Add ModelScope API-Inference provider support to Copilot

README.md

@@ -4,7 +4,8 @@ MGit是一款基于Python的Markdown笔记应用，结合了直观的编辑体
 
 ## 主要特点
 
-- **AI Copilot写作助手**：集成SiliconFlow大语言模型，提供智能写作辅助
+- **AI Copilot写作助手**：集成大语言模型，提供智能写作辅助
+  - **多提供商支持**：支持SiliconFlow和ModelScope API-Inference两种服务提供商
   - **行内补全**：根据上下文智能生成续写内容
   - **编辑模式**：根据指令优化和修改文本
   - **创作模式**：从提示词生成完整文档

docs/copilot_config.md

@@ -9,50 +9,118 @@ Typically located at: `~/.config/mgit/config.json` or similar location depending
 
 ## Configuration Keys
 
+### SiliconFlow Provider (Default)
+
 ```json
 {
   "copilot": {
     "enabled": true,
+    "provider": "siliconflow",
     "api_key": "your-siliconflow-api-key-here",
     "model": "Qwen/Qwen2.5-7B-Instruct"
   }
 }
 ```
 
+### ModelScope API-Inference Provider
+
+```json
+{
+  "copilot": {
+    "enabled": true,
+    "provider": "modelscope",
+    "api_key": "your-modelscope-api-key-here",
+    "model": "qwen/Qwen2.5-7B-Instruct"
+  }
+}
+```
+
+## Providers
+
+MGit Copilot now supports two LLM service providers:
+
+### 1. SiliconFlow (Default)
+- **Website**: https://siliconflow.cn
+- **API Documentation**: https://docs.siliconflow.cn
+- **Features**: Fast inference, multiple models, competitive pricing
+- **Provider ID**: `siliconflow`
+
+### 2. ModelScope API-Inference (New!)
+- **Website**: https://www.modelscope.cn
+- **API Documentation**: https://www.modelscope.cn/docs/model-service/API-Inference/intro
+- **Features**: Official Alibaba Cloud service, reliable infrastructure
+- **Provider ID**: `modelscope`
+
 ## Available Models
 
+### SiliconFlow Models
+
 You can use any of these models with SiliconFlow:
 
-### Qwen Series (Recommended)
+#### Qwen Series (Recommended)
 - `Qwen/Qwen2.5-7B-Instruct` - Best balance of speed and quality (default)
 - `Qwen/Qwen2.5-14B-Instruct` - Higher quality, slightly slower
 - `Qwen/Qwen2.5-32B-Instruct` - Highest quality, slower
 
-### DeepSeek
+#### DeepSeek
 - `deepseek-ai/DeepSeek-V2.5` - Strong reasoning, good for technical content
 
-### Meta LLaMA
+#### Meta LLaMA
 - `meta-llama/Meta-Llama-3.1-8B-Instruct` - Open source alternative
 
+### ModelScope API-Inference Models
+
+Available models with ModelScope:
+
+#### Qwen Series
+- `qwen/Qwen2.5-7B-Instruct` - Best balance of speed and quality (default)
+- `qwen/Qwen2.5-14B-Instruct` - Higher quality
+- `qwen/Qwen2.5-32B-Instruct` - Highest quality
+
+**Note**: Model names use lowercase namespace with ModelScope (e.g., `qwen/` instead of `Qwen/`)
+
 ## Getting API Key
 
+### For SiliconFlow
+
 1. Visit https://siliconflow.cn
 2. Register or login to your account
 3. Navigate to API Keys section
 4. Create a new API key
 5. Copy the key to Copilot settings in MGit
 
+### For ModelScope
+
+1. Visit https://www.modelscope.cn
+2. Register or login with your Alibaba Cloud account
+3. Navigate to the API-Inference section
+4. Create a new API key
+5. Copy the key to Copilot settings in MGit
+
 ## Configuration via UI
 
 The recommended way to configure Copilot is through the UI:
 
 1. Open MGit
 2. Go to menu: `Copilot > Copilot设置`
-3. Enter your API key
-4. Select your preferred model
-5. Check "启用 Copilot"
-6. Click "测试连接" to verify
-7. Click "保存"
+3. Select your preferred provider (SiliconFlow or ModelScope)
+4. Enter your API key
+5. Select your preferred model
+6. Check "启用 Copilot"
+7. Click "测试连接" to verify
+8. Click "保存"
+
+## Switching Between Providers
+
+You can switch between SiliconFlow and ModelScope at any time:
+
+1. Open Copilot settings
+2. Select different provider from dropdown
+3. Enter the API key for the new provider
+4. Select an appropriate model for that provider
+5. Test connection and save
+
+**Note**: Each provider requires its own API key. Make sure to obtain and configure API keys for both providers if you plan to switch between them.
 
 ## Security Notes
 
@@ -65,29 +133,52 @@ The recommended way to configure Copilot is through the UI:
 
 You can also set API key via environment variable:
 
+### For SiliconFlow
 ```bash
 export SILICONFLOW_API_KEY="your-api-key-here"
 ```
 
+### For ModelScope
+```bash
+export MODELSCOPE_API_KEY="your-api-key-here"
+```
+
 Note: UI configuration takes precedence over environment variables.
 
 ## Troubleshooting
 
 ### API Key Not Working
 - Verify key is correct (no extra spaces)
-- Check account balance on SiliconFlow
+- Check account balance on your provider's website
 - Ensure network connectivity
 - Try "测试连接" in settings dialog
+- Verify you're using the correct API key for the selected provider
 
 ### Model Not Available
-- Some models may require specific API tier
-- Check SiliconFlow documentation for model availability
+- Some models may require specific API tier or region
+- Check provider documentation for model availability
 - Try using default model first
+- For ModelScope: Ensure model name uses correct case (lowercase namespace)
+- For SiliconFlow: Ensure model name uses correct case (uppercase namespace)
 
 ### Connection Timeout
 - Check internet connection
 - Verify firewall settings
 - Try increasing timeout in code if needed
+- Some regions may have slower access to certain providers
+
+### Provider-Specific Issues
+
+#### SiliconFlow
+- Check status at https://siliconflow.cn
+- Verify API tier and quotas
+- Review rate limits
+
+#### ModelScope
+- Check Alibaba Cloud account status
+- Verify API-Inference service is enabled
+- Review service quotas and limits
+- Check region availability
 
 ## Advanced Configuration
 
@@ -99,15 +190,22 @@ For advanced users, you can modify additional parameters by editing the source c
 
 ## Rate Limits
 
-Be aware of SiliconFlow API rate limits:
+Be aware of API rate limits from your provider:
+
+### SiliconFlow
 - Free tier: Limited requests per minute
 - Paid tiers: Higher limits
+- Check your plan at https://siliconflow.cn
 
-Check your plan at https://siliconflow.cn
+### ModelScope API-Inference
+- Limits depend on your Alibaba Cloud account tier
+- Check quotas in your ModelScope console
+- Visit https://www.modelscope.cn for details
 
 ## Support
 
 For issues or questions:
 - Check the [Copilot Guide](./copilot_guide.md)
-- Submit GitHub issue
-- Contact SiliconFlow support for API issues
+- Submit GitHub issue for MGit-related problems
+- Contact SiliconFlow support for SiliconFlow API issues
+- Contact Alibaba Cloud support for ModelScope API issues

docs/copilot_guide.md

@@ -53,13 +53,29 @@ MGit Copilot 是一个专注于写作的AI助手，集成了最新的大语言
 
 ### 1. 配置API密钥
 
+MGit Copilot 支持两种大模型服务提供商：
+
+#### 选项1: SiliconFlow (默认)
 1. 访问 [SiliconFlow官网](https://siliconflow.cn) 注册账号并获取API密钥
 2. 在MGit中打开 `Copilot > Copilot设置`
-3. 输入API密钥
-4. 选择模型（推荐使用 Qwen/Qwen2.5-7B-Instruct）
-5. 勾选"启用Copilot"
-6. 点击"测试连接"验证配置
-7. 点击"保存"
+3. 选择 "SiliconFlow" 作为提供商
+4. 输入API密钥
+5. 选择模型（推荐使用 Qwen/Qwen2.5-7B-Instruct）
+6. 勾选"启用Copilot"
+7. 点击"测试连接"验证配置
+8. 点击"保存"
+
+#### 选项2: ModelScope API-Inference (新增!)
+1. 访问 [ModelScope官网](https://www.modelscope.cn) 注册账号并获取API密钥
+2. 在MGit中打开 `Copilot > Copilot设置`
+3. 选择 "ModelScope" 作为提供商
+4. 输入API密钥
+5. 选择模型（推荐使用 qwen/Qwen2.5-7B-Instruct）
+6. 勾选"启用Copilot"
+7. 点击"测试连接"验证配置
+8. 点击"保存"
+
+**注意**: 两个提供商的模型名称格式略有不同（ModelScope使用小写命名空间）
 
 ### 2. 显示Copilot面板
 
@@ -208,8 +224,9 @@ MGit Copilot 是一个专注于写作的AI助手，集成了最新的大语言
 **A**: 
 1. 检查API密钥是否正确
 2. 确认网络连接正常
-3. 验证SiliconFlow账户余额
-4. 查看错误信息，按提示操作
+3. 验证服务提供商账户余额（SiliconFlow或ModelScope）
+4. 确保选择的提供商与API密钥匹配
+5. 查看错误信息，按提示操作
 
 ### Q: 生成的内容质量不理想？
 
@@ -253,12 +270,20 @@ MGit Copilot 是一个专注于写作的AI助手，集成了最新的大语言
 ## 技术支持
 
 - **文档**: 参考本指南和项目README
+- **配置指南**: [Copilot配置文档](./copilot_config.md)
 - **问题反馈**: 提交GitHub Issue
 - **功能建议**: 通过PR或Issue提出
-- **API文档**: https://docs.siliconflow.cn
+- **SiliconFlow API文档**: https://docs.siliconflow.cn
+- **ModelScope API文档**: https://www.modelscope.cn/docs/model-service/API-Inference/intro
 
 ## 更新日志
 
+### Version 1.1.0 (2026-01-07)
+
+- 新增ModelScope API-Inference服务提供商支持
+- 支持在SiliconFlow和ModelScope之间切换
+- 更新配置文档，详细说明两个提供商的使用
+
 ### Version 1.0.0 (2026-01-03)
 
 - 初始版本发布