A modern web application that analyzes CSV files for data quality issues using LLaMA 8B via Ollama. Features a clean, chat-like interface that displays only errors found in your data.
- 🤖 AI-Powered Analysis: Uses LLaMA 8B model via Ollama for intelligent data quality detection
- 🎨 Clean UI: Modern, responsive design with intuitive user interface
- 💬 Chat Interface: Errors displayed in a clean chat-like format
- 🔍 Smart Filtering: Filter errors by type (Critical, Warning, Info)
- 📊 Health Score: Visual health score indicator (0-100)
- ⚡ Fast Processing: Efficient chunk-based processing for large files
- Python 3.8+ installed
- Node.js 16+ and npm installed
- Ollama installed and running
- LLaMA 8B model installed in Ollama
- Download and install Ollama from https://ollama.ai
- Pull the LLaMA 8B model:
ollama pull llama3:8b
- Verify Ollama is running:
ollama list
-
Install Python dependencies:
pip install -r requirements.txt
-
Install Node.js dependencies:
npm install
Simply double-click run_app.bat or run:
run_app.batTerminal 1 - Backend:
python -m uvicorn main:app --reload --host 0.0.0.0 --port 8000Terminal 2 - Frontend:
npm run devThe application will be available at:
- Frontend: http://localhost:5173
- Backend API: http://localhost:8000
- Open http://localhost:5173 in your browser
- Upload a CSV file by dragging and dropping or clicking to browse
- Wait for the analysis to complete (this may take a few moments)
- Review the results:
- Health Score: Overall data quality score (0-100)
- Error Summary: AI-generated summary of issues found
- Error Chat: Detailed list of all errors in a chat interface
- Filter errors by type using the filter buttons
hackthom-aim-25-main/
├── main.py # FastAPI backend server
├── services.py # CSV analysis service using LLaMA 8B
├── requirements.txt # Python dependencies
├── package.json # Node.js dependencies
├── vite.config.js # Vite configuration
├── run_app.bat # Windows startup script
├── src/
│ ├── main.jsx # React entry point
│ ├── App.jsx # Main app component
│ ├── components/
│ │ ├── FileUpload.jsx # File upload component
│ │ ├── AnalysisSummary.jsx # Results summary component
│ │ └── ErrorChat.jsx # Error display component
│ └── *.css # Component styles
└── index.html # HTML template
- File Upload: User uploads a CSV file through the web interface
- Chunk Processing: The CSV is processed in small chunks (5 rows at a time) to fit LLaMA 8B's context window
- AI Analysis: Each chunk is sent to LLaMA 8B via Ollama API for error detection
- Error Aggregation: All errors are collected and categorized by severity
- Summary Generation: LLaMA 8B generates a user-friendly summary
- Display: Results are shown in a clean, chat-like interface
- CRITICAL: Serious data quality issues that need immediate attention
- WARNING: Potential issues that should be reviewed
- INFO: Informational messages about data patterns
- Make sure Ollama is running:
ollama serve - Verify the model is installed:
ollama list - Check if Ollama is accessible at http://localhost:11434
- Check if port 8000 is available
- Verify Python dependencies are installed:
pip install -r requirements.txt
- Check if port 5173 is available
- Verify Node.js dependencies:
npm install - Clear node_modules and reinstall if needed
- LLaMA 8B processing can be slow for large files
- Consider using a GPU-accelerated Ollama setup for better performance
- Reduce chunk size in
services.pyif memory is limited
The codebase is designed to be:
- Easy to Understand: Clear function names, comments, and structure
- Well-Organized: Separated concerns (backend, frontend, services)
- Maintainable: Modular components and services
See LICENSE file for details.
Feel free to submit issues and enhancement requests!