使用说明
🎓 XJTLU Academic Navigator
An MCP (Model-Context-Protocol) based AI assistant for XJTLU Economics students. This project demonstrates how commercial AI systems can integrate fragmented academic data sources without requiring official API access.
🔍 Project Vision
Current academic systems at XJTLU (Learning Mall, e-Bridge) operate in isolation, forcing students to manually synthesize information. Our solution provides:
- ✨ Course explanations contextualized to career goals (e.g., HKU MFWM preparation)
- 📅 Personalized semester planning based on academic background and credit constraints
- 🎯 Career pathway analysis mapping XJTLU Economics courses to target graduate programs
- 🔍 Prerequisite validation and workload assessment for optimal course selection
🏗️ MCP Architecture
Our system implements a strict Model-Context-Protocol separation:
| Component | Role | Implementation |
|---|---|---|
| Model | AI reasoning engine | DeepSeek API integration with professional fallbacks |
| Context | Academic knowledge base | Mock data derived from official XJTLU Economics programme specification |
| Protocol | Standardized communication | JSON message format with intent routing and validation rules |
Key Architectural Components:
- Dispatcher: Intent recognition and request routing
- Orchestrator: Multi-agent coordination for complex queries
- Course Service: Semantic search over curriculum data
- Planning Service: Rule-based semester planning with AI enhancement
- AI Service: Unified interface for LLM interactions with graceful degradation
🛡️ Compliance & Ethics
This project strictly adheres to ethical and compliance standards:
- ✅ No access to real student data - all data is synthetic/mock
- ✅ No scraping of XJTLU systems - course data derived from publicly available programme specifications
- ✅ Transparent AI sourcing - all AI-generated content is clearly marked
- ✅ Privacy by design - no user data persistence in demo version
- ✅ Academic integrity - all course descriptions accurately reflect official XJTLU curriculum
🚀 Getting Started
Prerequisites
- Python 3.9+
- Node.js (optional, for frontend enhancements)
- DeepSeek API Key (optional - mock mode available)
Installation
# Clone repository
git clone https://github.com/yourusername/xjtlu-academic-navigator.git
cd xjtlu-academic-navigator
# Create virtual environment
python -m venv venv
source venv/bin/activate # Linux/Mac
# venv\Scripts\activate # Windows
# Install dependencies
pip install -r requirements.txt
# Set up environment variables
cp .env.example .env
# Edit .env with your DeepSeek API key (or keep USE_MOCK_AI=true)
