while retries < max_retries:
             start_time = time.perf_counter()
             request_datetime = datetime.datetime.now()
-            current_attempt_key = api_key
-            final_api_key = current_attempt_key
+            current_attempt_key = final_api_key
+
             try:
-                tool_call_flag = False
-                async for line in self.api_client.stream_generate_content(
-                    payload, model, current_attempt_key
-                ):
-                    # print(line)
-                    if line.startswith("data:"):
-                        chunk = json.loads(line[6:])
-                        openai_chunk = self.response_handler.handle_response(
-                            chunk, model, stream=True, finish_reason=None
-                        )
-                        if openai_chunk:
-                            # 提取文本内容
-                            text = self._extract_text_from_openai_chunk(openai_chunk)
-                            if text and settings.STREAM_OPTIMIZER_ENABLED:
-                                # 使用流式输出优化器处理文本输出
-                                async for (
-                                    optimized_chunk
-                                ) in openai_optimizer.optimize_stream_output(
-                                    text,
-                                    lambda t: self._create_char_openai_chunk(
-                                        openai_chunk, t
-                                    ),
-                                    lambda c: f"data: {json.dumps(c)}\n\n",
-                                ):
-                                    yield optimized_chunk
-                            else:
-                                # 如果没有文本内容（如工具调用等），整块输出
-                                if "tool_calls" in json.dumps(openai_chunk):
-                                    tool_call_flag = True
-                                yield f"data: {json.dumps(openai_chunk)}\n\n"
-                if tool_call_flag:
-                    yield f"data: {json.dumps(self.response_handler.handle_response({}, model, stream=True, finish_reason='tool_calls'))}\n\n"
+                stream_generator = None
+                if settings.FAKE_STREAM_ENABLED:
+                    logger.info(
+                        f"Using fake stream logic for model: {model}, Attempt: {retries + 1}"
+                    )
+                    stream_generator = self._fake_stream_logic_impl(
+                        model, payload, current_attempt_key
+                    )
                 else:
-                    yield f"data: {json.dumps(self.response_handler.handle_response({}, model, stream=True, finish_reason='stop'))}\n\n"
+                    logger.info(
+                        f"Using real stream logic for model: {model}, Attempt: {retries + 1}"
+                    )
+                    stream_generator = self._real_stream_logic_impl(
+                        model, payload, current_attempt_key
+                    )
+
+                async for chunk_data in stream_generator:
+                    yield chunk_data
+
                 yield "data: [DONE]\n\n"
-                logger.info("Streaming completed successfully")
+                logger.info(
+                    f"Streaming completed successfully for model: {model}, FakeStream: {settings.FAKE_STREAM_ENABLED}, Attempt: {retries + 1}"
+                )
                 is_success = True
                 status_code = 200
-                break  # 成功后退出循环
+                break
+
             except Exception as e:
                 retries += 1
                 is_success = False
                 error_log_msg = str(e)
                 logger.warning(
-                    f"Streaming API call failed with error: {error_log_msg}. Attempt {retries} of {max_retries}"
+                    f"Streaming API call failed with error: {error_log_msg}. Attempt {retries} of {max_retries} with key {current_attempt_key}"
                 )
-                # Parse error code for logging
-                match = re.search(r"status code (\d+)", error_log_msg)
+
+                match = re.search(r"status code (\\d+)", error_log_msg)
                 if match:
                     status_code = int(match.group(1))
                 else:
-                    status_code = 500
+                    if isinstance(e, asyncio.TimeoutError):
+                        status_code = 408
+                    else:
+                        status_code = 500
 
-                # Log error to error log table
                 await add_error_log(
                     gemini_key=current_attempt_key,
                     model_name=model,
@@ -319,42 +460,48 @@ class OpenAIChatService:
                     request_msg=payload,
                 )
 
-                # Attempt to switch API Key
-                # Ensure key_manager is available (might need adjustment if not always passed)
                 if self.key_manager:
-                    api_key = await self.key_manager.handle_api_failure(
+                    new_api_key = await self.key_manager.handle_api_failure(
                         current_attempt_key, retries
                     )
-                    if api_key:
-                        logger.info(f"Switched to new API key: {api_key}")
-                    else:
-                        logger.error(
-                            f"No valid API key available after {retries} retries."
+                    if new_api_key and new_api_key != current_attempt_key:
+                        final_api_key = new_api_key
+                        logger.info(
+                            f"Switched to new API key for next attempt: {final_api_key}"
                         )
-                        break 
+                    elif not new_api_key:
+                        logger.error(
+                            f"No valid API key available after {retries} retries, ceasing attempts for this request."
+                        )
+                        break
                 else:
-                    logger.error("KeyManager not available for retry logic.")
-                    break 
+                    logger.error(
+                        "KeyManager not available, cannot switch API key. Ceasing attempts for this request."
+                    )
+                    break
 
                 if retries >= max_retries:
-                    logger.error(f"Max retries ({max_retries}) reached for streaming.")
-                    break
+                    logger.error(
+                        f"Max retries ({max_retries}) reached for streaming model {model}."
+                    )
             finally:
-                # Log the final outcome of the streaming request
                 end_time = time.perf_counter()
                 latency_ms = int((end_time - start_time) * 1000)
                 await add_request_log(
                     model_name=model,
-                    api_key=final_api_key,
+                    api_key=current_attempt_key,
                     is_success=is_success,
                     status_code=status_code,
                     latency_ms=latency_ms,
                     request_time=request_datetime,
                 )
-                # If the loop finished due to failure, yield error and DONE
-                if not is_success and retries >= max_retries:
-                    yield f"data: {json.dumps({'error': 'Streaming failed after retries'})}\n\n"
-                    yield "data: [DONE]\n\n"
+
+        if not is_success:
+            logger.error(
+                f"Streaming failed permanently for model {model} after {retries} attempts."
+            )
+            yield f"data: {json.dumps({'error': f'Streaming failed after {retries} retries.'})}\n\n"
+            yield "data: [DONE]\n\n"
 
     async def create_image_chat_completion(
         self, request: ChatRequest, api_key: str
@@ -382,7 +529,7 @@ class OpenAIChatService:
         start_time = time.perf_counter()
         request_datetime = datetime.datetime.now()
         is_success = False
-        status_code = None 
+        status_code = None
 
         try:
             if image_data:
@@ -416,16 +563,14 @@ class OpenAIChatService:
             is_success = False
             error_log_msg = f"Stream image completion failed for model {model}: {e}"
             logger.error(error_log_msg)
-            status_code = 500 
+            status_code = 500
             await add_error_log(
                 gemini_key=api_key,
                 model_name=model,
                 error_type="openai-image-stream",
                 error_log=error_log_msg,
                 error_code=status_code,
-                request_msg={
-                    "image_data_truncated": image_data[:1000]
-                }, 
+                request_msg={"image_data_truncated": image_data[:1000]},
             )
             yield f"data: {json.dumps({'error': error_log_msg})}\n\n"
             yield "data: [DONE]\n\n"
@@ -445,13 +590,13 @@ class OpenAIChatService:
             )
 
     async def _handle_normal_image_completion(
-        self, model: str, image_data: str, api_key: str 
+        self, model: str, image_data: str, api_key: str
     ) -> Dict[str, Any]:
         logger.info(f"Starting normal image completion for model: {model}")
         start_time = time.perf_counter()
-        request_datetime = datetime.datetime.now() 
+        request_datetime = datetime.datetime.now()
         is_success = False
-        status_code = None 
+        status_code = None
         result = None
 
         try:
@@ -475,11 +620,8 @@ class OpenAIChatService:
                 error_type="openai-image-non-stream",
                 error_log=error_log_msg,
                 error_code=status_code,
-                request_msg={
-                    "image_data_truncated": image_data[:1000]
-                },
+                request_msg={"image_data_truncated": image_data[:1000]},
             )
-            # Re-raise the exception so the caller knows about the failure
             raise e
         finally:
             end_time = time.perf_counter()

app/service/chat/vertex_express_chat_service.py

@@ -0,0 +1,328 @@
+# app/services/chat_service.py
+
+import json
+import re
+import datetime
+import time
+from typing import Any, AsyncGenerator, Dict, List
+from app.config.config import settings
+from app.core.constants import GEMINI_2_FLASH_EXP_SAFETY_SETTINGS
+from app.domain.gemini_models import GeminiRequest
+from app.handler.response_handler import GeminiResponseHandler
+from app.handler.stream_optimizer import gemini_optimizer
+from app.log.logger import get_gemini_logger
+from app.service.client.api_client import GeminiApiClient
+from app.service.key.key_manager import KeyManager
+from app.database.services import add_error_log, add_request_log
+
+logger = get_gemini_logger()
+
+
+def _has_image_parts(contents: List[Dict[str, Any]]) -> bool:
+    """判断消息是否包含图片部分"""
+    for content in contents:
+        if "parts" in content:
+            for part in content["parts"]:
+                if "image_url" in part or "inline_data" in part:
+                    return True
+    return False
+
+
+def _clean_json_schema_properties(obj: Any) -> Any:
+    """清理JSON Schema中Gemini API不支持的字段"""
+    if not isinstance(obj, dict):
+        return obj
+    
+    # Gemini API不支持的JSON Schema字段
+    unsupported_fields = {
+        "exclusiveMaximum", "exclusiveMinimum", "const", "examples", 
+        "contentEncoding", "contentMediaType", "if", "then", "else",
+        "allOf", "anyOf", "oneOf", "not", "definitions", "$schema",
+        "$id", "$ref", "$comment", "readOnly", "writeOnly"
+    }
+    
+    cleaned = {}
+    for key, value in obj.items():
+        if key in unsupported_fields:
+            continue
+        if isinstance(value, dict):
+            cleaned[key] = _clean_json_schema_properties(value)
+        elif isinstance(value, list):
+            cleaned[key] = [_clean_json_schema_properties(item) for item in value]
+        else:
+            cleaned[key] = value
+    
+    return cleaned
+
+
+def _build_tools(model: str, payload: Dict[str, Any]) -> List[Dict[str, Any]]:
+    """构建工具"""
+    
+    def _merge_tools(tools: List[Dict[str, Any]]) -> Dict[str, Any]:
+        record = dict()
+        for item in tools:
+            if not item or not isinstance(item, dict):
+                continue
+
+            for k, v in item.items():
+                if k == "functionDeclarations" and v and isinstance(v, list):
+                    functions = record.get("functionDeclarations", [])
+                    # 清理每个函数声明中的不支持字段
+                    cleaned_functions = []
+                    for func in v:
+                        if isinstance(func, dict):
+                            cleaned_func = _clean_json_schema_properties(func)
+                            cleaned_functions.append(cleaned_func)
+                        else:
+                            cleaned_functions.append(func)
+                    functions.extend(cleaned_functions)
+                    record["functionDeclarations"] = functions
+                else:
+                    record[k] = v
+        return record
+
+    tool = dict()
+    if payload and isinstance(payload, dict) and "tools" in payload:
+        if payload.get("tools") and isinstance(payload.get("tools"), dict):
+            payload["tools"] = [payload.get("tools")]
+        items = payload.get("tools", [])
+        if items and isinstance(items, list):
+            tool.update(_merge_tools(items))
+
+    if (
+        settings.TOOLS_CODE_EXECUTION_ENABLED
+        and not (model.endswith("-search") or "-thinking" in model)
+        and not _has_image_parts(payload.get("contents", []))
+    ):
+        tool["codeExecution"] = {}
+    if model.endswith("-search"):
+        tool["googleSearch"] = {}
+
+    # 解决 "Tool use with function calling is unsupported" 问题
+    if tool.get("functionDeclarations"):
+        tool.pop("googleSearch", None)
+        tool.pop("codeExecution", None)
+
+    return [tool] if tool else []
+
+
+def _get_safety_settings(model: str) -> List[Dict[str, str]]:
+    """获取安全设置"""
+    if model == "gemini-2.0-flash-exp":
+        return GEMINI_2_FLASH_EXP_SAFETY_SETTINGS
+    return settings.SAFETY_SETTINGS
+
+
+def _build_payload(model: str, request: GeminiRequest) -> Dict[str, Any]:
+    """构建请求payload"""
+    request_dict = request.model_dump()
+    if request.generationConfig:
+        if request.generationConfig.maxOutputTokens is None:
+            # 如果未指定最大输出长度，则不传递该字段，解决截断的问题
+            request_dict["generationConfig"].pop("maxOutputTokens")
+    
+    payload = {
+        "contents": request_dict.get("contents", []),
+        "tools": _build_tools(model, request_dict),
+        "safetySettings": _get_safety_settings(model),
+        "generationConfig": request_dict.get("generationConfig"),
+        "systemInstruction": request_dict.get("systemInstruction"),
+    }
+
+    if model.endswith("-image") or model.endswith("-image-generation"):
+        payload.pop("systemInstruction")
+        payload["generationConfig"]["responseModalities"] = ["Text", "Image"]
+        
+    # 处理思考配置：优先使用客户端提供的配置，否则使用默认配置
+    client_thinking_config = None
+    if request.generationConfig and request.generationConfig.thinkingConfig:
+        client_thinking_config = request.generationConfig.thinkingConfig
+    
+    if client_thinking_config is not None:
+        # 客户端提供了思考配置，直接使用
+        payload["generationConfig"]["thinkingConfig"] = client_thinking_config
+    else:
+        # 客户端没有提供思考配置，使用默认配置    
+        if model.endswith("-non-thinking"):
+            payload["generationConfig"]["thinkingConfig"] = {"thinkingBudget": 0} 
+        elif model in settings.THINKING_BUDGET_MAP:
+            if settings.SHOW_THINKING_PROCESS:
+                payload["generationConfig"]["thinkingConfig"] = {
+                    "thinkingBudget": settings.THINKING_BUDGET_MAP.get(model,1000),
+                    "includeThoughts": True
+                }
+            else:
+                payload["generationConfig"]["thinkingConfig"] = {"thinkingBudget": settings.THINKING_BUDGET_MAP.get(model,1000)}
+
+    return payload
+
+
+class GeminiChatService:
+    """聊天服务"""
+
+    def __init__(self, base_url: str, key_manager: KeyManager):
+        self.api_client = GeminiApiClient(base_url, settings.TIME_OUT)
+        self.key_manager = key_manager
+        self.response_handler = GeminiResponseHandler()
+
+    def _extract_text_from_response(self, response: Dict[str, Any]) -> str:
+        """从响应中提取文本内容"""
+        if not response.get("candidates"):
+            return ""
+
+        candidate = response["candidates"][0]
+        content = candidate.get("content", {})
+        parts = content.get("parts", [])
+
+        if parts and "text" in parts[0]:
+            return parts[0].get("text", "")
+        return ""
+
+    def _create_char_response(
+        self, original_response: Dict[str, Any], text: str
+    ) -> Dict[str, Any]:
+        """创建包含指定文本的响应"""
+        response_copy = json.loads(json.dumps(original_response))  # 深拷贝
+        if response_copy.get("candidates") and response_copy["candidates"][0].get(
+            "content", {}
+        ).get("parts"):
+            response_copy["candidates"][0]["content"]["parts"][0]["text"] = text
+        return response_copy
+
+    async def generate_content(
+        self, model: str, request: GeminiRequest, api_key: str
+    ) -> Dict[str, Any]:
+        """生成内容"""
+        payload = _build_payload(model, request)
+        start_time = time.perf_counter()
+        request_datetime = datetime.datetime.now()
+        is_success = False
+        status_code = None
+        response = None
+
+        try:
+            response = await self.api_client.generate_content(payload, model, api_key)
+            is_success = True
+            status_code = 200
+            return self.response_handler.handle_response(response, model, stream=False)
+        except Exception as e:
+            is_success = False
+            error_log_msg = str(e)
+            logger.error(f"Normal API call failed with error: {error_log_msg}")
+            match = re.search(r"status code (\d+)", error_log_msg)
+            if match:
+                status_code = int(match.group(1))
+            else:
+                status_code = 500
+
+            await add_error_log(
+                gemini_key=api_key,
+                model_name=model,
+                error_type="gemini-chat-non-stream",
+                error_log=error_log_msg,
+                error_code=status_code,
+                request_msg=payload
+            )
+            raise e
+        finally:
+            end_time = time.perf_counter()
+            latency_ms = int((end_time - start_time) * 1000)
+            await add_request_log(
+                model_name=model,
+                api_key=api_key,
+                is_success=is_success,
+                status_code=status_code,
+                latency_ms=latency_ms,
+                request_time=request_datetime
+            )
+
+    async def stream_generate_content(
+        self, model: str, request: GeminiRequest, api_key: str
+    ) -> AsyncGenerator[str, None]:
+        """流式生成内容"""
+        retries = 0
+        max_retries = settings.MAX_RETRIES
+        payload = _build_payload(model, request)
+        is_success = False
+        status_code = None
+        final_api_key = api_key
+
+        while retries < max_retries:
+            request_datetime = datetime.datetime.now()
+            start_time = time.perf_counter()
+            current_attempt_key = api_key
+            final_api_key = current_attempt_key # Update final key used
+            try:
+                async for line in self.api_client.stream_generate_content(
+                    payload, model, current_attempt_key
+                ):
+                    # print(line)
+                    if line.startswith("data:"):
+                        line = line[6:]
+                        response_data = self.response_handler.handle_response(
+                            json.loads(line), model, stream=True
+                        )
+                        text = self._extract_text_from_response(response_data)
+                        # 如果有文本内容，且开启了流式输出优化器，则使用流式输出优化器处理
+                        if text and settings.STREAM_OPTIMIZER_ENABLED:
+                            # 使用流式输出优化器处理文本输出
+                            async for (
+                                optimized_chunk
+                            ) in gemini_optimizer.optimize_stream_output(
+                                text,
+                                lambda t: self._create_char_response(response_data, t),
+                                lambda c: "data: " + json.dumps(c) + "\n\n",
+                            ):
+                                yield optimized_chunk
+                        else:
+                            # 如果没有文本内容（如工具调用等），整块输出
+                            yield "data: " + json.dumps(response_data) + "\n\n"
+                logger.info("Streaming completed successfully")
+                is_success = True
+                status_code = 200
+                break
+            except Exception as e:
+                retries += 1
+                is_success = False
+                error_log_msg = str(e)
+                logger.warning(
+                    f"Streaming API call failed with error: {error_log_msg}. Attempt {retries} of {max_retries}"
+                )
+                match = re.search(r"status code (\d+)", error_log_msg)
+                if match:
+                    status_code = int(match.group(1))
+                else:
+                    status_code = 500
+
+                await add_error_log(
+                    gemini_key=current_attempt_key,
+                    model_name=model,
+                    error_type="gemini-chat-stream",
+                    error_log=error_log_msg,
+                    error_code=status_code,
+                    request_msg=payload
+                )
+
+                api_key = await self.key_manager.handle_api_failure(current_attempt_key, retries)
+                if api_key:
+                    logger.info(f"Switched to new API key: {api_key}")
+                else:
+                    logger.error(f"No valid API key available after {retries} retries.")
+                    break
+
+                if retries >= max_retries:
+                    logger.error(
+                        f"Max retries ({max_retries}) reached for streaming."
+                    )
+                    break
+            finally:
+                end_time = time.perf_counter()
+                latency_ms = int((end_time - start_time) * 1000)
+                await add_request_log(
+                    model_name=model,
+                    api_key=final_api_key,
+                    is_success=is_success,
+                    status_code=status_code,
+                    latency_ms=latency_ms,
+                    request_time=request_datetime
+                )

app/service/client/api_client.py

@@ -40,29 +40,38 @@ class GeminiApiClient(ApiClient):
             model = model[:-20]
         return model
 
+    def _prepare_headers(self) -> Dict[str, str]:
+        headers = {}
+        if settings.CUSTOM_HEADERS:
+            headers.update(settings.CUSTOM_HEADERS)
+            logger.info(f"Using custom headers: {settings.CUSTOM_HEADERS}")
+        return headers
+
     async def get_models(self, api_key: str) -> Optional[Dict[str, Any]]:
         """获取可用的 Gemini 模型列表"""
         timeout = httpx.Timeout(timeout=5)
         
         proxy_to_use = None
         if settings.PROXIES:
-            proxy_to_use = random.choice(settings.PROXIES)
+            if settings.PROXIES_USE_CONSISTENCY_HASH_BY_API_KEY:
+                proxy_to_use = settings.PROXIES[hash(api_key) % len(settings.PROXIES)]
+            else:
+                proxy_to_use = random.choice(settings.PROXIES)
             logger.info(f"Using proxy for getting models: {proxy_to_use}")
 
+        headers = self._prepare_headers()
         async with httpx.AsyncClient(timeout=timeout, proxy=proxy_to_use) as client:
-            url = f"{self.base_url}/models?key={api_key}"
+            url = f"{self.base_url}/models?key={api_key}&pageSize=1000"
             try:
-                response = await client.get(url)
-                response.raise_for_status()  # 如果状态码不是 2xx，则引发 HTTPStatusError
+                response = await client.get(url, headers=headers)
+                response.raise_for_status()
                 return response.json()
             except httpx.HTTPStatusError as e:
                 logger.error(f"获取模型列表失败: {e.response.status_code}")
                 logger.error(e.response.text)
-                # 返回 None 而不是抛出异常，以便上层处理
                 return None
             except httpx.RequestError as e:
                 logger.error(f"请求模型列表失败: {e}")
-                # 返回 None 而不是抛出异常
                 return None
             
     async def generate_content(self, payload: Dict[str, Any], model: str, api_key: str) -> Dict[str, Any]:
@@ -71,12 +80,16 @@ class GeminiApiClient(ApiClient):
 
         proxy_to_use = None
         if settings.PROXIES:
-            proxy_to_use = random.choice(settings.PROXIES)
-            logger.info(f"Using proxy: {proxy_to_use}")
+            if settings.PROXIES_USE_CONSISTENCY_HASH_BY_API_KEY:
+                proxy_to_use = settings.PROXIES[hash(api_key) % len(settings.PROXIES)]
+            else:
+                proxy_to_use = random.choice(settings.PROXIES)
+            logger.info(f"Using proxy for getting models: {proxy_to_use}")
             
+        headers = self._prepare_headers()
         async with httpx.AsyncClient(timeout=timeout, proxy=proxy_to_use) as client:
             url = f"{self.base_url}/models/{model}:generateContent?key={api_key}"
-            response = await client.post(url, json=payload)
+            response = await client.post(url, json=payload, headers=headers)
             if response.status_code != 200:
                 error_content = response.text
                 raise Exception(f"API call failed with status code {response.status_code}, {error_content}")
@@ -88,12 +101,16 @@ class GeminiApiClient(ApiClient):
         
         proxy_to_use = None
         if settings.PROXIES:
-            proxy_to_use = random.choice(settings.PROXIES)
-            logger.info(f"Using proxy: {proxy_to_use}")
+            if settings.PROXIES_USE_CONSISTENCY_HASH_BY_API_KEY:
+                proxy_to_use = settings.PROXIES[hash(api_key) % len(settings.PROXIES)]
+            else:
+                proxy_to_use = random.choice(settings.PROXIES)
+            logger.info(f"Using proxy for getting models: {proxy_to_use}")
 
+        headers = self._prepare_headers()
         async with httpx.AsyncClient(timeout=timeout, proxy=proxy_to_use) as client:
             url = f"{self.base_url}/models/{model}:streamGenerateContent?alt=sse&key={api_key}"
-            async with client.stream(method="POST", url=url, json=payload) as response:
+            async with client.stream(method="POST", url=url, json=payload, headers=headers) as response:
                 if response.status_code != 200:
                     error_content = await response.aread()
                     error_msg = error_content.decode("utf-8")
@@ -101,6 +118,27 @@ class GeminiApiClient(ApiClient):
                 async for line in response.aiter_lines():
                     yield line
 
+    async def count_tokens(self, payload: Dict[str, Any], model: str, api_key: str) -> Dict[str, Any]:
+        timeout = httpx.Timeout(self.timeout, read=self.timeout)
+        model = self._get_real_model(model)
+
+        proxy_to_use = None
+        if settings.PROXIES:
+            if settings.PROXIES_USE_CONSISTENCY_HASH_BY_API_KEY:
+                proxy_to_use = settings.PROXIES[hash(api_key) % len(settings.PROXIES)]
+            else:
+                proxy_to_use = random.choice(settings.PROXIES)
+            logger.info(f"Using proxy for counting tokens: {proxy_to_use}")
+
+        headers = self._prepare_headers()
+        async with httpx.AsyncClient(timeout=timeout, proxy=proxy_to_use) as client:
+            url = f"{self.base_url}/models/{model}:countTokens?key={api_key}"
+            response = await client.post(url, json=payload, headers=headers)
+            if response.status_code != 200:
+                error_content = response.text
+                raise Exception(f"API call failed with status code {response.status_code}, {error_content}")
+            return response.json()
+
 
 class OpenaiApiClient(ApiClient):
     """OpenAI API客户端"""
@@ -109,11 +147,27 @@ class OpenaiApiClient(ApiClient):
         self.base_url = base_url
         self.timeout = timeout
         
+    def _prepare_headers(self, api_key: str) -> Dict[str, str]:
+        headers = {"Authorization": f"Bearer {api_key}"}
+        if settings.CUSTOM_HEADERS:
+            headers.update(settings.CUSTOM_HEADERS)
+            logger.info(f"Using custom headers: {settings.CUSTOM_HEADERS}")
+        return headers
+
     async def get_models(self, api_key: str) -> Dict[str, Any]:
         timeout = httpx.Timeout(self.timeout, read=self.timeout)
-        async with httpx.AsyncClient(timeout=timeout) as client:
+
+        proxy_to_use = None
+        if settings.PROXIES:
+            if settings.PROXIES_USE_CONSISTENCY_HASH_BY_API_KEY:
+                proxy_to_use = settings.PROXIES[hash(api_key) % len(settings.PROXIES)]
+            else:
+                proxy_to_use = random.choice(settings.PROXIES)
+            logger.info(f"Using proxy for getting models: {proxy_to_use}")
+
+        headers = self._prepare_headers(api_key)
+        async with httpx.AsyncClient(timeout=timeout, proxy=proxy_to_use) as client:
             url = f"{self.base_url}/openai/models"
-            headers = {"Authorization": f"Bearer {api_key}"}
             response = await client.get(url, headers=headers)
             if response.status_code != 200:
                 error_content = response.text
@@ -122,15 +176,18 @@ class OpenaiApiClient(ApiClient):
 
     async def generate_content(self, payload: Dict[str, Any], api_key: str) -> Dict[str, Any]:
         timeout = httpx.Timeout(self.timeout, read=self.timeout)
-        
+        logger.info(f"settings.PROXIES_USE_CONSISTENCY_HASH_BY_API_KEY: {settings.PROXIES_USE_CONSISTENCY_HASH_BY_API_KEY}")
         proxy_to_use = None
         if settings.PROXIES:
-            proxy_to_use = random.choice(settings.PROXIES)
-            logger.info(f"Using proxy: {proxy_to_use}")
+            if settings.PROXIES_USE_CONSISTENCY_HASH_BY_API_KEY:
+                proxy_to_use = settings.PROXIES[hash(api_key) % len(settings.PROXIES)]
+            else:
+                proxy_to_use = random.choice(settings.PROXIES)
+            logger.info(f"Using proxy for getting models: {proxy_to_use}")
 
+        headers = self._prepare_headers(api_key)
         async with httpx.AsyncClient(timeout=timeout, proxy=proxy_to_use) as client:
             url = f"{self.base_url}/openai/chat/completions"
-            headers = {"Authorization": f"Bearer {api_key}"}
             response = await client.post(url, json=payload, headers=headers)
             if response.status_code != 200:
                 error_content = response.text
@@ -139,15 +196,17 @@ class OpenaiApiClient(ApiClient):
 
     async def stream_generate_content(self, payload: Dict[str, Any], api_key: str) -> AsyncGenerator[str, None]:
         timeout = httpx.Timeout(self.timeout, read=self.timeout)
-        
         proxy_to_use = None
         if settings.PROXIES:
-            proxy_to_use = random.choice(settings.PROXIES)
-            logger.info(f"Using proxy: {proxy_to_use}")
+            if settings.PROXIES_USE_CONSISTENCY_HASH_BY_API_KEY:
+                proxy_to_use = settings.PROXIES[hash(api_key) % len(settings.PROXIES)]
+            else:
+                proxy_to_use = random.choice(settings.PROXIES)
+            logger.info(f"Using proxy for getting models: {proxy_to_use}")
 
+        headers = self._prepare_headers(api_key)
         async with httpx.AsyncClient(timeout=timeout, proxy=proxy_to_use) as client:
             url = f"{self.base_url}/openai/chat/completions"
-            headers = {"Authorization": f"Bearer {api_key}"}
             async with client.stream(method="POST", url=url, json=payload, headers=headers) as response:
                 if response.status_code != 200:
                     error_content = await response.aread()
@@ -161,12 +220,15 @@ class OpenaiApiClient(ApiClient):
         
         proxy_to_use = None
         if settings.PROXIES:
-            proxy_to_use = random.choice(settings.PROXIES)
-            logger.info(f"Using proxy: {proxy_to_use}")
+            if settings.PROXIES_USE_CONSISTENCY_HASH_BY_API_KEY:
+                proxy_to_use = settings.PROXIES[hash(api_key) % len(settings.PROXIES)]
+            else:
+                proxy_to_use = random.choice(settings.PROXIES)
+            logger.info(f"Using proxy for getting models: {proxy_to_use}")
 
+        headers = self._prepare_headers(api_key)
         async with httpx.AsyncClient(timeout=timeout, proxy=proxy_to_use) as client:
             url = f"{self.base_url}/openai/embeddings"
-            headers = {"Authorization": f"Bearer {api_key}"}
             payload = {
                 "input": input,
                 "model": model,
@@ -179,15 +241,18 @@ class OpenaiApiClient(ApiClient):
                     
     async def generate_images(self, payload: Dict[str, Any], api_key: str) -> Dict[str, Any]:
         timeout = httpx.Timeout(self.timeout, read=self.timeout)
-        
+
         proxy_to_use = None
         if settings.PROXIES:
-            proxy_to_use = random.choice(settings.PROXIES)
-            logger.info(f"Using proxy: {proxy_to_use}")
+            if settings.PROXIES_USE_CONSISTENCY_HASH_BY_API_KEY:
+                proxy_to_use = settings.PROXIES[hash(api_key) % len(settings.PROXIES)]
+            else:
+                proxy_to_use = random.choice(settings.PROXIES)
+            logger.info(f"Using proxy for getting models: {proxy_to_use}")
 
+        headers = self._prepare_headers(api_key)
         async with httpx.AsyncClient(timeout=timeout, proxy=proxy_to_use) as client:
             url = f"{self.base_url}/openai/images/generations"
-            headers = {"Authorization": f"Bearer {api_key}"}
             response = await client.post(url, json=payload, headers=headers)
             if response.status_code != 200:
                 error_content = response.text

app/service/config/config_service.py

@@ -1,41 +1,49 @@
 """
 配置服务模块
 """
+
 import datetime
 import json
 from typing import Any, Dict, List
 
 from dotenv import find_dotenv, load_dotenv
+from fastapi import HTTPException
 from sqlalchemy import insert, update
 
+from app.config.config import Settings as ConfigSettings
 from app.config.config import settings
 from app.database.connection import database
 from app.database.models import Settings
-from app.config.config import Settings as ConfigSettings
 from app.database.services import get_all_settings
-from app.service.key.key_manager import get_key_manager_instance, reset_key_manager_instance
 from app.log.logger import get_config_routes_logger
+from app.service.key.key_manager import (
+    get_key_manager_instance,
+    reset_key_manager_instance,
+)
+from app.service.model.model_service import ModelService
 
 logger = get_config_routes_logger()
 
 
 class ConfigService:
     """配置服务类，用于管理应用程序配置"""
-    
+
     @staticmethod
     async def get_config() -> Dict[str, Any]:
         return settings.model_dump()
-    
+
     @staticmethod
     async def update_config(config_data: Dict[str, Any]) -> Dict[str, Any]:
         for key, value in config_data.items():
             if hasattr(settings, key):
                 setattr(settings, key, value)
-                logger.debug(f"Updated setting in memory: {key}") 
-        
+                logger.debug(f"Updated setting in memory: {key}")
+
         # 获取现有设置
         existing_settings_raw: List[Dict[str, Any]] = await get_all_settings()
-        existing_settings_map: Dict[str, Dict[str, Any]] = {s['key']: s for s in existing_settings_raw}
+        existing_settings_map: Dict[str, Dict[str, Any]] = {
+            s["key"]: s for s in existing_settings_raw
+        }
         existing_keys = set(existing_settings_map.keys())
 
         settings_to_update: List[Dict[str, Any]] = []
@@ -47,7 +55,7 @@ class ConfigService:
             # 处理不同类型的值
             if isinstance(value, list):
                 db_value = json.dumps(value)
-            elif isinstance(value, dict): # 新增对 dict 类型的处理
+            elif isinstance(value, dict):
                 db_value = json.dumps(value)
             elif isinstance(value, bool):
                 db_value = str(value).lower()
@@ -55,24 +63,25 @@ class ConfigService:
                 db_value = str(value)
 
             # 仅当值发生变化时才更新
-            if key in existing_keys and existing_settings_map[key]['value'] == db_value:
-                continue 
+            if key in existing_keys and existing_settings_map[key]["value"] == db_value:
+                continue
 
-            description = f"{key}配置项" 
+            description = f"{key}配置项"
 
             data = {
-                'key': key,
-                'value': db_value,
-                'description': description,
-                'updated_at': now
+                "key": key,
+                "value": db_value,
+                "description": description,
+                "updated_at": now,
             }
 
             if key in existing_keys:
-                # Preserve original description if not explicitly provided
-                data['description'] = existing_settings_map[key].get('description', description)
+                data["description"] = existing_settings_map[key].get(
+                    "description", description
+                )
                 settings_to_update.append(data)
             else:
-                data['created_at'] = now
+                data["created_at"] = now
                 settings_to_insert.append(data)
 
         # 在事务中执行批量插入和更新
@@ -82,37 +91,109 @@ class ConfigService:
                     if settings_to_insert:
                         query_insert = insert(Settings).values(settings_to_insert)
                         await database.execute(query=query_insert)
-                        logger.info(f"Bulk inserted {len(settings_to_insert)} settings.")
+                        logger.info(
+                            f"Bulk inserted {len(settings_to_insert)} settings."
+                        )
 
                     if settings_to_update:
                         for setting_data in settings_to_update:
                             query_update = (
                                 update(Settings)
-                                .where(Settings.key == setting_data['key'])
+                                .where(Settings.key == setting_data["key"])
                                 .values(
-                                    value=setting_data['value'],
-                                    description=setting_data['description'],
-                                    updated_at=setting_data['updated_at']
+                                    value=setting_data["value"],
+                                    description=setting_data["description"],
+                                    updated_at=setting_data["updated_at"],
                                 )
                             )
                             await database.execute(query=query_update)
                         logger.info(f"Updated {len(settings_to_update)} settings.")
             except Exception as e:
                 logger.error(f"Failed to bulk update/insert settings: {str(e)}")
-                raise  # Re-raise the exception after logging
+                raise
 
         # 重置并重新初始化 KeyManager
         try:
             await reset_key_manager_instance()
-            await get_key_manager_instance(settings.API_KEYS)
+            await get_key_manager_instance(settings.API_KEYS, settings.VERTEX_API_KEYS)
             logger.info("KeyManager instance re-initialized with updated settings.")
         except Exception as e:
             logger.error(f"Failed to re-initialize KeyManager: {str(e)}")
-            # Decide if this error should prevent returning the updated config
-            # For now, we log the error and continue
 
         return await ConfigService.get_config()
-    
+
+    @staticmethod
+    async def delete_key(key_to_delete: str) -> Dict[str, Any]:
+        """删除单个API密钥"""
+        # 确保 settings.API_KEYS 是一个列表
+        if not isinstance(settings.API_KEYS, list):
+            settings.API_KEYS = []
+
+        original_keys_count = len(settings.API_KEYS)
+        # 创建一个不包含待删除密钥的新列表
+        updated_api_keys = [k for k in settings.API_KEYS if k != key_to_delete]
+
+        if len(updated_api_keys) < original_keys_count:
+            # 密钥已找到并从列表中移除
+            settings.API_KEYS = updated_api_keys  # 首先更新内存中的 settings
+            # 使用 update_config 持久化更改，它同时处理数据库和 KeyManager
+            await ConfigService.update_config({"API_KEYS": settings.API_KEYS})
+            logger.info(f"密钥 '{key_to_delete}' 已成功删除。")
+            return {"success": True, "message": f"密钥 '{key_to_delete}' 已成功删除。"}
+        else:
+            # 未找到密钥
+            logger.warning(f"尝试删除密钥 '{key_to_delete}'，但未找到该密钥。")
+            return {"success": False, "message": f"未找到密钥 '{key_to_delete}'。"}
+
+    @staticmethod
+    async def delete_selected_keys(keys_to_delete: List[str]) -> Dict[str, Any]:
+        """批量删除选定的API密钥"""
+        if not isinstance(settings.API_KEYS, list):
+            settings.API_KEYS = []
+
+        deleted_count = 0
+        not_found_keys: List[str] = []
+
+        current_api_keys = list(settings.API_KEYS)
+        keys_actually_removed: List[str] = []
+
+        for key_to_del in keys_to_delete:
+            if key_to_del in current_api_keys:
+                current_api_keys.remove(key_to_del)
+                keys_actually_removed.append(key_to_del)
+                deleted_count += 1
+            else:
+                not_found_keys.append(key_to_del)
+
+        if deleted_count > 0:
+            settings.API_KEYS = current_api_keys
+            await ConfigService.update_config({"API_KEYS": settings.API_KEYS})
+            logger.info(
+                f"成功删除 {deleted_count} 个密钥。密钥: {keys_actually_removed}"
+            )
+            message = f"成功删除 {deleted_count} 个密钥。"
+            if not_found_keys:
+                message += f" {len(not_found_keys)} 个密钥未找到: {not_found_keys}。"
+            return {
+                "success": True,
+                "message": message,
+                "deleted_count": deleted_count,
+                "not_found_keys": not_found_keys,
+            }
+        else:
+            message = "没有密钥被删除。"
+            if not_found_keys:
+                message = f"所有 {len(not_found_keys)} 个指定的密钥均未找到: {not_found_keys}。"
+            elif not keys_to_delete:
+                message = "未指定要删除的密钥。"
+            logger.warning(message)
+            return {
+                "success": False,
+                "message": message,
+                "deleted_count": 0,
+                "not_found_keys": not_found_keys,
+            }
+
     @staticmethod
     async def reset_config() -> Dict[str, Any]:
         """
@@ -124,7 +205,9 @@ class ConfigService:
         """
         # 1. 重新加载配置对象，它应该处理环境变量和 .env 的优先级
         _reload_settings()
-        logger.info("Settings object reloaded, prioritizing system environment variables then .env file.")
+        logger.info(
+            "Settings object reloaded, prioritizing system environment variables then .env file."
+        )
 
         # 2. 重置并重新初始化 KeyManager
         try:
@@ -140,6 +223,34 @@ class ConfigService:
         # 3. 返回更新后的配置
         return await ConfigService.get_config()
 
+    @staticmethod
+    async def fetch_ui_models() -> List[Dict[str, Any]]:
+        """获取用于UI显示的模型列表"""
+        try:
+            key_manager = await get_key_manager_instance()
+            model_service = ModelService()
+
+            api_key = await key_manager.get_first_valid_key()
+            if not api_key:
+                logger.error("No valid API keys available to fetch model list for UI.")
+                raise HTTPException(
+                    status_code=500,
+                    detail="No valid API keys available to fetch model list.",
+                )
+
+            models = await model_service.get_gemini_openai_models(api_key)
+            return models
+        except HTTPException as e:
+            raise e
+        except Exception as e:
+            logger.error(
+                f"Failed to fetch models for UI in ConfigService: {e}", exc_info=True
+            )
+            raise HTTPException(
+                status_code=500, detail=f"Failed to fetch models for UI: {str(e)}"
+            )
+
+
 # 重新加载配置的函数
 def _reload_settings():
     """重新加载环境变量并更新配置"""
@@ -147,4 +258,4 @@ def _reload_settings():
     load_dotenv(find_dotenv(), override=True)
     # 更新现有 settings 对象的属性，而不是新建实例
     for key, value in ConfigSettings().model_dump().items():
-        setattr(settings, key, value)
+        setattr(settings, key, value)

app/service/embedding/embedding_service.py

@@ -1,15 +1,15 @@
 import datetime
 import time
-import re # For potential status code parsing from generic errors
+import re
 from typing import List, Union
 
 import openai
-from openai import APIStatusError # Import specific error type
+from openai import APIStatusError
 from openai.types import CreateEmbeddingResponse
 
 from app.config.config import settings
 from app.log.logger import get_embeddings_logger
-from app.database.services import add_error_log, add_request_log # Import DB logging functions
+from app.database.services import add_error_log, add_request_log
 
 logger = get_embeddings_logger()
 
@@ -26,7 +26,6 @@ class EmbeddingService:
         status_code = None
         response = None
         error_log_msg = ""
-        # Prepare request message for logging (truncate if list or long string)
         if isinstance(input_text, list):
             request_msg_log = {"input_truncated": [str(item)[:100] + "..." if len(str(item)) > 100 else str(item) for item in input_text[:5]]}
             if len(input_text) > 5:
@@ -39,39 +38,36 @@ class EmbeddingService:
             client = openai.OpenAI(api_key=api_key, base_url=settings.BASE_URL)
             response = client.embeddings.create(input=input_text, model=model)
             is_success = True
-            status_code = 200 # Assume 200 OK on success
+            status_code = 200
             return response
         except APIStatusError as e:
             is_success = False
             status_code = e.status_code
             error_log_msg = f"OpenAI API error: {e}"
             logger.error(f"Error creating embedding (APIStatusError): {error_log_msg}")
-            raise e # Re-raise the specific error
+            raise e
         except Exception as e:
             is_success = False
             error_log_msg = f"Generic error: {e}"
             logger.error(f"Error creating embedding (Exception): {error_log_msg}")
-            # Try to parse status code from generic error (less reliable)
             match = re.search(r"status code (\d+)", str(e))
             if match:
                 status_code = int(match.group(1))
             else:
-                status_code = 500 # Default if parsing fails
-            raise e # Re-raise the generic error
+                status_code = 500
+            raise e
         finally:
             end_time = time.perf_counter()
             latency_ms = int((end_time - start_time) * 1000)
             if not is_success:
-                 # Log error to database if it failed
-                 await add_error_log(
-                     gemini_key=api_key, # Using gemini_key parameter name for consistency
-                     model_name=model,
-                     error_type="openai-embedding",
-                     error_log=error_log_msg,
-                     error_code=status_code,
-                     request_msg=request_msg_log
+                await add_error_log(
+                    gemini_key=api_key,
+                    model_name=model,
+                    error_type="openai-embedding",
+                    error_log=error_log_msg,
+                    error_code=status_code,
+                    request_msg=request_msg_log
                  )
-            # Log request outcome to database regardless of success/failure
             await add_request_log(
                 model_name=model,
                 api_key=api_key,

app/service/error_log/error_log_service.py

@@ -0,0 +1,178 @@
+from datetime import datetime, timedelta, timezone
+from typing import Any, Dict, List, Optional
+
+from sqlalchemy import delete, func, select
+
+from app.config.config import settings
+from app.database import services as db_services
+from app.database.connection import database
+from app.database.models import ErrorLog
+from app.log.logger import get_error_log_logger
+
+logger = get_error_log_logger()
+
+
+async def delete_old_error_logs():
+    """
+    Deletes error logs older than a specified number of days,
+    based on the AUTO_DELETE_ERROR_LOGS_ENABLED and AUTO_DELETE_ERROR_LOGS_DAYS settings.
+    """
+    if not settings.AUTO_DELETE_ERROR_LOGS_ENABLED:
+        logger.info("Auto-deletion of error logs is disabled. Skipping.")
+        return
+
+    days_to_keep = settings.AUTO_DELETE_ERROR_LOGS_DAYS
+    if not isinstance(days_to_keep, int) or days_to_keep <= 0:
+        logger.error(
+            f"Invalid AUTO_DELETE_ERROR_LOGS_DAYS value: {days_to_keep}. Must be a positive integer. Skipping deletion."
+        )
+        return
+
+    cutoff_date = datetime.now(timezone.utc) - timedelta(days=days_to_keep)
+
+    logger.info(
+        f"Attempting to delete error logs older than {days_to_keep} days (before {cutoff_date.strftime('%Y-%m-%d %H:%M:%S %Z')})."
+    )
+
+    try:
+        if not database.is_connected:
+            await database.connect()
+            logger.info("Database connection established for deleting error logs.")
+
+        # First, count how many logs will be deleted (optional, for logging)
+        count_query = select(func.count(ErrorLog.id)).where(
+            ErrorLog.request_time < cutoff_date
+        )
+        num_logs_to_delete = await database.fetch_val(count_query)
+
+        if num_logs_to_delete == 0:
+            logger.info(
+                "No error logs found older than the specified period. No deletion needed."
+            )
+            return
+
+        logger.info(f"Found {num_logs_to_delete} error logs to delete.")
+
+        # Perform the deletion
+        query = delete(ErrorLog).where(ErrorLog.request_time < cutoff_date)
+        await database.execute(query)
+        logger.info(
+            f"Successfully deleted {num_logs_to_delete} error logs older than {days_to_keep} days."
+        )
+
+    except Exception as e:
+        logger.error(
+            f"Error during automatic deletion of error logs: {e}", exc_info=True
+        )
+
+
+async def process_get_error_logs(
+    limit: int,
+    offset: int,
+    key_search: Optional[str],
+    error_search: Optional[str],
+    error_code_search: Optional[str],
+    start_date: Optional[datetime],
+    end_date: Optional[datetime],
+    sort_by: str,
+    sort_order: str,
+) -> Dict[str, Any]:
+    """
+    处理错误日志的检索，支持分页和过滤。
+    """
+    try:
+        logs_data = await db_services.get_error_logs(
+            limit=limit,
+            offset=offset,
+            key_search=key_search,
+            error_search=error_search,
+            error_code_search=error_code_search,
+            start_date=start_date,
+            end_date=end_date,
+            sort_by=sort_by,
+            sort_order=sort_order,
+        )
+        total_count = await db_services.get_error_logs_count(
+            key_search=key_search,
+            error_search=error_search,
+            error_code_search=error_code_search,
+            start_date=start_date,
+            end_date=end_date,
+        )
+        return {"logs": logs_data, "total": total_count}
+    except Exception as e:
+        logger.error(f"Service error in process_get_error_logs: {e}", exc_info=True)
+        raise
+
+
+async def process_get_error_log_details(log_id: int) -> Optional[Dict[str, Any]]:
+    """
+    处理特定错误日志详细信息的检索。
+    如果未找到，则返回 None。
+    """
+    try:
+        log_details = await db_services.get_error_log_details(log_id=log_id)
+        return log_details
+    except Exception as e:
+        logger.error(
+            f"Service error in process_get_error_log_details for ID {log_id}: {e}",
+            exc_info=True,
+        )
+        raise
+
+
+async def process_delete_error_logs_by_ids(log_ids: List[int]) -> int:
+    """
+    按 ID 批量删除错误日志。
+    返回尝试删除的日志数量。
+    """
+    if not log_ids:
+        return 0
+    try:
+        deleted_count = await db_services.delete_error_logs_by_ids(log_ids)
+        return deleted_count
+    except Exception as e:
+        logger.error(
+            f"Service error in process_delete_error_logs_by_ids for IDs {log_ids}: {e}",
+            exc_info=True,
+        )
+        raise
+
+
+async def process_delete_error_log_by_id(log_id: int) -> bool:
+    """
+    按 ID 删除单个错误日志。
+    如果删除成功（或找到日志并尝试删除），则返回 True，否则返回 False。
+    """
+    try:
+        success = await db_services.delete_error_log_by_id(log_id)
+        return success
+    except Exception as e:
+        logger.error(
+            f"Service error in process_delete_error_log_by_id for ID {log_id}: {e}",
+            exc_info=True,
+        )
+        raise
+
+
+async def process_delete_all_error_logs() -> int:
+    """
+    处理删除所有错误日志的请求。
+    返回删除的日志数量。
+    """
+    try:
+        if not database.is_connected:
+            await database.connect()
+            logger.info("Database connection established for deleting all error logs.")
+
+        deleted_count = await db_services.delete_all_error_logs()
+        logger.info(
+            f"Successfully processed request to delete all error logs. Count: {deleted_count}"
+        )
+        return deleted_count
+    except Exception as e:
+        logger.error(
+            f"Service error in process_delete_all_error_logs: {e}",
+            exc_info=True,
+        )
+        raise

app/service/image/image_create_service.py

@@ -121,6 +121,7 @@ class ImageCreateService:
                             provider=settings.UPLOAD_PROVIDER,
                             base_url=settings.CLOUDFLARE_IMGBED_URL,
                             auth_code=settings.CLOUDFLARE_IMGBED_AUTH_CODE,
+                            upload_folder=settings.CLOUDFLARE_IMGBED_UPLOAD_FOLDER,
                         )
                     else:
                         raise ValueError(
@@ -137,7 +138,7 @@ class ImageCreateService:
                     )
 
             response_data = {
-                "created": int(time.time()),  # Current timestamp
+                "created": int(time.time()),
                 "data": images_data,
             }
             return response_data

app/service/key/key_manager.py

@@ -1,7 +1,6 @@
 import asyncio
 from itertools import cycle
-from typing import Dict
-
+from typing import Dict, Union
 
 from app.config.config import settings
 from app.log.logger import get_key_manager_logger
@@ -10,12 +9,19 @@ logger = get_key_manager_logger()
 
 
 class KeyManager:
-    def __init__(self, api_keys: list):
+    def __init__(self, api_keys: list, vertex_api_keys: list):
         self.api_keys = api_keys
+        self.vertex_api_keys = vertex_api_keys
         self.key_cycle = cycle(api_keys)
+        self.vertex_key_cycle = cycle(vertex_api_keys)
         self.key_cycle_lock = asyncio.Lock()
+        self.vertex_key_cycle_lock = asyncio.Lock()
         self.failure_count_lock = asyncio.Lock()
+        self.vertex_failure_count_lock = asyncio.Lock()
         self.key_failure_counts: Dict[str, int] = {key: 0 for key in api_keys}
+        self.vertex_key_failure_counts: Dict[str, int] = {
+            key: 0 for key in vertex_api_keys
+        }
         self.MAX_FAILURES = settings.MAX_FAILURES
         self.paid_key = settings.PAID_KEY
 
@@ -27,17 +33,33 @@ class KeyManager:
         async with self.key_cycle_lock:
             return next(self.key_cycle)
 
+    async def get_next_vertex_key(self) -> str:
+        """获取下一个 Vertex Express API key"""
+        async with self.vertex_key_cycle_lock:
+            return next(self.vertex_key_cycle)
+
     async def is_key_valid(self, key: str) -> bool:
         """检查key是否有效"""
         async with self.failure_count_lock:
             return self.key_failure_counts[key] < self.MAX_FAILURES
 
+    async def is_vertex_key_valid(self, key: str) -> bool:
+        """检查 Vertex key 是否有效"""
+        async with self.vertex_failure_count_lock:
+            return self.vertex_key_failure_counts[key] < self.MAX_FAILURES
+
     async def reset_failure_counts(self):
         """重置所有key的失败计数"""
         async with self.failure_count_lock:
             for key in self.key_failure_counts:
                 self.key_failure_counts[key] = 0
-                
+
+    async def reset_vertex_failure_counts(self):
+        """重置所有 Vertex key 的失败计数"""
+        async with self.vertex_failure_count_lock:
+            for key in self.vertex_key_failure_counts:
+                self.vertex_key_failure_counts[key] = 0
+
     async def reset_key_failure_count(self, key: str) -> bool:
         """重置指定key的失败计数"""
         async with self.failure_count_lock:
@@ -45,7 +67,21 @@ class KeyManager:
                 self.key_failure_counts[key] = 0
                 logger.info(f"Reset failure count for key: {key}")
                 return True
-            logger.warning(f"Attempt to reset failure count for non-existent key: {key}")
+            logger.warning(
+                f"Attempt to reset failure count for non-existent key: {key}"
+            )
+            return False
+
+    async def reset_vertex_key_failure_count(self, key: str) -> bool:
+        """重置指定 Vertex key 的失败计数"""
+        async with self.vertex_failure_count_lock:
+            if key in self.vertex_key_failure_counts:
+                self.vertex_key_failure_counts[key] = 0
+                logger.info(f"Reset failure count for Vertex key: {key}")
+                return True
+            logger.warning(
+                f"Attempt to reset failure count for non-existent Vertex key: {key}"
+            )
             return False
 
     async def get_next_working_key(self) -> str:
@@ -59,10 +95,22 @@ class KeyManager:
 
             current_key = await self.get_next_key()
             if current_key == initial_key:
-                # await self.reset_failure_counts() 取消重置
                 return current_key
 
-    async def handle_api_failure(self, api_key: str,retries: int) -> str:
+    async def get_next_working_vertex_key(self) -> str:
+        """获取下一可用的 Vertex Express API key"""
+        initial_key = await self.get_next_vertex_key()
+        current_key = initial_key
+
+        while True:
+            if await self.is_vertex_key_valid(current_key):
+                return current_key
+
+            current_key = await self.get_next_vertex_key()
+            if current_key == initial_key:
+                return current_key
+
+    async def handle_api_failure(self, api_key: str, retries: int) -> str:
         """处理API调用失败"""
         async with self.failure_count_lock:
             self.key_failure_counts[api_key] += 1
@@ -72,13 +120,26 @@ class KeyManager:
                 )
         if retries < settings.MAX_RETRIES:
             return await self.get_next_working_key()
-        else: 
+        else:
             return ""
 
+    async def handle_vertex_api_failure(self, api_key: str, retries: int) -> str:
+        """处理 Vertex Express API 调用失败"""
+        async with self.vertex_failure_count_lock:
+            self.vertex_key_failure_counts[api_key] += 1
+            if self.vertex_key_failure_counts[api_key] >= self.MAX_FAILURES:
+                logger.warning(
+                    f"Vertex Express API key {api_key} has failed {self.MAX_FAILURES} times"
+                )
+
     def get_fail_count(self, key: str) -> int:
         """获取指定密钥的失败次数"""
         return self.key_failure_counts.get(key, 0)
 
+    def get_vertex_fail_count(self, key: str) -> int:
+        """获取指定 Vertex 密钥的失败次数"""
+        return self.vertex_key_failure_counts.get(key, 0)
+
     async def get_keys_by_status(self) -> dict:
         """获取分类后的API key列表，包括失败次数"""
         valid_keys = {}
@@ -94,40 +155,309 @@ class KeyManager:
 
         return {"valid_keys": valid_keys, "invalid_keys": invalid_keys}
 
+    async def get_vertex_keys_by_status(self) -> dict:
+        """获取分类后的 Vertex Express API key 列表，包括失败次数"""
+        valid_keys = {}
+        invalid_keys = {}
+
+        async with self.vertex_failure_count_lock:
+            for key in self.vertex_api_keys:
+                fail_count = self.vertex_key_failure_counts[key]
+                if fail_count < self.MAX_FAILURES:
+                    valid_keys[key] = fail_count
+                else:
+                    invalid_keys[key] = fail_count
+        return {"valid_keys": valid_keys, "invalid_keys": invalid_keys}
+
     async def get_first_valid_key(self) -> str:
         """获取第一个有效的API key"""
         async with self.failure_count_lock:
             for key in self.key_failure_counts:
                 if self.key_failure_counts[key] < self.MAX_FAILURES:
                     return key
+        if self.api_keys:
+            return self.api_keys[0]
+        if not self.api_keys:
+            logger.warning("API key list is empty, cannot get first valid key.")
+            return ""
         return self.api_keys[0]
 
+
 _singleton_instance = None
 _singleton_lock = asyncio.Lock()
+_preserved_failure_counts: Union[Dict[str, int], None] = None
+_preserved_vertex_failure_counts: Union[Dict[str, int], None] = None
+_preserved_old_api_keys_for_reset: Union[list, None] = None
+_preserved_vertex_old_api_keys_for_reset: Union[list, None] = None
+_preserved_next_key_in_cycle: Union[str, None] = None
+_preserved_vertex_next_key_in_cycle: Union[str, None] = None
 
 
-async def get_key_manager_instance(api_keys: list = None) -> KeyManager:
+async def get_key_manager_instance(
+    api_keys: list = None, vertex_api_keys: list = None
+) -> KeyManager:
     """
     获取 KeyManager 单例实例。
 
-    如果尚未创建实例，将使用提供的 api_keys 初始化 KeyManager。
+    如果尚未创建实例，将使用提供的 api_keys,vertex_api_keys 初始化 KeyManager。
     如果已创建实例，则忽略 api_keys 参数，返回现有单例。
+    如果在重置后调用，会尝试恢复之前的状态（失败计数、循环位置）。
     """
-    global _singleton_instance
+    global _singleton_instance, _preserved_failure_counts, _preserved_vertex_failure_counts, _preserved_old_api_keys_for_reset, _preserved_vertex_old_api_keys_for_reset, _preserved_next_key_in_cycle, _preserved_vertex_next_key_in_cycle
 
     async with _singleton_lock:
         if _singleton_instance is None:
             if api_keys is None:
-                raise ValueError("API keys are required to initialize the KeyManager")
-            _singleton_instance = KeyManager(api_keys)
-            logger.info("KeyManager instance created.")
+                raise ValueError(
+                    "API keys are required to initialize or re-initialize the KeyManager instance."
+                )
+            if vertex_api_keys is None:
+                raise ValueError(
+                    "Vertex Express API keys are required to initialize or re-initialize the KeyManager instance."
+                )
+
+            if not api_keys:
+                logger.warning(
+                    "Initializing KeyManager with an empty list of API keys."
+                )
+            if not vertex_api_keys:
+                logger.warning(
+                    "Initializing KeyManager with an empty list of Vertex Express API keys."
+                )
+
+            _singleton_instance = KeyManager(api_keys, vertex_api_keys)
+            logger.info(
+                f"KeyManager instance created/re-created with {len(api_keys)} API keys and {len(vertex_api_keys)} Vertex Express API keys."
+            )
+
+            # 1. 恢复失败计数
+            if _preserved_failure_counts:
+                current_failure_counts = {
+                    key: 0 for key in _singleton_instance.api_keys
+                }
+                for key, count in _preserved_failure_counts.items():
+                    if key in current_failure_counts:
+                        current_failure_counts[key] = count
+                _singleton_instance.key_failure_counts = current_failure_counts
+                logger.info("Inherited failure counts for applicable keys.")
+            _preserved_failure_counts = None
+
+            if _preserved_vertex_failure_counts:
+                current_vertex_failure_counts = {
+                    key: 0 for key in _singleton_instance.vertex_api_keys
+                }
+                for key, count in _preserved_vertex_failure_counts.items():
+                    if key in current_vertex_failure_counts:
+                        current_vertex_failure_counts[key] = count
+                _singleton_instance.vertex_key_failure_counts = (
+                    current_vertex_failure_counts
+                )
+                logger.info("Inherited failure counts for applicable Vertex keys.")
+            _preserved_vertex_failure_counts = None
+
+            # 2. 调整 key_cycle 的起始点
+            start_key_for_new_cycle = None
+            if (
+                _preserved_old_api_keys_for_reset
+                and _preserved_next_key_in_cycle
+                and _singleton_instance.api_keys
+            ):
+                try:
+                    start_idx_in_old = _preserved_old_api_keys_for_reset.index(
+                        _preserved_next_key_in_cycle
+                    )
+
+                    for i in range(len(_preserved_old_api_keys_for_reset)):
+                        current_old_key_idx = (start_idx_in_old + i) % len(
+                            _preserved_old_api_keys_for_reset
+                        )
+                        key_candidate = _preserved_old_api_keys_for_reset[
+                            current_old_key_idx
+                        ]
+                        if key_candidate in _singleton_instance.api_keys:
+                            start_key_for_new_cycle = key_candidate
+                            break
+                except ValueError:
+                    logger.warning(
+                        f"Preserved next key '{_preserved_next_key_in_cycle}' not found in preserved old API keys. "
+                        "New cycle will start from the beginning of the new list."
+                    )
+                except Exception as e:
+                    logger.error(
+                        f"Error determining start key for new cycle from preserved state: {e}. "
+                        "New cycle will start from the beginning."
+                    )
+
+            if start_key_for_new_cycle and _singleton_instance.api_keys:
+                try:
+                    target_idx = _singleton_instance.api_keys.index(
+                        start_key_for_new_cycle
+                    )
+                    for _ in range(target_idx):
+                        next(_singleton_instance.key_cycle)
+                    logger.info(
+                        f"Key cycle in new instance advanced. Next call to get_next_key() will yield: {start_key_for_new_cycle}"
+                    )
+                except ValueError:
+                    logger.warning(
+                        f"Determined start key '{start_key_for_new_cycle}' not found in new API keys during cycle advancement. "
+                        "New cycle will start from the beginning."
+                    )
+                except StopIteration:
+                    logger.error(
+                        "StopIteration while advancing key cycle, implies empty new API key list previously missed."
+                    )
+                except Exception as e:
+                    logger.error(
+                        f"Error advancing new key cycle: {e}. Cycle will start from beginning."
+                    )
+            else:
+                if _singleton_instance.api_keys:
+                    logger.info(
+                        "New key cycle will start from the beginning of the new API key list (no specific start key determined or needed)."
+                    )
+                else:
+                    logger.info(
+                        "New key cycle not applicable as the new API key list is empty."
+                    )
+
+            # 清理所有保存的状态
+            _preserved_old_api_keys_for_reset = None
+            _preserved_next_key_in_cycle = None
+
+            # 3. 调整 vertex_key_cycle 的起始点
+            start_key_for_new_vertex_cycle = None
+            if (
+                _preserved_vertex_old_api_keys_for_reset
+                and _preserved_vertex_next_key_in_cycle
+                and _singleton_instance.vertex_api_keys
+            ):
+                try:
+                    start_idx_in_old = _preserved_vertex_old_api_keys_for_reset.index(
+                        _preserved_vertex_next_key_in_cycle
+                    )
+
+                    for i in range(len(_preserved_vertex_old_api_keys_for_reset)):
+                        current_old_key_idx = (start_idx_in_old + i) % len(
+                            _preserved_vertex_old_api_keys_for_reset
+                        )
+                        key_candidate = _preserved_vertex_old_api_keys_for_reset[
+                            current_old_key_idx
+                        ]
+                        if key_candidate in _singleton_instance.vertex_api_keys:
+                            start_key_for_new_vertex_cycle = key_candidate
+                            break
+                except ValueError:
+                    logger.warning(
+                        f"Preserved next key '{_preserved_vertex_next_key_in_cycle}' not found in preserved old Vertex Express API keys. "
+                        "New cycle will start from the beginning of the new list."
+                    )
+                except Exception as e:
+                    logger.error(
+                        f"Error determining start key for new Vertex key cycle from preserved state: {e}. "
+                        "New cycle will start from the beginning."
+                    )
+
+            if start_key_for_new_vertex_cycle and _singleton_instance.vertex_api_keys:
+                try:
+                    target_idx = _singleton_instance.vertex_api_keys.index(
+                        start_key_for_new_vertex_cycle
+                    )
+                    for _ in range(target_idx):
+                        next(_singleton_instance.vertex_key_cycle)
+                    logger.info(
+                        f"Vertex key cycle in new instance advanced. Next call to get_next_vertex_key() will yield: {start_key_for_new_vertex_cycle}"
+                    )
+                except ValueError:
+                    logger.warning(
+                        f"Determined start key '{start_key_for_new_vertex_cycle}' not found in new Vertex Express API keys during cycle advancement. "
+                        "New cycle will start from the beginning."
+                    )
+                except StopIteration:
+                    logger.error(
+                        "StopIteration while advancing Vertex key cycle, implies empty new Vertex Express API key list previously missed."
+                    )
+                except Exception as e:
+                    logger.error(
+                        f"Error advancing new Vertex key cycle: {e}. Cycle will start from beginning."
+                    )
+            else:
+                if _singleton_instance.vertex_api_keys:
+                    logger.info(
+                        "New Vertex key cycle will start from the beginning of the new Vertex Express API key list (no specific start key determined or needed)."
+                    )
+                else:
+                    logger.info(
+                        "New Vertex key cycle not applicable as the new Vertex Express API key list is empty."
+                    )
+
+            # 清理所有保存的状态
+            _preserved_vertex_old_api_keys_for_reset = None
+            _preserved_vertex_next_key_in_cycle = None
+
         return _singleton_instance
 
-    
+
 async def reset_key_manager_instance():
-    """重置 KeyManager 单例实例"""
-    global _singleton_instance
+    """
+    重置 KeyManager 单例实例。
+    将保存当前实例的状态（失败计数、旧 API keys、下一个 key 提示）
+    以供下一次 get_key_manager_instance 调用时恢复。
+    """
+    global _singleton_instance, _preserved_failure_counts, _preserved_vertex_failure_counts, _preserved_old_api_keys_for_reset, _preserved_vertex_old_api_keys_for_reset, _preserved_next_key_in_cycle, _preserved_vertex_next_key_in_cycle
     async with _singleton_lock:
         if _singleton_instance:
+            # 1. 保存失败计数
+            _preserved_failure_counts = _singleton_instance.key_failure_counts.copy()
+            _preserved_vertex_failure_counts = (
+                _singleton_instance.vertex_key_failure_counts.copy()
+            )
+
+            # 2. 保存旧的 API keys 列表
+            _preserved_old_api_keys_for_reset = _singleton_instance.api_keys.copy()
+            _preserved_vertex_old_api_keys_for_reset = (
+                _singleton_instance.vertex_api_keys.copy()
+            )
+
+            # 3. 保存 key_cycle 的下一个 key 提示
+            try:
+                if _singleton_instance.api_keys:
+                    _preserved_next_key_in_cycle = (
+                        await _singleton_instance.get_next_key()
+                    )
+                else:
+                    _preserved_next_key_in_cycle = None
+            except StopIteration:
+                logger.warning(
+                    "Could not preserve next key hint: key cycle was empty or exhausted in old instance."
+                )
+                _preserved_next_key_in_cycle = None
+            except Exception as e:
+                logger.error(f"Error preserving next key hint during reset: {e}")
+                _preserved_next_key_in_cycle = None
+
+            # 4. 保存 vertex_key_cycle 的下一个 key 提示
+            try:
+                if _singleton_instance.vertex_api_keys:
+                    _preserved_vertex_next_key_in_cycle = (
+                        await _singleton_instance.get_next_vertex_key()
+                    )
+                else:
+                    _preserved_vertex_next_key_in_cycle = None
+            except StopIteration:
+                logger.warning(
+                    "Could not preserve next key hint: Vertex key cycle was empty or exhausted in old instance."
+                )
+                _preserved_vertex_next_key_in_cycle = None
+            except Exception as e:
+                logger.error(f"Error preserving next key hint during reset: {e}")
+                _preserved_vertex_next_key_in_cycle = None
+
             _singleton_instance = None
-            logger.info("KeyManager instance reset.")
+            logger.info(
+                "KeyManager instance has been reset. State (failure counts, old keys, next key hint) preserved for next instantiation."
+            )
+        else:
+            logger.info(
+                "KeyManager instance was not set (or already reset), no reset action performed."
+            )

app/service/model/model_service.py

@@ -10,8 +10,7 @@ logger = get_model_logger()
 
 class ModelService:
     async def get_gemini_models(self, api_key: str) -> Optional[Dict[str, Any]]:
-        """使用 GeminiApiClient 获取并过滤模型列表"""
-        api_client = GeminiApiClient(base_url=settings.BASE_URL) # 实例化客户端
+        api_client = GeminiApiClient(base_url=settings.BASE_URL)
         gemini_models = await api_client.get_models(api_key)
 
         if gemini_models is None:

app/service/openai_compatiable/openai_compatiable_service.py

@@ -79,7 +79,6 @@ class OpenAICompatiableService:
             is_success = False
             error_log_msg = str(e)
             logger.error(f"Normal API call failed with error: {error_log_msg}")
-            # Try to parse status code from exception
             match = re.search(r"status code (\d+)", error_log_msg)
             if match:
                 status_code = int(match.group(1))
@@ -132,7 +131,7 @@ class OpenAICompatiableService:
                 logger.info("Streaming completed successfully")
                 is_success = True
                 status_code = 200
-                break  # 成功后退出循环
+                break
             except Exception as e:
                 retries += 1
                 is_success = False
@@ -140,14 +139,12 @@ class OpenAICompatiableService:
                 logger.warning(
                     f"Streaming API call failed with error: {error_log_msg}. Attempt {retries} of {max_retries}"
                 )
-                # Parse error code for logging
                 match = re.search(r"status code (\d+)", error_log_msg)
                 if match:
                     status_code = int(match.group(1))
                 else:
                     status_code = 500
 
-                # Log error to error log table
                 await add_error_log(
                     gemini_key=current_attempt_key,
                     model_name=model,
@@ -157,8 +154,6 @@ class OpenAICompatiableService:
                     request_msg=payload,
                 )
 
-                # Attempt to switch API Key
-                # Ensure key_manager is available (might need adjustment if not always passed)
                 if self.key_manager:
                     api_key = await self.key_manager.handle_api_failure(
                         current_attempt_key, retries
@@ -178,7 +173,6 @@ class OpenAICompatiableService:
                     logger.error(f"Max retries ({max_retries}) reached for streaming.")
                     break
             finally:
-                # Log the final outcome of the streaming request
                 end_time = time.perf_counter()
                 latency_ms = int((end_time - start_time) * 1000)
                 await add_request_log(
@@ -189,7 +183,6 @@ class OpenAICompatiableService:
                     latency_ms=latency_ms,
                     request_time=request_datetime,
                 )
-                # If the loop finished due to failure, yield error and DONE
                 if not is_success and retries >= max_retries:
                     yield f"data: {json.dumps({'error': 'Streaming failed after retries'})}\n\n"
                     yield "data: [DONE]\n\n"

app/service/request_log/request_log_service.py

@@ -0,0 +1,50 @@
+"""
+Service for request log operations.
+"""
+
+from datetime import datetime, timedelta, timezone
+
+from sqlalchemy import delete
+
+from app.database.connection import database
+from app.config.config import settings
+from app.database.models import RequestLog
+from app.log.logger import get_request_log_logger
+
+logger = get_request_log_logger()
+
+
+async def delete_old_request_logs_task():
+    """
+    定时删除旧的请求日志。
+    """
+    if not settings.AUTO_DELETE_REQUEST_LOGS_ENABLED:
+        logger.info(
+            "Auto-delete for request logs is disabled by settings. Skipping task."
+        )
+        return
+
+    days_to_keep = settings.AUTO_DELETE_REQUEST_LOGS_DAYS
+    logger.info(
+        f"Starting scheduled task to delete old request logs older than {days_to_keep} days."
+    )
+
+    try:
+        cutoff_date = datetime.now(timezone.utc) - timedelta(days=days_to_keep)
+
+        query = delete(RequestLog).where(RequestLog.request_time < cutoff_date)
+
+        if not database.is_connected:
+            logger.info("Connecting to database for request log deletion.")
+            await database.connect()
+
+        result = await database.execute(query)
+        logger.info(
+            f"Request logs older than {cutoff_date} potentially deleted. Rows affected: {result.rowcount if result else 'N/A'}"
+        )
+
+    except Exception as e:
+        logger.error(
+            f"An error occurred during the scheduled request log deletion: {str(e)}",
+            exc_info=True,
+        )

app/service/stats/stats_service.py

@@ -1,7 +1,9 @@
 # app/service/stats_service.py
 
 import datetime
-from sqlalchemy import select, func
+from typing import Union
+
+from sqlalchemy import and_, case, func, or_, select
 
 from app.database.connection import database
 from app.database.models import RequestLog
@@ -13,66 +15,129 @@ logger = get_stats_logger()
 class StatsService:
     """Service class for handling statistics related operations."""
 
-    async def get_calls_in_last_seconds(self, seconds: int) -> int:
-        """获取过去 N 秒内的调用次数 (包括成功和失败)"""
+    async def get_calls_in_last_seconds(self, seconds: int) -> dict[str, int]:
+        """获取过去 N 秒内的调用次数 (总数、成功、失败)"""
         try:
             cutoff_time = datetime.datetime.now() - datetime.timedelta(seconds=seconds)
-            query = select(func.count(RequestLog.id)).where(
-                RequestLog.request_time >= cutoff_time
-            )
-            count_result = await database.fetch_one(query)
-            return count_result[0] if count_result else 0
+            query = select(
+                func.count(RequestLog.id).label("total"),
+                func.sum(
+                    case(
+                        (
+                            and_(
+                                RequestLog.status_code >= 200,
+                                RequestLog.status_code < 300,
+                            ),
+                            1,
+                        ),
+                        else_=0,
+                    )
+                ).label("success"),
+                func.sum(
+                    case(
+                        (
+                            or_(
+                                RequestLog.status_code < 200,
+                                RequestLog.status_code >= 300,
+                            ),
+                            1,
+                        ),
+                        (RequestLog.status_code is None, 1),
+                        else_=0,
+                    )
+                ).label("failure"),
+            ).where(RequestLog.request_time >= cutoff_time)
+            result = await database.fetch_one(query)
+            if result:
+                return {
+                    "total": result["total"] or 0,
+                    "success": result["success"] or 0,
+                    "failure": result["failure"] or 0,
+                }
+            return {"total": 0, "success": 0, "failure": 0}
         except Exception as e:
             logger.error(f"Failed to get calls in last {seconds} seconds: {e}")
-            return 0 # Return 0 on error
+            return {"total": 0, "success": 0, "failure": 0}
 
-    async def get_calls_in_last_minutes(self, minutes: int) -> int:
-        """获取过去 N 分钟内的调用次数 (包括成功和失败)"""
+    async def get_calls_in_last_minutes(self, minutes: int) -> dict[str, int]:
+        """获取过去 N 分钟内的调用次数 (总数、成功、失败)"""
         return await self.get_calls_in_last_seconds(minutes * 60)
 
-    async def get_calls_in_last_hours(self, hours: int) -> int:
-        """获取过去 N 小时内的调用次数 (包括成功和失败)"""
+    async def get_calls_in_last_hours(self, hours: int) -> dict[str, int]:
+        """获取过去 N 小时内的调用次数 (总数、成功、失败)"""
         return await self.get_calls_in_last_seconds(hours * 3600)
 
-    async def get_calls_in_current_month(self) -> int:
-        """获取当前自然月内的调用次数 (包括成功和失败)"""
+    async def get_calls_in_current_month(self) -> dict[str, int]:
+        """获取当前自然月内的调用次数 (总数、成功、失败)"""
         try:
             now = datetime.datetime.now()
-            start_of_month = now.replace(day=1, hour=0, minute=0, second=0, microsecond=0)
-            query = select(func.count(RequestLog.id)).where(
-                RequestLog.request_time >= start_of_month
+            start_of_month = now.replace(
+                day=1, hour=0, minute=0, second=0, microsecond=0
             )
-            count_result = await database.fetch_one(query)
-            return count_result[0] if count_result else 0
+            query = select(
+                func.count(RequestLog.id).label("total"),
+                func.sum(
+                    case(
+                        (
+                            and_(
+                                RequestLog.status_code >= 200,
+                                RequestLog.status_code < 300,
+                            ),
+                            1,
+                        ),
+                        else_=0,
+                    )
+                ).label("success"),
+                func.sum(
+                    case(
+                        (
+                            or_(
+                                RequestLog.status_code < 200,
+                                RequestLog.status_code >= 300,
+                            ),
+                            1,
+                        ),
+                        (RequestLog.status_code is None, 1),
+                        else_=0,
+                    )
+                ).label("failure"),
+            ).where(RequestLog.request_time >= start_of_month)
+            result = await database.fetch_one(query)
+            if result:
+                return {
+                    "total": result["total"] or 0,
+                    "success": result["success"] or 0,
+                    "failure": result["failure"] or 0,
+                }
+            return {"total": 0, "success": 0, "failure": 0}
         except Exception as e:
             logger.error(f"Failed to get calls in current month: {e}")
-            return 0 # Return 0 on error
+            return {"total": 0, "success": 0, "failure": 0}
 
     async def get_api_usage_stats(self) -> dict:
-        """获取所有需要的 API 使用统计数据"""
+        """获取所有需要的 API 使用统计数据 (总数、成功、失败)"""
         try:
-            calls_1m = await self.get_calls_in_last_minutes(1)
-            calls_1h = await self.get_calls_in_last_hours(1)
-            calls_24h = await self.get_calls_in_last_hours(24)
-            calls_month = await self.get_calls_in_current_month()
+            stats_1m = await self.get_calls_in_last_minutes(1)
+            stats_1h = await self.get_calls_in_last_hours(1)
+            stats_24h = await self.get_calls_in_last_hours(24)
+            stats_month = await self.get_calls_in_current_month()
 
             return {
-                "calls_1m": calls_1m,
-                "calls_1h": calls_1h,
-                "calls_24h": calls_24h,
-                "calls_month": calls_month,
+                "calls_1m": stats_1m,
+                "calls_1h": stats_1h,
+                "calls_24h": stats_24h,
+                "calls_month": stats_month,
             }
         except Exception as e:
             logger.error(f"Failed to get API usage stats: {e}")
-            # Return default values on error
+            default_stat = {"total": 0, "success": 0, "failure": 0}
             return {
-                "calls_1m": 0,
-                "calls_1h": 0,
-                "calls_24h": 0,
-                "calls_month": 0,
+                "calls_1m": default_stat.copy(),
+                "calls_1h": default_stat.copy(),
+                "calls_24h": default_stat.copy(),
+                "calls_month": default_stat.copy(),
             }
 
-
     async def get_api_call_details(self, period: str) -> list[dict]:
         """
         获取指定时间段内的 API 调用详情
@@ -87,48 +152,55 @@ class StatsService:
             ValueError: 如果 period 无效
         """
         now = datetime.datetime.now()
-        if period == '1m':
+        if period == "1m":
             start_time = now - datetime.timedelta(minutes=1)
-        elif period == '1h':
+        elif period == "1h":
             start_time = now - datetime.timedelta(hours=1)
-        elif period == '24h':
+        elif period == "24h":
             start_time = now - datetime.timedelta(hours=24)
         else:
             raise ValueError(f"无效的时间段标识: {period}")
 
         try:
-            query = select(
-                RequestLog.request_time.label("timestamp"),
-                RequestLog.api_key.label("key"),
-                RequestLog.model_name.label("model"),
-                RequestLog.status_code # We might need to map this to 'success'/'failure' later
-            ).where(
-                RequestLog.request_time >= start_time
-            ).order_by(RequestLog.request_time.desc()) # Order by most recent first
+            query = (
+                select(
+                    RequestLog.request_time.label("timestamp"),
+                    RequestLog.api_key.label("key"),
+                    RequestLog.model_name.label("model"),
+                    RequestLog.status_code,
+                )
+                .where(RequestLog.request_time >= start_time)
+                .order_by(RequestLog.request_time.desc())
+            )
 
             results = await database.fetch_all(query)
 
-            # Convert results to list of dicts and map status_code
             details = []
             for row in results:
-                status = 'failure' # 默认状态为 failure，如果 status_code 有效且在 200-299 范围内则更新为 success
-                if row['status_code'] is not None: # 检查 status_code 是否为空
-                    status = 'success' if 200 <= row['status_code'] < 300 else 'failure'
-                details.append({
-                    "timestamp": row['timestamp'].isoformat(), # Use ISO format for JS compatibility
-                    "key": row['key'],
-                    "model": row['model'],
-                    "status": status
-                })
-            logger.info(f"Retrieved {len(details)} API call details for period '{period}'")
+                status = "failure"
+                if row["status_code"] is not None:
+                    status = "success" if 200 <= row["status_code"] < 300 else "failure"
+                details.append(
+                    {
+                        "timestamp": row[
+                            "timestamp"
+                        ].isoformat(),
+                        "key": row["key"],
+                        "model": row["model"],
+                        "status": status,
+                    }
+                )
+            logger.info(
+                f"Retrieved {len(details)} API call details for period '{period}'"
+            )
             return details
 
         except Exception as e:
-            logger.error(f"Failed to get API call details for period '{period}': {e}")
-            # Re-raise the exception to be handled by the route
+            logger.error(
+                f"Failed to get API call details for period '{period}': {e}")
             raise
 
-    async def get_key_usage_details_last_24h(self, key: str) -> dict | None:
+    async def get_key_usage_details_last_24h(self, key: str) -> Union[dict, None]:
         """
         获取指定 API 密钥在过去 24 小时内按模型统计的调用次数。
 
@@ -140,35 +212,44 @@ class StatsService:
             如果查询出错或没有找到记录，可能返回 None 或空字典。
             Example: {"gemini-pro": 10, "gemini-1.5-pro-latest": 5}
         """
-        logger.info(f"Fetching usage details for key ending in ...{key[-4:]} for the last 24h.")
+        logger.info(
+            f"Fetching usage details for key ending in ...{key[-4:]} for the last 24h."
+        )
         cutoff_time = datetime.datetime.now() - datetime.timedelta(hours=24)
 
         try:
-            query = select(
-                RequestLog.model_name,
-                func.count(RequestLog.id).label("call_count")
-            ).where(
-                RequestLog.api_key == key,
-                RequestLog.request_time >= cutoff_time,
-                RequestLog.model_name.isnot(None) # Ensure model_name is not null
-            ).group_by(
-                RequestLog.model_name
-            ).order_by(
-                func.count(RequestLog.id).desc() # Order by count descending
+            query = (
+                select(
+                    RequestLog.model_name, func.count(
+                        RequestLog.id).label("call_count")
+                )
+                .where(
+                    RequestLog.api_key == key,
+                    RequestLog.request_time >= cutoff_time,
+                    RequestLog.model_name.isnot(None),
+                )
+                .group_by(RequestLog.model_name)
+                .order_by(func.count(RequestLog.id).desc())
             )
 
             results = await database.fetch_all(query)
 
             if not results:
-                logger.info(f"No usage details found for key ending in ...{key[-4:]} in the last 24h.")
-                return {} # Return empty dict if no records found
+                logger.info(
+                    f"No usage details found for key ending in ...{key[-4:]} in the last 24h."
+                )
+                return {}
 
-            usage_details = {row['model_name']: row['call_count'] for row in results}
-            logger.info(f"Successfully fetched usage details for key ending in ...{key[-4:]}: {usage_details}")
+            usage_details = {row["model_name"]: row["call_count"]
+                             for row in results}
+            logger.info(
+                f"Successfully fetched usage details for key ending in ...{key[-4:]}: {usage_details}"
+            )
             return usage_details
 
         except Exception as e:
-            logger.error(f"Failed to get key usage details for key ending in ...{key[-4:]}: {e}", exc_info=True)
-            # Depending on requirements, you might return None or raise the exception
-            # Raising allows the route handler to return a 500 error.
-            raise # Re-raise the exception
+            logger.error(
+                f"Failed to get key usage details for key ending in ...{key[-4:]}: {e}",
+                exc_info=True,
+            )
+            raise

app/service/tts/tts_service.py

@@ -0,0 +1,95 @@
+import datetime
+import io
+import re
+import time
+import wave
+from typing import Optional
+
+from google import genai
+
+from app.config.config import settings
+from app.core.constants import TTS_VOICE_NAMES
+from app.database.services import add_error_log, add_request_log
+from app.domain.openai_models import TTSRequest
+from app.log.logger import get_openai_logger
+
+logger = get_openai_logger()
+
+
+def _create_wav_file(audio_data: bytes) -> bytes:
+    """Creates a WAV file in memory from raw audio data."""
+    with io.BytesIO() as wav_file:
+        with wave.open(wav_file, "wb") as wf:
+            wf.setnchannels(1)  # Mono
+            wf.setsampwidth(2)  # 16-bit
+            wf.setframerate(24000)  # 24kHz sample rate
+            wf.writeframes(audio_data)
+        return wav_file.getvalue()
+
+
+class TTSService:
+    async def create_tts(self, request: TTSRequest, api_key: str) -> Optional[bytes]:
+        """
+        使用 Google Gemini SDK 创建音频。
+        """
+        start_time = time.perf_counter()
+        request_datetime = datetime.datetime.now()
+        is_success = False
+        status_code = None
+        response = None
+        error_log_msg = ""
+        try:
+            client = genai.Client(api_key=api_key)
+            response =await client.aio.models.generate_content(
+                model=settings.TTS_MODEL,
+                contents=f"Speak in a {settings.TTS_SPEED} speed voice: {request.input}",
+                config={
+                    "response_modalities": ["Audio"],
+                    "speech_config": {
+                        "voice_config": {
+                            "prebuilt_voice_config": {
+                                "voice_name": request.voice if request.voice in TTS_VOICE_NAMES else settings.TTS_VOICE_NAME
+                            }
+                        }
+                    },
+                },
+            )
+            if (
+                response.candidates
+                and response.candidates[0].content.parts
+                and response.candidates[0].content.parts[0].inline_data
+            ):
+                raw_audio_data = response.candidates[0].content.parts[0].inline_data.data
+                is_success = True
+                status_code = 200
+                return _create_wav_file(raw_audio_data)
+        except Exception as e:
+            is_success = False
+            error_log_msg = f"Generic error: {e}"
+            logger.error(f"An error occurred in TTSService: {error_log_msg}")
+            match = re.search(r"status code (\d+)", str(e))
+            if match:
+                status_code = int(match.group(1))
+            else:
+                status_code = 500
+            raise
+        finally:
+            end_time = time.perf_counter()
+            latency_ms = int((end_time - start_time) * 1000)
+            if not is_success:
+                await add_error_log(
+                    gemini_key=api_key,
+                    model_name=settings.TTS_MODEL,
+                    error_type="google-tts",
+                    error_log=error_log_msg,
+                    error_code=status_code,
+                    request_msg=request.input
+                 )
+            await add_request_log(
+                model_name=settings.TTS_MODEL,
+                api_key=api_key,
+                is_success=is_success,
+                status_code=status_code,
+                latency_ms=latency_ms,
+                request_time=request_datetime
+            )

app/service/update/update_service.py

@@ -7,11 +7,7 @@ from app.log.logger import get_update_logger
 
 logger = get_update_logger()
 
-# GitHub repository details are read from settings (defined in app/config/config.py or environment variables)
-
-# GITHUB_API_URL will be constructed inside the function to ensure settings are loaded
-
-VERSION_FILE_PATH = "VERSION" # Path relative to project root
+VERSION_FILE_PATH = "VERSION"
 
 async def check_for_updates() -> Tuple[bool, Optional[str], Optional[str]]:
     """
@@ -24,9 +20,6 @@ async def check_for_updates() -> Tuple[bool, Optional[str], Optional[str]]:
             - Optional[str]: 如果检查失败，则为错误消息，否则为 None。
     """
     try:
-        # Read current version from VERSION file
-        # Ensure the path is correct relative to the execution context or use absolute path if needed
-        # Assuming execution from project root d:/develop/pythonProjects/gemini-balance
         with open(VERSION_FILE_PATH, 'r', encoding='utf-8') as f:
             current_v = f.read().strip()
         if not current_v:
@@ -41,25 +34,22 @@ async def check_for_updates() -> Tuple[bool, Optional[str], Optional[str]]:
 
     logger.info(f"当前应用程序版本 (from {VERSION_FILE_PATH}): {current_v}")
 
-    # Check if repository details are configured in settings
     if not settings.GITHUB_REPO_OWNER or not settings.GITHUB_REPO_NAME or \
        settings.GITHUB_REPO_OWNER == "your_owner" or settings.GITHUB_REPO_NAME == "your_repo":
         logger.warning("GitHub repository owner/name not configured in settings. Skipping update check.")
         return False, None, "Update check skipped: Repository not configured in settings."
 
-    # Construct the API URL inside the function to ensure settings are loaded
     github_api_url = f"https://api.github.com/repos/{settings.GITHUB_REPO_OWNER}/{settings.GITHUB_REPO_NAME}/releases/latest"
-    logger.debug(f"Checking for updates at URL: {github_api_url}") # Log the URL for debugging
+    logger.debug(f"Checking for updates at URL: {github_api_url}")
 
     try:
         async with httpx.AsyncClient(timeout=10.0) as client:
-            # 添加 User-Agent 头，GitHub API 可能需要
             headers = {
                 "Accept": "application/vnd.github.v3+json",
-                "User-Agent": f"{settings.GITHUB_REPO_NAME}-UpdateChecker/1.0" # Use repo name from settings for User-Agent
+                "User-Agent": f"{settings.GITHUB_REPO_NAME}-UpdateChecker/1.0"
             }
-            response = await client.get(github_api_url, headers=headers) # Use the locally constructed URL
-            response.raise_for_status() # 对错误的 HTTP 状态码（4xx 或 5xx）抛出异常
+            response = await client.get(github_api_url, headers=headers)
+            response.raise_for_status()
 
             latest_release = response.json()
             latest_v_str = latest_release.get("tag_name")
@@ -68,7 +58,6 @@ async def check_for_updates() -> Tuple[bool, Optional[str], Optional[str]]:
                 logger.warning("在最新的 GitHub release 响应中找不到 'tag_name'。")
                 return False, None, "无法从 GitHub 解析最新版本。"
 
-            # 移除 tag 名称中可能存在的 'v' 前缀
             if latest_v_str.startswith('v'):
                 latest_v_str = latest_v_str[1:]
 
@@ -98,8 +87,6 @@ async def check_for_updates() -> Tuple[bool, Optional[str], Optional[str]]:
         logger.error(f"检查更新时发生网络错误: {e}")
         return False, None, "更新检查期间发生网络错误。"
     except version.InvalidVersion:
-        # Note: latest_v_str might not be defined if the error occurs before fetching it.
-        # Consider adding a check or default value for logging.
         latest_v_str_for_log = latest_v_str if 'latest_v_str' in locals() else 'N/A'
         logger.error(f"发现无效的版本格式。当前 (from {VERSION_FILE_PATH}): '{current_v}', 最新: '{latest_v_str_for_log}'")
         return False, None, "遇到无效的版本格式。"

app/templates/auth.html

@@ -6,13 +6,14 @@
 <style>
     /* auth.html specific styles */
     .auth-glass-card { /* Renamed to avoid conflict if base.html has .glass-card */
-        background: rgba(255, 255, 255, 0.85); /* Increased opacity */
+        background: rgba(255, 255, 255, 0.95); /* High opacity white for light theme */
         backdrop-filter: blur(20px);
         -webkit-backdrop-filter: blur(20px);
-        border: 1px solid rgba(255, 255, 255, 0.2);
+        border: 1px solid rgba(0, 0, 0, 0.08);
+        box-shadow: 0 10px 25px -5px rgba(0, 0, 0, 0.1), 0 10px 10px -5px rgba(0, 0, 0, 0.04);
     }
     .auth-bg-gradient { /* Renamed to avoid conflict if base.html has .bg-gradient */
-         background: linear-gradient(135deg, #4F46E5 0%, #7C3AED 50%, #EC4899 100%);
+         background: #f8fafc; /* Light gray background for auth page */
     }
     /* .input-icon class removed, using direct Tailwind classes now */
     /* Keep button ripple effect if needed, or remove if base provides similar */
@@ -49,7 +50,7 @@
             </div>
         </div>
         
-        <h2 class="text-3xl font-extrabold text-center text-transparent bg-clip-text bg-gradient-to-r from-primary-600 to-primary-700 mb-8 animate-slide-down">
+        <h2 class="text-3xl font-extrabold text-center text-gray-800 mb-8 animate-slide-down">
              <img src="/static/icons/logo.png" alt="Gemini Balance Logo" class="h-9 inline-block align-middle mr-2">
              Gemini Balance
         </h2>
@@ -67,9 +68,9 @@
                 >
             </div>
             
-            <button 
-                type="submit" 
-                class="w-full py-4 rounded-xl bg-gradient-to-r from-primary-600 to-primary-700 text-white font-semibold transition duration-300 transform hover:-translate-y-1 hover:shadow-lg"
+            <button
+                type="submit"
+                class="w-full py-4 rounded-xl bg-blue-600 hover:bg-blue-700 text-white font-semibold transition duration-300 transform hover:-translate-y-1 hover:shadow-lg"
             >
                 登录
             </button>

app/templates/base.html

@@ -1,316 +1,642 @@
 <!DOCTYPE html>
 <html lang="zh-CN">
-<head>
-    <meta charset="UTF-8">
-    <meta name="viewport" content="width=device-width, initial-scale=1.0">
+  <head>
+    <meta charset="UTF-8" />
+    <meta name="viewport" content="width=device-width, initial-scale=1.0" />
     <title>{% block title %}Gemini Balance{% endblock %}</title>
-    <link rel="manifest" href="/static/manifest.json">
-    <meta name="theme-color" content="#4F46E5">
-    <meta name="apple-mobile-web-app-capable" content="yes">
-    <meta name="apple-mobile-web-app-status-bar-style" content="black">
-    <meta name="apple-mobile-web-app-title" content="GBalance">
-    <link rel="icon" href="/static/icons/icon-192x192.png">
-    <link href="https://fonts.googleapis.com/css2?family=Inter:wght@300;400;500;600;700&display=swap" rel="stylesheet">
-    <link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/font-awesome/6.4.0/css/all.min.css">
+    <link rel="manifest" href="/static/manifest.json" />
+    <meta name="theme-color" content="#4F46E5" />
+    <meta name="apple-mobile-web-app-capable" content="yes" />
+    <meta name="apple-mobile-web-app-status-bar-style" content="black" />
+    <meta name="apple-mobile-web-app-title" content="GBalance" />
+    <link rel="icon" href="/static/icons/icon-192x192.png" />
+    <link
+      href="https://fonts.googleapis.com/css2?family=Inter:wght@300;400;500;600;700&display=swap"
+      rel="stylesheet"
+    />
+    <link
+      rel="stylesheet"
+      href="https://cdnjs.cloudflare.com/ajax/libs/font-awesome/6.4.0/css/all.min.css"
+    />
     <script src="https://cdn.tailwindcss.com"></script>
     <script>
-        tailwind.config = {
-            theme: {
-                extend: {
-                    colors: {
-                        primary: {
-                            50: '#eef2ff',
-                            100: '#e0e7ff',
-                            200: '#c7d2fe',
-                            300: '#a5b4fc',
-                            400: '#818cf8',
-                            500: '#6366f1',
-                            600: '#4f46e5',
-                            700: '#4338ca',
-                            800: '#3730a3',
-                            900: '#312e81',
-                        },
-                        success: {
-                            50: '#ecfdf5',
-                            500: '#10b981',
-                            600: '#059669'
-                        },
-                        danger: {
-                            50: '#fef2f2',
-                            500: '#ef4444',
-                            600: '#dc2626'
-                        }
-                    },
-                    fontFamily: {
-                        sans: ['Inter', 'sans-serif'],
-                        mono: ['JetBrains Mono', 'SFMono-Regular', 'Menlo', 'Monaco', 'Consolas', 'monospace'],
-                    },
-                    animation: {
-                        'fade-in': 'fadeIn 0.5s ease-out',
-                        'slide-up': 'slideUp 0.5s ease-out',
-                        'slide-down': 'slideDown 0.5s ease-out',
-                        'shake': 'shake 0.5s ease-in-out',
-                        'spin': 'spin 1s linear infinite',
-                    },
-                    keyframes: {
-                        fadeIn: {
-                            '0%': { opacity: '0' },
-                            '100%': { opacity: '1' },
-                        },
-                        slideUp: {
-                            '0%': { transform: 'translateY(20px)', opacity: '0' },
-                            '100%': { transform: 'translateY(0)', opacity: '1' },
-                        },
-                        slideDown: {
-                            '0%': { transform: 'translateY(-20px)', opacity: '0' },
-                            '100%': { transform: 'translateY(0)', opacity: '1' },
-                        },
-                        shake: {
-                            '0%, 100%': { transform: 'translateX(0)' },
-                            '25%': { transform: 'translateX(-5px)' },
-                            '75%': { transform: 'translateX(5px)' },
-                        },
-                         spin: {
-                            '0%': { transform: 'rotate(0deg)' },
-                            '100%': { transform: 'rotate(360deg)' },
-                        },
-                    },
-                }
-            }
-        }
+      tailwind.config = {
+        theme: {
+          extend: {
+            colors: {
+              primary: {
+                50: "#eef2ff",
+                100: "#e0e7ff",
+                200: "#c7d2fe",
+                300: "#a5b4fc",
+                400: "#818cf8",
+                500: "#6366f1",
+                600: "#4f46e5",
+                700: "#4338ca",
+                800: "#3730a3",
+                900: "#312e81",
+              },
+              success: {
+                50: "#ecfdf5",
+                500: "#10b981",
+                600: "#059669",
+              },
+              danger: {
+                50: "#fef2f2",
+                500: "#ef4444",
+                600: "#dc2626",
+              },
+            },
+            fontFamily: {
+              sans: ["Inter", "sans-serif"],
+              mono: [
+                "JetBrains Mono",
+                "SFMono-Regular",
+                "Menlo",
+                "Monaco",
+                "Consolas",
+                "monospace",
+              ],
+            },
+            animation: {
+              "fade-in": "fadeIn 0.5s ease-out",
+              "slide-up": "slideUp 0.5s ease-out",
+              "slide-down": "slideDown 0.5s ease-out",
+              shake: "shake 0.5s ease-in-out",
+              spin: "spin 1s linear infinite",
+            },
+            keyframes: {
+              fadeIn: {
+                "0%": { opacity: "0" },
+                "100%": { opacity: "1" },
+              },
+              slideUp: {
+                "0%": { transform: "translateY(20px)", opacity: "0" },
+                "100%": { transform: "translateY(0)", opacity: "1" },
+              },
+              slideDown: {
+                "0%": { transform: "translateY(-20px)", opacity: "0" },
+                "100%": { transform: "translateY(0)", opacity: "1" },
+              },
+              shake: {
+                "0%, 100%": { transform: "translateX(0)" },
+                "25%": { transform: "translateX(-5px)" },
+                "75%": { transform: "translateX(5px)" },
+              },
+              spin: {
+                "0%": { transform: "rotate(0deg)" },
+                "100%": { transform: "rotate(360deg)" },
+              },
+            },
+          },
+        },
+      };
     </script>
     <style>
-        .glass-card {
-            background: rgba(255, 255, 255, 0.85); /* Slightly increased opacity for better readability */
-            backdrop-filter: blur(16px);
-            -webkit-backdrop-filter: blur(16px);
-            border: 1px solid rgba(255, 255, 255, 0.18); /* Subtle border */
-        }
-        .bg-gradient {
-            background: linear-gradient(135deg, #4F46E5 0%, #7C3AED 50%, #EC4899 100%);
-        }
-        /* Scrollbar styling */
-        ::-webkit-scrollbar {
-            width: 8px;
-            height: 8px;
-        }
-        ::-webkit-scrollbar-track {
-            background: rgba(243, 244, 246, 0.8); /* bg-gray-100 with opacity */
-            border-radius: 10px;
-        }
-        ::-webkit-scrollbar-thumb {
-            background: rgba(79, 70, 229, 0.4); /* primary-600 with opacity */
-            border-radius: 10px;
-        }
-        ::-webkit-scrollbar-thumb:hover {
-            background: rgba(79, 70, 229, 0.6); /* primary-600 with more opacity */
-        }
-        /* Basic modal styles */
-        .modal {
-            display: none;
-            position: fixed;
-            z-index: 50;
-            left: 0;
-            top: 0;
-            width: 100%;
-            height: 100%;
-            background-color: rgba(0,0,0,0.5);
-            backdrop-filter: blur(4px);
-        }
-        .modal.show {
-            display: flex;
-            align-items: center;
-            justify-content: center;
-        }
-        /* Loading spinner */
-        .loading-spin {
-            animation: spin 1s linear infinite;
-        }
-        @keyframes spin {
-            from { transform: rotate(0deg); }
-            to { transform: rotate(360deg); }
-        }
-        /* Notification */
-        .notification {
-            position: fixed;
-            bottom: 5rem; /* Adjusted from bottom-20 */
-            left: 50%;
-            transform: translateX(-50%);
-            padding: 0.75rem 1.25rem; /* px-5 py-3 */
-            border-radius: 0.5rem; /* rounded-lg */
-            background-color: rgba(0, 0, 0, 0.8);
-            color: white;
-            font-weight: 500; /* font-medium */
-            z-index: 1000; /* Increased z-index */
-            opacity: 0;
-            transition: opacity 0.3s ease-in-out, transform 0.3s ease-in-out;
-        }
-        .notification.show {
-            opacity: 1;
-            transform: translate(-50%, 0);
-        }
-        .notification.error {
-             background-color: rgba(220, 38, 38, 0.8); /* danger-600 with opacity */
-        }
-        /* Scroll buttons */
-        .scroll-buttons {
-            position: fixed;
-            right: 1.25rem; /* right-5 */
-            bottom: 5rem; /* bottom-20 */
-            display: flex;
-            flex-direction: column;
-            gap: 0.5rem; /* gap-2 */
-            z-index: 10;
-        }
-        .scroll-button {
-            width: 2.5rem; /* w-10 */
-            height: 2.5rem; /* h-10 */
-            background-color: #4f46e5; /* bg-primary-600 */
-            color: white;
-            border-radius: 9999px; /* rounded-full */
-            box-shadow: 0 4px 6px -1px rgba(0, 0, 0, 0.1), 0 2px 4px -1px rgba(0, 0, 0, 0.06); /* shadow-md */
-            display: flex;
-            align-items: center;
-            justify-content: center;
-            transition: all 0.3s ease-in-out;
-        }
-        .scroll-button:hover {
-            background-color: #4338ca; /* hover:bg-primary-700 */
-            box-shadow: 0 10px 15px -3px rgba(0, 0, 0, 0.1), 0 4px 6px -2px rgba(0, 0, 0, 0.05); /* hover:shadow-lg */
-        }
-        {% block head_extra_styles %}
-        {% endblock %}
+      .glass-card {
+          background: rgba(255, 255, 255, 0.95); /* High opacity white for light theme */
+          backdrop-filter: blur(16px);
+          -webkit-backdrop-filter: blur(16px);
+          border: 1px solid rgba(0, 0, 0, 0.08); /* Light gray border */
+          box-shadow: 0 4px 6px -1px rgba(0, 0, 0, 0.1), 0 2px 4px -1px rgba(0, 0, 0, 0.06);
+      }
+      .bg-gradient {
+          background: #ffffff; /* Clean white background */
+      }
+      /* Scrollbar styling */
+      ::-webkit-scrollbar {
+          width: 8px;
+          height: 8px;
+      }
+      ::-webkit-scrollbar-track {
+          background: rgba(243, 244, 246, 0.8); /* bg-gray-100 with opacity */
+          border-radius: 10px;
+      }
+      ::-webkit-scrollbar-thumb {
+          background: rgba(107, 114, 128, 0.6); /* gray-500 for light theme */
+          border-radius: 10px;
+      }
+      ::-webkit-scrollbar-thumb:hover {
+          background: rgba(75, 85, 99, 0.8); /* gray-600 for light theme */
+      }
+      /* Basic modal styles */
+      .modal {
+          display: none;
+          position: fixed;
+          z-index: 50;
+          left: 0;
+          top: 0;
+          width: 100%;
+          height: 100%;
+          background-color: rgba(0,0,0,0.5);
+          backdrop-filter: blur(4px);
+      }
+      .modal.show {
+          display: flex;
+          align-items: center;
+          justify-content: center;
+      }
+
+      /* Global modal content styling for light theme consistency */
+      .modal .w-full[style*="background-color: rgba(70, 50, 150"],
+      .modal .w-full[style*="background-color: rgba(80, 60, 160"] {
+          background-color: rgba(255, 255, 255, 0.98) !important;
+          color: #374151 !important; /* gray-700 */
+          border: 1px solid rgba(0, 0, 0, 0.08) !important;
+          box-shadow: 0 20px 25px -5px rgba(0, 0, 0, 0.1), 0 10px 10px -5px rgba(0, 0, 0, 0.04) !important;
+      }
+
+      /* Global modal text color fixes */
+      .modal .text-gray-100, .modal h2.text-gray-100, .modal h3.text-gray-100 {
+          color: #1f2937 !important; /* gray-800 */
+          font-weight: 600 !important;
+      }
+
+      .modal .text-gray-200, .modal .text-gray-300 {
+          color: #6b7280 !important; /* gray-500 */
+      }
+
+      .modal .text-gray-300:hover {
+          color: #374151 !important; /* gray-700 */
+      }
+
+      /* Global modal button styling */
+      .modal .bg-violet-600, .modal button.bg-violet-600 {
+          background-color: #3b82f6 !important; /* blue-500 - light blue */
+          color: #ffffff !important;
+      }
+
+      .modal .bg-violet-600:hover, .modal button.bg-violet-600:hover {
+          background-color: #2563eb !important; /* blue-600 - darker light blue */
+      }
+
+      /* Global modal blue button styling */
+      .modal .bg-blue-500, .modal button.bg-blue-500,
+      .modal .bg-blue-600, .modal button.bg-blue-600,
+      .modal .bg-blue-700, .modal button.bg-blue-700 {
+          background-color: #3b82f6 !important; /* blue-500 - light blue */
+          color: #ffffff !important;
+      }
+
+      .modal .bg-blue-500:hover, .modal button.bg-blue-500:hover,
+      .modal .bg-blue-600:hover, .modal button.bg-blue-600:hover,
+      .modal .bg-blue-700:hover, .modal button.bg-blue-700:hover {
+          background-color: #2563eb !important; /* blue-600 - darker light blue */
+      }
+
+      /* Global modal red button styling */
+      .modal .bg-red-500, .modal button.bg-red-500,
+      .modal .bg-red-600, .modal button.bg-red-600,
+      .modal .bg-red-700, .modal button.bg-red-700 {
+          background-color: #f87171 !important; /* red-400 - bright light red */
+          color: #ffffff !important;
+      }
+
+      .modal .bg-red-500:hover, .modal button.bg-red-500:hover,
+      .modal .bg-red-600:hover, .modal button.bg-red-600:hover,
+      .modal .bg-red-700:hover, .modal button.bg-red-700:hover {
+          background-color: #ef4444 !important; /* red-500 - darker bright light red */
+      }
+
+      /* Global modal gray button styling */
+      .modal .bg-gray-500, .modal button.bg-gray-500,
+      .modal .bg-gray-600, .modal button.bg-gray-600,
+      .modal .bg-gray-700, .modal button.bg-gray-700 {
+          background-color: #e5e7eb !important; /* gray-200 - light gray */
+          color: #374151 !important; /* gray-700 - dark text for contrast */
+      }
+
+      .modal .bg-gray-500:hover, .modal button.bg-gray-500:hover,
+      .modal .bg-gray-600:hover, .modal button.bg-gray-600:hover,
+      .modal .bg-gray-700:hover, .modal button.bg-gray-700:hover {
+          background-color: #d1d5db !important; /* gray-300 - darker light gray */
+          color: #374151 !important; /* gray-700 - dark text for contrast */
+      }
+
+      /* Comprehensive button contrast fixes */
+      /* Ensure all dark background buttons have white text */
+      .bg-blue-500, .bg-blue-600, .bg-blue-700, .bg-blue-800, .bg-blue-900,
+      .bg-red-500, .bg-red-600, .bg-red-700, .bg-red-800, .bg-red-900,
+      .bg-green-500, .bg-green-600, .bg-green-700, .bg-green-800, .bg-green-900,
+      .bg-purple-500, .bg-purple-600, .bg-purple-700, .bg-purple-800, .bg-purple-900,
+      .bg-indigo-500, .bg-indigo-600, .bg-indigo-700, .bg-indigo-800, .bg-indigo-900,
+      .bg-violet-500, .bg-violet-600, .bg-violet-700, .bg-violet-800, .bg-violet-900,
+      .bg-sky-500, .bg-sky-600, .bg-sky-700, .bg-sky-800, .bg-sky-900,
+      .bg-teal-500, .bg-teal-600, .bg-teal-700, .bg-teal-800, .bg-teal-900,
+      .bg-gray-700, .bg-gray-800, .bg-gray-900,
+      .bg-slate-500, .bg-slate-600, .bg-slate-700, .bg-slate-800, .bg-slate-900 {
+          color: #ffffff !important;
+      }
+
+      /* Ensure all light background buttons have dark text */
+      .bg-gray-50, .bg-gray-100, .bg-gray-200, .bg-gray-300,
+      .bg-white, .bg-transparent {
+          color: #374151 !important; /* gray-700 */
+      }
+
+      /* Fix button children text inheritance */
+      .bg-blue-500 *, .bg-blue-600 *, .bg-blue-700 *, .bg-blue-800 *, .bg-blue-900 *,
+      .bg-red-500 *, .bg-red-600 *, .bg-red-700 *, .bg-red-800 *, .bg-red-900 *,
+      .bg-green-500 *, .bg-green-600 *, .bg-green-700 *, .bg-green-800 *, .bg-green-900 *,
+      .bg-purple-500 *, .bg-purple-600 *, .bg-purple-700 *, .bg-purple-800 *, .bg-purple-900 *,
+      .bg-violet-500 *, .bg-violet-600 *, .bg-violet-700 *, .bg-violet-800 *, .bg-violet-900 *,
+      .bg-sky-500 *, .bg-sky-600 *, .bg-sky-700 *, .bg-sky-800 *, .bg-sky-900 *,
+      .bg-teal-500 *, .bg-teal-600 *, .bg-teal-700 *, .bg-teal-800 *, .bg-teal-900 *,
+      .bg-gray-700 *, .bg-gray-800 *, .bg-gray-900 *,
+      .bg-slate-500 *, .bg-slate-600 *, .bg-slate-700 *, .bg-slate-800 *, .bg-slate-900 * {
+          color: inherit !important;
+      }
+
+      /* Global form element styling for consistency */
+      select, input[type="text"], input[type="number"], input[type="search"],
+      input[type="email"], input[type="password"], input[type="datetime-local"],
+      textarea, .form-input, .form-select {
+          background-color: rgba(255, 255, 255, 0.95) !important;
+          color: #374151 !important; /* gray-700 */
+          border: 1px solid rgba(0, 0, 0, 0.12) !important;
+          border-radius: 0.375rem !important; /* rounded-md */
+      }
+
+      select:focus, input:focus, textarea:focus,
+      .form-input:focus, .form-select:focus {
+          border-color: #3b82f6 !important; /* blue-500 */
+          box-shadow: 0 0 0 3px rgba(59, 130, 246, 0.1) !important;
+          outline: none !important;
+      }
+
+      /* Fix dropdown option styling */
+      select option {
+          background-color: rgba(255, 255, 255, 0.98) !important;
+          color: #374151 !important; /* gray-700 */
+          padding: 8px !important;
+      }
+
+      /* Fix pagination controls globally */
+      .pagination-button, .pagination a, .pagination button {
+          background-color: rgba(255, 255, 255, 0.9) !important;
+          color: #374151 !important; /* gray-700 */
+          border: 1px solid rgba(0, 0, 0, 0.08) !important;
+          transition: all 0.15s ease-in-out !important;
+      }
+
+      .pagination-button:hover, .pagination a:hover, .pagination button:hover {
+          background-color: rgba(229, 231, 235, 1) !important; /* gray-200 */
+          border-color: rgba(0, 0, 0, 0.12) !important;
+          transform: translateY(-1px) !important;
+      }
+
+      .pagination-button.active, .pagination a.active, .pagination button.active {
+          background-color: #3b82f6 !important; /* blue-500 - light blue */
+          color: #ffffff !important;
+          border-color: #2563eb !important; /* blue-600 - darker light blue */
+          font-weight: 600 !important;
+      }
+      /* Loading spinner */
+      .loading-spin {
+          animation: spin 1s linear infinite;
+      }
+      @keyframes spin {
+          from { transform: rotate(0deg); }
+          to { transform: rotate(360deg); }
+      }
+      /* Notification */
+      .notification {
+          position: fixed;
+          bottom: 5rem; /* Adjusted from bottom-20 */
+          left: 50%;
+          transform: translateX(-50%);
+          padding: 0.75rem 1.25rem; /* px-5 py-3 */
+          border-radius: 0.5rem; /* rounded-lg */
+          background-color: rgba(34, 197, 94, 0.95); /* green-500 for success */
+          color: white;
+          font-weight: 500; /* font-medium */
+          z-index: 1000; /* Increased z-index */
+          opacity: 0;
+          transition: opacity 0.3s ease-in-out, transform 0.3s ease-in-out;
+          box-shadow: 0 10px 15px -3px rgba(0, 0, 0, 0.1), 0 4px 6px -2px rgba(0, 0, 0, 0.05);
+      }
+      .notification.show {
+          opacity: 1;
+          transform: translate(-50%, 0);
+      }
+      .notification.error {
+           background-color: rgba(239, 68, 68, 0.95); /* red-500 for error */
+      }
+      /* Scroll buttons */
+      .scroll-buttons {
+          position: fixed;
+          right: 1.25rem; /* right-5 */
+          bottom: 5rem; /* bottom-20 */
+          display: flex;
+          flex-direction: column;
+          gap: 0.5rem; /* gap-2 */
+          z-index: 10;
+      }
+      .scroll-button {
+          width: 2.5rem; /* w-10 */
+          height: 2.5rem; /* h-10 */
+          background-color: #3b82f6; /* blue-500 - light blue */
+          color: white;
+          border-radius: 9999px; /* rounded-full */
+          box-shadow: 0 4px 6px -1px rgba(0, 0, 0, 0.1), 0 2px 4px -1px rgba(0, 0, 0, 0.06); /* shadow-md */
+          display: flex;
+          align-items: center;
+          justify-content: center;
+          transition: all 0.3s ease-in-out;
+      }
+      .scroll-button:hover {
+          background-color: #2563eb; /* blue-600 - darker light blue */
+          box-shadow: 0 10px 15px -3px rgba(0, 0, 0, 0.1), 0 4px 6px -2px rgba(0, 0, 0, 0.05); /* hover:shadow-lg */
+      }
+
+      /* Global overrides for light theme consistency */
+      .text-gray-200, .text-gray-300, .text-gray-400 {
+          color: #6b7280 !important; /* gray-500 for better contrast */
+      }
+
+      /* Navigation and header improvements */
+      .bg-primary-600, .bg-primary-700 {
+          background-color: #3b82f6 !important; /* blue-500 - light blue */
+      }
+
+      .text-primary-600, .text-primary-700 {
+          color: #3b82f6 !important; /* blue-500 - light blue */
+      }
+
+      .border-primary-500, .focus\\:border-primary-500 {
+          border-color: #3b82f6 !important; /* blue-500 */
+      }
+
+      .ring-primary-200, .focus\\:ring-primary-200 {
+          --tw-ring-color: rgba(59, 130, 246, 0.2) !important; /* blue-500 with opacity */
+      }
+
+      /* Global purple to blue conversion */
+      .bg-violet-50, .bg-violet-100, .bg-violet-200, .bg-violet-300, .bg-violet-400, .bg-violet-500, .bg-violet-600, .bg-violet-700, .bg-violet-800, .bg-violet-900 {
+          background-color: #3b82f6 !important; /* blue-500 - light blue */
+      }
+
+      .text-violet-50, .text-violet-100, .text-violet-200, .text-violet-300, .text-violet-400, .text-violet-500, .text-violet-600, .text-violet-700, .text-violet-800, .text-violet-900 {
+          color: #3b82f6 !important; /* blue-500 - light blue */
+      }
+
+      .border-violet-50, .border-violet-100, .border-violet-200, .border-violet-300, .border-violet-400, .border-violet-500, .border-violet-600, .border-violet-700, .border-violet-800, .border-violet-900 {
+          border-color: #3b82f6 !important; /* blue-500 - light blue */
+      }
+
+      /* Global button color overrides */
+      /* Blue buttons to light blue */
+      .bg-blue-500, .bg-blue-600, .bg-blue-700, .bg-blue-800, .bg-blue-900,
+      button.bg-blue-500, button.bg-blue-600, button.bg-blue-700, button.bg-blue-800, button.bg-blue-900 {
+          background-color: #3b82f6 !important; /* blue-500 - light blue */
+      }
+
+      .bg-blue-500:hover, .bg-blue-600:hover, .bg-blue-700:hover, .bg-blue-800:hover, .bg-blue-900:hover,
+      button.bg-blue-500:hover, button.bg-blue-600:hover, button.bg-blue-700:hover, button.bg-blue-800:hover, button.bg-blue-900:hover,
+      .hover\\:bg-blue-600:hover, .hover\\:bg-blue-700:hover, .hover\\:bg-blue-800:hover {
+          background-color: #2563eb !important; /* blue-600 - darker light blue */
+      }
+
+      /* Red buttons to bright light red */
+      .bg-red-500, .bg-red-600, .bg-red-700, .bg-red-800, .bg-red-900,
+      button.bg-red-500, button.bg-red-600, button.bg-red-700, button.bg-red-800, button.bg-red-900 {
+          background-color: #f87171 !important; /* red-400 - bright light red */
+      }
+
+      .bg-red-500:hover, .bg-red-600:hover, .bg-red-700:hover, .bg-red-800:hover, .bg-red-900:hover,
+      button.bg-red-500:hover, button.bg-red-600:hover, button.bg-red-700:hover, button.bg-red-800:hover, button.bg-red-900:hover,
+      .hover\\:bg-red-600:hover, .hover\\:bg-red-700:hover, .hover\\:bg-red-800:hover {
+          background-color: #ef4444 !important; /* red-500 - darker bright light red */
+      }
+
+      /* Gray buttons to light gray */
+      .bg-gray-500, .bg-gray-600, .bg-gray-700, .bg-gray-800, .bg-gray-900,
+      button.bg-gray-500, button.bg-gray-600, button.bg-gray-700, button.bg-gray-800, button.bg-gray-900 {
+          background-color: #e5e7eb !important; /* gray-200 - light gray */
+          color: #374151 !important; /* gray-700 - dark text for contrast */
+      }
+
+      .bg-gray-500:hover, .bg-gray-600:hover, .bg-gray-700:hover, .bg-gray-800:hover, .bg-gray-900:hover,
+      button.bg-gray-500:hover, button.bg-gray-600:hover, button.bg-gray-700:hover, button.bg-gray-800:hover, button.bg-gray-900:hover,
+      .hover\\:bg-gray-600:hover, .hover\\:bg-gray-700:hover, .hover\\:bg-gray-800:hover {
+          background-color: #d1d5db !important; /* gray-300 - darker light gray */
+          color: #374151 !important; /* gray-700 - dark text for contrast */
+      }
+
+      /* Ensure all text has proper contrast in light theme */
+      .text-white {
+          color: #374151 !important; /* gray-700 for better contrast on light backgrounds */
+      }
+
+      /* Fix dark button text - ensure white text on dark backgrounds */
+      .bg-blue-500, .bg-blue-600, .bg-blue-700, .bg-blue-800, .bg-blue-900,
+      .bg-red-500, .bg-red-600, .bg-red-700, .bg-red-800, .bg-red-900,
+      .bg-green-500, .bg-green-600, .bg-green-700, .bg-green-800, .bg-green-900,
+      .bg-purple-500, .bg-purple-600, .bg-purple-700, .bg-purple-800, .bg-purple-900,
+      .bg-indigo-500, .bg-indigo-600, .bg-indigo-700, .bg-indigo-800, .bg-indigo-900,
+      .bg-gray-700, .bg-gray-800, .bg-gray-900,
+      .bg-sky-500, .bg-sky-600, .bg-sky-700, .bg-sky-800, .bg-sky-900 {
+          color: #ffffff !important;
+      }
+
+      /* Ensure buttons with dark backgrounds have white text */
+      button.bg-blue-500, button.bg-blue-600, button.bg-blue-700,
+      button.bg-red-500, button.bg-red-600, button.bg-red-700,
+      button.bg-green-500, button.bg-green-600, button.bg-green-700,
+      button.bg-sky-500, button.bg-sky-600, button.bg-sky-700,
+      .btn-primary, .btn-danger, .btn-success, .btn-info {
+          color: #ffffff !important;
+      }
+
+      /* Override any nested text color rules for dark buttons */
+      .bg-blue-500 *, .bg-blue-600 *, .bg-blue-700 *,
+      .bg-red-500 *, .bg-red-600 *, .bg-red-700 *,
+      .bg-green-500 *, .bg-green-600 *, .bg-green-700 *,
+      .bg-sky-500 *, .bg-sky-600 *, .bg-sky-700 * {
+          color: inherit !important;
+      }
+
+      {% block head_extra_styles %}
+      {% endblock %}
     </style>
     {% block head_extra_scripts %}{% endblock %}
-</head>
-<body class="bg-gradient min-h-screen text-gray-800 pt-6 pb-16">
+  </head>
+  <body class="bg-white min-h-screen text-gray-900 pt-6 pb-16">
     {% block content %}{% endblock %}
 
     <!-- 底部版权 -->
-    <div class="fixed bottom-0 left-0 w-full py-3 bg-white bg-opacity-80 backdrop-blur-md text-center text-sm text-gray-600 border-t border-gray-200">
-        © <span id="copyright-year"></span> by
-        <a href="https://linux.do/u/snaily" target="_blank" class="text-primary-600 hover:text-primary-800 transition duration-300">
-            <img src="https://linux.do/user_avatar/linux.do/snaily/288/306510_2.gif" alt="snaily" class="inline-block w-5 h-5 rounded-full align-middle mr-1">snaily
-        </a> |
-        <a href="https://github.com/snailyp/gemini-balance" target="_blank" class="text-primary-600 hover:text-primary-800 transition duration-300">
-            <i class="fab fa-github"></i> GitHub
-        </a> |
-        <a href="https://afdian.com/a/snaily" target="_blank" class="text-primary-600 hover:text-primary-800 transition duration-300">
-            <i class="fas fa-drumstick-bite text-yellow-600"></i> 给作者加鸡腿
+    <div
+      class="fixed bottom-0 left-0 w-full py-3 bg-white bg-opacity-95 backdrop-blur-md text-sm text-gray-800 border-t border-gray-200 flex flex-col items-center space-y-1"
+    >
+      <!-- 第一行 -->
+      <div class="flex items-center justify-center space-x-2">
+        <span>© <span id="copyright-year"></span> by</span>
+        <a
+          href="https://linux.do/u/snaily"
+          target="_blank"
+          class="text-primary-600 hover:text-primary-800 transition duration-300 flex items-center"
+        >
+          <img
+            src="https://linux.do/user_avatar/linux.do/snaily/288/306510_2.gif"
+            alt="snaily"
+            class="inline-block w-5 h-5 rounded-full align-middle mr-1"
+          />snaily
         </a>
-        <span class="mx-1">|</span>
-        <span class="text-xs text-yellow-600 font-semibold">
-            <i class="fas fa-exclamation-triangle mr-1"></i>免费项目，谨防诈骗
+        <span class="text-gray-400">|</span>
+        <a
+          href="https://github.com/snailyp/gemini-balance"
+          target="_blank"
+          class="text-primary-600 hover:text-primary-800 transition duration-300 flex items-center"
+        >
+          <i class="fab fa-github mr-1"></i> GitHub
+        </a>
+      </div>
+      <!-- 第二行 -->
+      <div class="flex items-center justify-center space-x-2 text-xs">
+        <a
+          href="https://gb-docs.snaily.top/guide/supportme.html"
+          target="_blank"
+          class="text-primary-600 hover:text-primary-800 transition duration-300 flex items-center"
+        >
+          <i class="fas fa-drumstick-bite text-yellow-600 mr-1"></i> 给作者加鸡腿
+        </a>
+        <span class="text-gray-400">|</span>
+        <a
+          href="https://gb-docs.snaily.top"
+          target="_blank"
+          class="text-primary-600 hover:text-primary-800 transition duration-300 flex items-center"
+        >
+          <i class="fas fa-book mr-1"></i> 在线文档
+        </a>
+<span class="text-gray-400">|</span>
+        <a
+          href="https://t.me/+soaHax5lyI0wZDVl"
+          target="_blank"
+          class="text-primary-600 hover:text-primary-800 transition duration-300 flex items-center"
+        >
+          <i class="fab fa-telegram-plane mr-1"></i> 交流群
+        </a>
+        <span class="text-gray-400">|</span>
+        <span class="text-yellow-600 font-semibold flex items-center">
+          <i class="fas fa-exclamation-triangle mr-1"></i>免费项目，谨防诈骗
         </span>
-        <span id="version-info-container" class="inline-block">
-            <!-- Version info will be loaded here by JavaScript -->
+        <span id="version-info-container" class="inline-flex items-center">
+          <!-- Version info will be loaded here by JavaScript -->
         </span>
+      </div>
     </div>
 
     <!-- 通用JS -->
     <script>
-        // 设置版权年份
-        document.getElementById('copyright-year').textContent = new Date().getFullYear();
+      // 设置版权年份
+      document.getElementById("copyright-year").textContent =
+        new Date().getFullYear();
 
-        // 滚动到顶部/底部函数 (如果页面需要)
-        function scrollToTop() {
-            window.scrollTo({ top: 0, behavior: 'smooth' });
-        }
-        function scrollToBottom() {
-            window.scrollTo({ top: document.body.scrollHeight, behavior: 'smooth' });
+      // 滚动到顶部/底部函数 (如果页面需要)
+      function scrollToTop() {
+        window.scrollTo({ top: 0, behavior: "smooth" });
+      }
+      function scrollToBottom() {
+        window.scrollTo({
+          top: document.body.scrollHeight,
+          behavior: "smooth",
+        });
+      }
+
+      // 显示通知
+      function showNotification(message, type = "success", duration = 3000) {
+        const notification =
+          document.getElementById("notification") ||
+          createNotificationElement();
+        if (!notification) return;
+
+        notification.textContent = message;
+        notification.className = "notification show"; // Reset classes
+        if (type === "error") {
+          notification.classList.add("error");
         }
 
-        // 显示通知
-        function showNotification(message, type = 'success', duration = 3000) {
-            const notification = document.getElementById('notification') || createNotificationElement();
-            if (!notification) return;
-
-            notification.textContent = message;
-            notification.className = 'notification show'; // Reset classes
-            if (type === 'error') {
-                notification.classList.add('error');
-            }
-
-            // Clear previous timeout if exists
-            if (notification.timeoutId) {
-                clearTimeout(notification.timeoutId);
-            }
-
-            notification.timeoutId = setTimeout(() => {
-                notification.classList.remove('show');
-                // Optional: remove the element after fade out if dynamically created
-                // setTimeout(() => notification.remove(), 300);
-            }, duration);
+        // Clear previous timeout if exists
+        if (notification.timeoutId) {
+          clearTimeout(notification.timeoutId);
         }
 
-        // Helper to create notification element if it doesn't exist
-        function createNotificationElement() {
-            let notification = document.getElementById('notification');
-            if (!notification) {
-                notification = document.createElement('div');
-                notification.id = 'notification';
-                notification.className = 'notification';
-                document.body.appendChild(notification);
-            }
-            return notification;
+        notification.timeoutId = setTimeout(() => {
+          notification.classList.remove("show");
+          // Optional: remove the element after fade out if dynamically created
+          // setTimeout(() => notification.remove(), 300);
+        }, duration);
+      }
+
+      // Helper to create notification element if it doesn't exist
+      function createNotificationElement() {
+        let notification = document.getElementById("notification");
+        if (!notification) {
+          notification = document.createElement("div");
+          notification.id = "notification";
+          notification.className = "notification";
+          document.body.appendChild(notification);
         }
+        return notification;
+      }
 
-        // 页面刷新带加载状态
-        function refreshPage(button) {
-            if (button) {
-                const icon = button.querySelector('i');
-                if (icon) {
-                    icon.classList.add('loading-spin');
-                }
-            }
-            setTimeout(() => {
-                window.location.reload();
-            }, 300); // Short delay to show spinner
+      // 页面刷新带加载状态
+      function refreshPage(button) {
+        if (button) {
+          const icon = button.querySelector("i");
+          if (icon) {
+            icon.classList.add("loading-spin");
+          }
         }
+        setTimeout(() => {
+          window.location.reload();
+        }, 300); // Short delay to show spinner
+      }
 
-        // --- Version Check ---
-        const versionInfoContainer = document.getElementById('version-info-container');
+      // --- Version Check ---
+      const versionInfoContainer = document.getElementById(
+        "version-info-container"
+      );
 
-        async function fetchVersionInfo() {
-            if (!versionInfoContainer) return;
-            versionInfoContainer.innerHTML = '<span class="mx-1">|</span><span class="text-xs text-gray-400">检查更新中...</span>'; // Initial loading state
+      async function fetchVersionInfo() {
+        if (!versionInfoContainer) return;
+        versionInfoContainer.innerHTML =
+          '<span class="mx-1">|</span><span class="text-xs text-gray-700">检查更新中...</span>'; // Initial loading state
 
-            try {
-                const response = await fetch('/api/version/check');
-                if (!response.ok) {
-                    throw new Error(`HTTP error! status: ${response.status}`);
-                }
-                const data = await response.json();
+        try {
+          const response = await fetch("/api/version/check");
+          if (!response.ok) {
+            throw new Error(`HTTP error! status: ${response.status}`);
+          }
+          const data = await response.json();
 
-                let versionHtml = `<span class="mx-1">|</span><span class="text-xs text-gray-500">v${data.current_version}</span>`;
-                if (data.update_available) {
-                    versionHtml += `
+          let versionHtml = `<span class="mx-1">|</span><span class="text-xs text-gray-800">v${data.current_version}</span>`;
+          if (data.update_available) {
+            versionHtml += `
                         <span class="mx-1">|</span>
                         <a href="https://github.com/snailyp/gemini-balance/releases/latest" target="_blank" class="text-yellow-600 hover:text-yellow-800 transition duration-300 animate-pulse">
                             <i class="fas fa-arrow-up"></i> 新版本: v${data.latest_version}
                         </a>`;
-                } else if (data.error_message) {
-                     versionHtml += `
+          } else if (data.error_message) {
+            versionHtml += `
                         <span class="mx-1">|</span>
                         <span class="text-xs text-red-500" title="${data.error_message}">更新检查失败</span>`;
-                } else {
-                     versionHtml += `<span class="mx-1">|</span><span class="text-xs text-green-500">已是最新</span>`; // Indicate up-to-date
-                }
-                versionInfoContainer.innerHTML = versionHtml;
-
-            } catch (error) {
-                console.error('Error fetching version info:', error);
-                versionInfoContainer.innerHTML = `<span class="mx-1">|</span><span class="text-xs text-red-500" title="无法连接到服务器或解析响应">更新检查失败</span>`;
-            }
+          } else {
+            versionHtml += `<span class="mx-1">|</span><span class="text-xs text-green-500">已是最新</span>`; // Indicate up-to-date
+          }
+          versionInfoContainer.innerHTML = versionHtml;
+        } catch (error) {
+          console.error("Error fetching version info:", error);
+          versionInfoContainer.innerHTML = `<span class="mx-1">|</span><span class="text-xs text-red-500" title="无法连接到服务器或解析响应">更新检查失败</span>`;
         }
+      }
 
-        // Fetch immediately on load
-        fetchVersionInfo();
-
-        // Fetch periodically (e.g., every hour)
-        setInterval(fetchVersionInfo, 3600000); // 3600000 ms = 1 hour
+      // Fetch immediately on load
+      fetchVersionInfo();
 
+      // Fetch periodically (e.g., every hour)
+      setInterval(fetchVersionInfo, 3600000); // 3600000 ms = 1 hour
     </script>
     {% block body_scripts %}{% endblock %}
-</body>
-</html>
+  </body>
+</html>

app/utils/helpers.py

@@ -7,15 +7,12 @@ import base64
 import requests
 from typing import Dict, Any, List, Optional, Tuple
 from pathlib import Path
-import logging # Import logging
+import logging
 
 from app.core.constants import DATA_URL_PATTERN, IMAGE_URL_PATTERN, VALID_IMAGE_RATIOS
 
-# Define logger for helper functions if needed, or use specific loggers
-helper_logger = logging.getLogger("app.utils") # Or use a more specific logger if available
+helper_logger = logging.getLogger("app.utils")
 
-# Define project root and version file path here for get_current_version
-# Assuming this file is at app/utils/helpers.py
 PROJECT_ROOT = Path(__file__).resolve().parent.parent.parent
 VERSION_FILE_PATH = PROJECT_ROOT / "VERSION"
 
@@ -159,9 +156,8 @@ def is_valid_api_key(key: str) -> bool:
 
 def get_current_version(default_version: str = "0.0.0") -> str:
     """Reads the current version from the VERSION file."""
-    version_file = VERSION_FILE_PATH # Use Path object defined above
+    version_file = VERSION_FILE_PATH
     try:
-        # Use Path object's open method
         with version_file.open('r', encoding='utf-8') as f:
             version = f.read().strip()
         if not version:

app/utils/uploader.py

@@ -261,18 +261,20 @@ class PicGoUploader(ImageUploader):
 
 class CloudFlareImgBedUploader(ImageUploader):
     """CloudFlare图床上传器"""
-    
-    def __init__(self, auth_code: str, api_url: str):
+
+    def __init__(self, auth_code: str, api_url: str, upload_folder: str = ""):
         """
         初始化CloudFlare图床上传器
         
         Args:
             auth_code: 认证码
             api_url: 上传API地址
+            upload_folder: 上传文件夹路径（可选）
         """
         self.auth_code = auth_code
         self.api_url = api_url
-        
+        self.upload_folder = upload_folder
+
     def upload(self, file: bytes, filename: str) -> UploadResponse:
         """
         上传图片到CloudFlare图床
@@ -288,12 +290,16 @@ class CloudFlareImgBedUploader(ImageUploader):
             UploadError: 上传失败时抛出异常
         """
         try:
-            # 准备请求URL（添加认证码参数，如果存在）
+            # 准备请求URL参数
+            params = []
+            if self.upload_folder:
+                params.append(f"uploadFolder={self.upload_folder}")
             if self.auth_code:
-                request_url = f"{self.api_url}?authCode={self.auth_code}&uploadNameType=origin"
-            else:
-                request_url = f"{self.api_url}?uploadNameType=origin"
-            
+                params.append(f"authCode={self.auth_code}")
+            params.append("uploadNameType=origin")
+
+            request_url = f"{self.api_url}?{'&'.join(params)}"
+
             # 准备文件数据
             files = {
                 "file": (filename, file)
@@ -388,6 +394,7 @@ class ImageUploaderFactory:
         elif provider == "cloudflare_imgbed":
             return CloudFlareImgBedUploader(
                 credentials["auth_code"],
-                credentials["base_url"]
+                credentials["base_url"],
+                credentials.get("upload_folder", ""),
             )
         raise ValueError(f"Unknown provider: {provider}")

files/dataocean.svg

@@ -0,0 +1,71 @@
+<?xml version="1.0" encoding="utf-8"?>
+<!-- Generator: Adobe Illustrator 19.0.0, SVG Export Plug-In . SVG Version: 6.00 Build 0)  -->
+<svg version="1.1" id="Layer_1" xmlns="http://www.w3.org/2000/svg" xmlns:xlink="http://www.w3.org/1999/xlink" x="0px" y="0px"
+	 viewBox="0 0 603 103" style="enable-background:new 0 0 603 103;" xml:space="preserve">
+<style type="text/css">
+	.st0{fill:#0080FF;}
+	.st1{fill-rule:evenodd;clip-rule:evenodd;fill:#0080FF;}
+</style>
+<g id="XMLID_2369_">
+	<g id="XMLID_2638_">
+		<g id="XMLID_2639_">
+			<g>
+				<g id="XMLID_44_">
+					<g id="XMLID_48_">
+						<path id="XMLID_49_" class="st0" d="M52.1,102.1l0-19.6c20.8,0,36.8-20.6,28.9-42.4C78,32,71.6,25.5,63.5,22.6
+							c-21.8-7.9-42.4,8.1-42.4,28.9c0,0,0,0,0,0l-19.6,0c0-33.1,32-58.9,66.7-48.1c15.2,4.7,27.2,16.8,31.9,31.9
+							C110.9,70.1,85.2,102.1,52.1,102.1z"/>
+					</g>
+					<polygon id="XMLID_47_" class="st1" points="52.1,82.5 32.6,82.5 32.6,63 32.6,63 52.1,63 52.1,63 					"/>
+					<polygon id="XMLID_46_" class="st1" points="32.6,97.5 17.6,97.5 17.6,97.5 17.6,82.5 32.6,82.5 32.6,97.5 					"/>
+					<polygon id="XMLID_45_" class="st1" points="17.6,82.5 5,82.5 5,82.5 5,70 5,70 17.6,70 17.6,70 					"/>
+				</g>
+			</g>
+		</g>
+	</g>
+	<g id="XMLID_2370_">
+		<path id="XMLID_2635_" class="st0" d="M181.5,30.2c-5.8-4-13-6.1-21.4-6.1h-18.3v58.1h18.3c8.4,0,15.6-2.1,21.4-6.4
+			c3.2-2.2,5.7-5.4,7.4-9.3c1.7-3.9,2.6-8.5,2.6-13.7c0-5.1-0.9-9.7-2.6-13.6C187.2,35.4,184.7,32.3,181.5,30.2z M152.5,34h5.8
+			c6.4,0,11.7,1.3,15.7,3.7c4.4,2.7,6.7,7.8,6.7,15.1c0,7.6-2.3,12.9-6.7,15.8h0c-3.8,2.5-9.1,3.8-15.6,3.8h-5.8V34z"/>
+		<path id="XMLID_2634_" class="st0" d="M204.3,23.4c-1.8,0-3.3,0.6-4.5,1.8c-1.2,1.2-1.9,2.7-1.9,4.4c0,1.8,0.6,3.3,1.9,4.5
+			c1.2,1.2,2.7,1.9,4.5,1.9c1.8,0,3.3-0.6,4.5-1.9c1.2-1.2,1.9-2.8,1.9-4.5c0-1.8-0.6-3.3-1.9-4.4C207.6,24,206,23.4,204.3,23.4z"/>
+		<rect id="XMLID_2564_" x="199" y="41.3" class="st0" width="10.3" height="41"/>
+		<path id="XMLID_2561_" class="st0" d="M246.8,44.7c-3.1-2.8-6.6-4.4-10.3-4.4c-5.7,0-10.4,2-14.1,5.8c-3.7,3.8-5.5,8.8-5.5,14.7
+			c0,5.8,1.8,10.7,5.5,14.7c3.7,3.8,8.4,5.8,14.1,5.8c4,0,7.4-1.1,10.2-3.3V79c0,3.4-0.9,6-2.7,7.9c-1.8,1.8-4.3,2.7-7.4,2.7
+			c-4.8,0-7.7-1.9-11.4-6.8l-7,6.7l0.2,0.3c1.5,2.1,3.8,4.2,6.9,6.2c3.1,2,6.9,3,11.5,3c6.1,0,11.1-1.9,14.7-5.6
+			c3.7-3.7,5.5-8.7,5.5-14.9V41.3h-10.1V44.7z M244.1,68.9c-1.8,2-4.1,3-7.1,3c-3,0-5.3-1-7-3c-1.8-2-2.7-4.7-2.7-8
+			c0-3.3,0.9-6.1,2.7-8.1c1.8-2,4.1-3.1,7-3.1c3,0,5.3,1,7.1,3.1c1.8,2,2.7,4.8,2.7,8.1C246.8,64.2,245.8,66.9,244.1,68.9z"/>
+		<rect id="XMLID_2560_" x="265.7" y="41.3" class="st0" width="10.3" height="41"/>
+		<path id="XMLID_2552_" class="st0" d="M271,23.4c-1.8,0-3.3,0.6-4.5,1.8c-1.2,1.2-1.9,2.7-1.9,4.4c0,1.8,0.6,3.3,1.9,4.5
+			c1.2,1.2,2.7,1.9,4.5,1.9c1.8,0,3.3-0.6,4.5-1.9c1.2-1.2,1.9-2.8,1.9-4.5c0-1.8-0.6-3.3-1.9-4.4C274.3,24,272.7,23.4,271,23.4z"/>
+		<path id="XMLID_2509_" class="st0" d="M298.6,30.3h-10.1v11.1h-5.9v9.4h5.9v17c0,5.3,1.1,9.1,3.2,11.3c2.1,2.2,5.8,3.3,11.1,3.3
+			c1.7,0,3.4-0.1,5-0.2l0.5,0v-9.4l-3.5,0.2c-2.5,0-4.1-0.4-4.9-1.3c-0.8-0.9-1.2-2.7-1.2-5.4V50.7h9.6v-9.4h-9.6V30.3z"/>
+		<rect id="XMLID_2508_" x="356.5" y="24.1" class="st0" width="10.3" height="58.1"/>
+		<path id="XMLID_2470_" class="st0" d="M470.9,67.6c-1.8,2.1-3.7,3.9-5.2,4.8v0c-1.4,0.9-3.2,1.4-5.3,1.4c-3,0-5.5-1.1-7.5-3.4
+			c-2-2.3-3-5.2-3-8.7s1-6.4,2.9-8.6c2-2.3,4.4-3.4,7.4-3.4c3.3,0,6.8,2.1,9.8,5.6l6.8-6.5l0,0c-4.4-5.8-10.1-8.5-16.9-8.5
+			c-5.7,0-10.6,2.1-14.6,6.1c-4,4-6,9.2-6,15.3s2,11.2,6,15.3c4,4.1,8.9,6.1,14.6,6.1c7.5,0,13.5-3.2,17.5-9.1L470.9,67.6z"/>
+		<path id="XMLID_2460_" class="st0" d="M513.2,47c-1.5-2-3.5-3.7-5.9-4.9c-2.5-1.2-5.3-1.8-8.5-1.8c-5.8,0-10.5,2.1-14,6.3
+			c-3.4,4.2-5.2,9.3-5.2,15.4c0,6.2,1.9,11.3,5.7,15.3c3.7,3.9,8.8,5.9,14.9,5.9c6.9,0,12.7-2.8,16.9-8.4l0.2-0.3l-6.7-6.5l0,0
+			c-0.6,0.8-1.5,1.6-2.3,2.4c-1,1-2,1.7-3,2.2c-1.5,0.8-3.3,1.1-5.2,1.1c-2.9,0-5.2-0.8-7-2.5c-1.7-1.5-2.7-3.6-2.9-6.2h27.3
+			l0.1-3.8c0-2.7-0.4-5.2-1.1-7.6C515.8,51.3,514.7,49.1,513.2,47z M490.7,56.7c0.5-2,1.4-3.6,2.7-4.9c1.4-1.4,3.2-2.1,5.4-2.1
+			c2.5,0,4.4,0.7,5.7,2.1c1.2,1.3,1.9,2.9,2.1,4.8H490.7z"/>
+		<path id="XMLID_2456_" class="st0" d="M552.8,44.4L552.8,44.4c-3.1-2.7-7.4-4-12.8-4c-3.4,0-6.6,0.8-9.5,2.2
+			c-2.7,1.4-5.3,3.6-7,6.6l0.1,0.1l6.6,6.3c2.7-4.3,5.7-5.8,9.7-5.8c2.2,0,3.9,0.6,5.3,1.7c1.4,1.1,2,2.6,2,4.4v2
+			c-2.6-0.8-5.1-1.2-7.6-1.2c-5.1,0-9.3,1.2-12.4,3.6c-3.1,2.4-4.7,5.9-4.7,10.2c0,3.8,1.3,7,4,9.3c2.7,2.2,6,3.4,9.9,3.4
+			c3.9,0,7.6-1.6,10.9-4.3v3.4h10.1V55.9C557.6,51,556,47.1,552.8,44.4z M534.5,66.6c1.2-0.8,2.8-1.2,4.9-1.2c2.5,0,5.1,0.5,7.8,1.5
+			v4C545,73,542,74,538.3,74c-1.8,0-3.2-0.4-4.1-1.2c-0.9-0.8-1.4-1.7-1.4-3C532.8,68.5,533.4,67.4,534.5,66.6z"/>
+		<path id="XMLID_2454_" class="st0" d="M597.2,45.2c-2.9-3.2-6.9-4.8-12-4.8c-4.1,0-7.4,1.2-9.9,3.5v-2.5h-10.1v41h10.3V59.7
+			c0-3.1,0.7-5.6,2.2-7.3c1.5-1.8,3.4-2.6,6.1-2.6c2.3,0,4.1,0.8,5.4,2.3c1.3,1.6,2,3.7,2,6.4v23.7h10.3V58.5
+			C601.5,52.9,600.1,48.4,597.2,45.2z"/>
+		<path id="XMLID_2450_" class="st0" d="M343.6,44.4L343.6,44.4c-3.1-2.7-7.4-4-12.8-4c-3.4,0-6.6,0.8-9.5,2.2
+			c-2.7,1.4-5.3,3.6-7,6.6l0.1,0.1l6.6,6.3c2.7-4.3,5.7-5.8,9.7-5.8c2.2,0,3.9,0.6,5.3,1.7c1.4,1.1,2,2.6,2,4.4v2
+			c-2.6-0.8-5.1-1.2-7.6-1.2c-5.1,0-9.3,1.2-12.4,3.6c-3.1,2.4-4.7,5.9-4.7,10.2c0,3.8,1.3,7,4,9.3c2.7,2.2,6,3.4,9.9,3.4
+			c3.9,0,7.6-1.6,10.9-4.3v3.4h10.1V55.9C348.3,51,346.7,47.1,343.6,44.4z M325.3,66.6c1.2-0.8,2.8-1.2,4.9-1.2
+			c2.5,0,5.1,0.5,7.8,1.5v4c-2.2,2.1-5.2,3.1-8.9,3.1c-1.8,0-3.2-0.4-4.1-1.2c-0.9-0.8-1.4-1.7-1.4-3
+			C323.6,68.5,324.1,67.4,325.3,66.6z"/>
+		<path id="XMLID_2371_" class="st0" d="M404.2,83.1c-16.5,0-30-13.4-30-30s13.4-30,30-30c16.5,0,30,13.4,30,30
+			S420.7,83.1,404.2,83.1z M404.2,33.8c-10.7,0-19.4,8.7-19.4,19.4s8.7,19.4,19.4,19.4c10.7,0,19.4-8.7,19.4-19.4
+			S414.9,33.8,404.2,33.8z"/>
+	</g>
+</g>
+</svg>

requirements.txt

@@ -9,11 +9,11 @@ uvicorn
 google-genai
 jinja2
 python-multipart
-cryptography # 支持 MySQL 8+ caching_sha2_password 验证
-# 数据库相关依赖
+cryptography
 pymysql
 sqlalchemy
 aiomysql
+aiosqlite
 databases
 python-dotenv
 apscheduler