AI Code Review Assistant is a full-stack web application that helps you analyze code, detect plagiarism, visualize logic, and receive AI-driven suggestions across multiple programming languages. It is designed for students, educators, and developers who want private, smart, and visual code reviews without relying on cloud services.
✅ Static Code Analysis
- Supports Java, Python, JavaScript, C, C++
- Provides metrics: Lines of Code (LOC), functions, classes, loops, conditionals
- Code quality insights and suggestions
✅ Visualization
- Generates flowcharts of code execution paths
- Supports nested structures, loops, and conditions
- Interactive diagrams rendered via D3.js
✅ AI Suggestions (LLM Integration)
- Powered by TinyLLaMA (via
llama-cpp-python) - Provides concise, context-aware improvements
- Language-agnostic suggestions for best practices
✅ Plagiarism Detection
- Levenshtein Distance → Measures edit distance between two code files
- Token-based Jaccard Similarity → Compares structural/code token overlap
✅ Batch Processing
- Upload a GitHub repo link
- Analyze multiple files in one go
- Consolidated results and reports
✅ Report Generation
- Exportable reports (PDF)
✅ Frontend Features
- Built with Angular 20 + TailwindCSS
- Integrated Monaco Editor (VS Code editor) with syntax highlighting
- Tabbed interface for Analysis, Visualization, Suggestions, and Plagiarism
- D3.js integration for dynamic diagrams
✅ Backend Features
- Spring Boot (Java) → API gateway, plagiarism (Java), static analysis (Java),AI integration
- Flask Microservice (Python) → Multi-language static analysis & plagiarism detection
- Microservice-based architecture for modularity
| Layer | Stack |
|---|---|
| Frontend | Angular 20, TailwindCSS, Monaco Editor,D3.js |
| Backend | Spring Boot (Java), Flask (Python) |
| AI Engine | TinyLLaMA |
| Plagiarism | Levenshtein Distance, Token Jaccard |
| Analysis | JavaParser, Python AST, PyJsParser, Clang AST |
🎓 MCA Minor Project 👨💻 Murali Krishna (CHN24MCA-2039) 📌 For Academic & Demo Use Only