[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-72542":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":10,"language":11,"languages":10,"totalLinesOfCode":10,"stars":12,"forks":13,"watchers":14,"openIssues":15,"contributorsCount":16,"subscribersCount":16,"size":16,"stars1d":16,"stars7d":16,"stars30d":17,"stars90d":16,"forks30d":16,"starsTrendScore":16,"compositeScore":18,"rankGlobal":10,"rankLanguage":10,"license":19,"archived":20,"fork":20,"defaultBranch":21,"hasWiki":22,"hasPages":20,"topics":23,"createdAt":10,"pushedAt":10,"updatedAt":24,"readmeContent":25,"aiSummary":26,"trendingCount":16,"starSnapshotCount":16,"syncStatus":27,"lastSyncTime":28,"discoverSource":29},72542,"deepseek-engineer","Doriandarko\u002Fdeepseek-engineer","Doriandarko","A powerful coding assistant application that integrates with the DeepSeek API to process user conversations and generate structured JSON responses. Through an intuitive command-line interface, it can read local file contents, create new files, and apply diff edits to existing files in real time.","",null,"Python",2234,363,31,10,0,1,59.78,"MIT License",false,"main",true,[],"2026-06-12 04:01:06","# DeepSeek Engineer v2 🐋\n\n## Overview\n\nDeepSeek Engineer v2 is a powerful AI-powered coding assistant that provides an interactive terminal interface for seamless code development. It integrates with DeepSeek's advanced reasoning models to offer intelligent file operations, code analysis, and development assistance through natural conversation and function calling.\n\n## 🚀 Latest Update: Function Calling Architecture\n\n**Version 2.0** introduces a big upgrade from structured JSON output to native function calling, providing:\n- **Natural conversations** with the AI without rigid response formats\n- **Automatic file operations** through intelligent function calls\n- **Real-time reasoning visibility** with Chain of Thought (CoT) capabilities\n- **Enhanced reliability** and better error handling\n\n## Key Features\n\n### 🧠 **AI Capabilities**\n- **Elite Software Engineering**: Decades of experience across all programming domains\n- **Chain of Thought Reasoning**: Visible thought process before providing solutions\n- **Code Analysis & Discussion**: Expert-level insights and optimization suggestions\n- **Intelligent Problem Solving**: Automatic file reading and context understanding\n\n### 🛠️ **Function Calling Tools**\nThe AI can automatically execute these operations when needed:\n\n#### `read_file(file_path: str)`\n- Read single file content with automatic path normalization\n- Built-in error handling for missing or inaccessible files\n- **Automatic**: AI can read any file you mention or reference in conversation\n\n#### `read_multiple_files(file_paths: List[str])`\n- Batch read multiple files efficiently\n- Formatted output with clear file separators\n\n#### `create_file(file_path: str, content: str)`\n- Create new files or overwrite existing ones\n- Automatic directory creation and safety checks\n\n#### `create_multiple_files(files: List[Dict])`\n- Create multiple files in a single operation\n- Perfect for scaffolding projects or creating related files\n\n#### `edit_file(file_path: str, original_snippet: str, new_snippet: str)`\n- Precise snippet-based file editing\n- Safe replacement with exact matching\n\n### 📁 **File Operations**\n\n#### **Automatic File Reading (Recommended)**\nThe AI can automatically read files you mention:\n```\nYou> Can you review the main.py file and suggest improvements?\n→ AI automatically calls read_file(\"main.py\")\n\nYou> Look at src\u002Futils.py and tests\u002Ftest_utils.py\n→ AI automatically calls read_multiple_files([\"src\u002Futils.py\", \"tests\u002Ftest_utils.py\"])\n```\n\n#### **Manual Context Addition (Optional)**\nFor when you want to preload files into conversation context:\n- **`\u002Fadd path\u002Fto\u002Ffile`** - Include single file in conversation context\n- **`\u002Fadd path\u002Fto\u002Ffolder`** - Include entire directory (with smart filtering)\n\n**Note**: The `\u002Fadd` command is mainly useful when you want to provide extra context upfront. The AI can read files automatically via function calls whenever needed during the conversation.\n\n### 🎨 **Rich Terminal Interface**\n- **Color-coded feedback** (green for success, red for errors, yellow for warnings)\n- **Real-time streaming** with visible reasoning process\n- **Structured tables** for diff previews\n- **Progress indicators** for long operations\n\n### 🛡️ **Security & Safety**\n- **Path normalization** and validation\n- **Directory traversal protection**\n- **File size limits** (5MB per file)\n- **Binary file detection** and exclusion\n\n## Getting Started\n\n### Prerequisites\n1. **DeepSeek API Key**: Get your API key from [DeepSeek Platform](https:\u002F\u002Fplatform.deepseek.com)\n2. **Python 3.11+**: Required for optimal performance\n\n### Installation\n\n1. **Clone the repository**:\n   ```bash\n   git clone \u003Crepository-url>\n   cd deepseek-engineer\n   ```\n\n2. **Set up environment**:\n   ```bash\n   # Create .env file\n   echo \"DEEPSEEK_API_KEY=your_api_key_here\" > .env\n   ```\n\n3. **Install dependencies** (choose one method):\n\n   #### Using uv (recommended - faster)\n   ```bash\n   uv venv\n   uv run deepseek-eng.py\n   ```\n\n   #### Using pip\n   ```bash\n   pip install -r requirements.txt\n   python3 deepseek-eng.py\n   ```\n\n### Usage Examples\n\n#### **Natural Conversation with Automatic File Operations**\n```\nYou> Can you read the main.py file and create a test file for it?\n\n💭 Reasoning: I need to first read the main.py file to understand its structure...\n\n🤖 Assistant> I'll read the main.py file first to understand its structure.\n⚡ Executing 1 function call(s)...\n→ read_file\n✓ Read file 'main.py'\n\n🔄 Processing results...\nNow I'll create comprehensive tests based on the code structure I found.\n⚡ Executing 1 function call(s)...\n→ create_file\n✓ Created\u002Fupdated file at 'test_main.py'\n\nI've analyzed main.py and created comprehensive tests covering all the main functions...\n```\n\n#### **Automatic Multi-File Analysis**\n```\nYou> Compare the implementation in utils.py with the tests in test_utils.py\n\n💭 Reasoning: I need to read both files to compare them...\n\n🤖 Assistant> I'll read both files to analyze the implementation and tests.\n⚡ Executing 1 function call(s)...\n→ read_multiple_files\n✓ Read files: utils.py, test_utils.py\n\n🔄 Processing results...\nAfter analyzing both files, I can see several areas where the tests could be improved...\n```\n\n#### **Manual Context Loading (Optional)**\n```\nYou> \u002Fadd src\u002F\n\n✓ Added folder 'src\u002F' to conversation.\n📁 Added files: (15 files)\n  📄 src\u002Futils.py\n  📄 src\u002Fmodels.py\n  ...\n\nYou> Now review this codebase structure\n\n🤖 Assistant> I've reviewed your codebase and found several areas for improvement:\n\n1. **Error Handling**: The utils.py file could benefit from more robust error handling...\n```\n\n## Technical Details\n\n### **Model**: DeepSeek-Reasoner\n- Powered by DeepSeek-R1 with Chain-of-Thought reasoning\n- Real-time reasoning visibility during processing\n- Enhanced problem-solving capabilities\n\n### **Function Call Execution Flow**\n1. **User Input** → Natural language request\n2. **AI Reasoning** → Visible thought process (CoT)\n3. **Function Calls** → Automatic tool execution\n4. **Real-time Feedback** → Operation status and results\n5. **Follow-up Response** → AI processes results and responds\n\n### **Streaming Architecture**\n- **Triple-stream processing**: reasoning + content + tool_calls\n- **Real-time tool execution** during streaming\n- **Automatic follow-up** responses after tool completion\n- **Error recovery** and graceful degradation\n\n## Advanced Features\n\n### **Intelligent Context Management**\n- **Automatic file detection** from user messages\n- **Smart conversation cleanup** to prevent token overflow\n- **File content preservation** across conversation history\n- **Tool message integration** for complete operation tracking\n\n### **Batch Operations**\n```\nYou> Create a complete Flask API with models, routes, and tests\n\n🤖 Assistant> I'll create a complete Flask API structure for you.\n⚡ Executing 1 function call(s)...\n→ create_multiple_files\n✓ Created 4 files: app.py, models.py, routes.py, test_api.py\n```\n\n### **Project Analysis**\n```\nYou> \u002Fadd .\nYou> Analyze this entire project and suggest a refactoring plan\n\n🤖 Assistant> ⚡ Executing 1 function call(s)...\n→ read_multiple_files\nBased on my analysis of your project, here's a comprehensive refactoring plan...\n```\n\n## File Operations Comparison\n\n| Method | When to Use | How It Works |\n|--------|-------------|--------------|\n| **Automatic Reading** | Most cases - just mention files | AI automatically calls `read_file()` when you reference files |\n| **`\u002Fadd` Command** | Preload context, bulk operations | Manually adds files to conversation context upfront |\n\n**Recommendation**: Use natural conversation - the AI will automatically read files as needed. Use `\u002Fadd` only when you want to provide extra context upfront.\n\n## Troubleshooting\n\n### **Common Issues**\n\n**API Key Not Found**\n```bash\n# Make sure .env file exists with your API key\necho \"DEEPSEEK_API_KEY=your_key_here\" > .env\n```\n\n**Import Errors**\n```bash\n# Install dependencies\nuv sync  # or pip install -r requirements.txt\n```\n\n**File Permission Errors**\n- Ensure you have write permissions in the working directory\n- Check file paths are correct and accessible\n\n## Contributing\n\nThis is an experimental project showcasing DeepSeek reasoning model capabilities. Contributions are welcome!\n\n### **Development Setup**\n```bash\ngit clone \u003Crepository-url>\ncd deepseek-engineer\nuv venv\nuv sync\n```\n\n### **Run**\n```bash\n# Run the application (preferred)\nuv run deepseek-eng.py\n```\nor\n```bash\npython3 deepseek-eng.py\n```\n\n## License\n\nThis project is licensed under the MIT License - see the [LICENSE](LICENSE) file for details.\n\nThis project is experimental and developed for testing DeepSeek reasoning model capabilities.\n\n---\n\n> **Note**: This is an experimental project developed to explore the capabilities of DeepSeek's reasoning model with function calling. The AI can automatically read files you mention in conversation, while the `\u002Fadd` command is available for when you want to preload context. Use responsibly and enjoy the enhanced AI pair programming experience! 🚀\n\n","DeepSeek Engineer v2 是一款强大的AI编码助手，通过与DeepSeek API集成处理用户对话并生成结构化的JSON响应。它具备自然语言交流、自动文件操作和实时代码编辑等核心功能，支持通过直观的命令行界面读取本地文件内容、创建新文件以及对现有文件进行实时差异编辑。该项目采用Python编写，并引入了函数调用架构以提高交互性和可靠性。适用于需要高效开发流程、代码分析及优化建议的软件开发场景中，特别适合那些希望在日常编码工作中获得智能辅助的开发者使用。",2,"2026-06-11 03:42:29","high_star"]