Yosef-Ali/Amharic-Catholic-AI

Amharic Catholic AI # Amharic Catholic AI 🇪🇹 **User-Friendly Dataset Collector for Ethiopian Catholic Content** A comprehensive, interactive AI system for processing Amharic Catholic texts, generating Q&A pairs, and creating training datasets for language models focused on Ethiopian Catholic content. Features a **user-friendly interface** that guides you through the entire process step-by-step! ## 📁 Project Structure ``` amharic-catholic-ai/ ├── 🚀 launch.py # Simple launcher interface ├── 🎯 collect_dataset.py # Main interactive collector ├── 📋 CLAUDE.md # Detailed documentation ├── ⚙️ config.py # Configuration & API keys ├── 📦 requirements.txt # Python dependencies ├── data/ │ ├── raw/ # 📄 Source documents (PDFs, images) │ ├── processed/ # 📝 Extracted text files │ ├── final/ # 🎯 Training-ready Q&A datasets │ └── validation_results/ # 📊 Quality reports & analysis ├── scripts/ │ ├── enhanced_scraper.py # 🌐 Advanced web scraping │ ├── enhanced_ocr.py # 🔍 DeepSeek OCR processor │ ├── data_validator.py # ✅ Quality validation system │ ├── dashboard.py # 📊 Progress monitoring │ ├── qa_generator.py # ❓ Q&A pair generation │ ├── pdf_processor.py # 📄 PDF text extraction │ ├── run_pipeline.py # 🔄 Automated pipeline │ └── utils.py # 🔧 Helper utilities ├── fine_tuning/ │ └── enhanced_fine_tuning.py # 🧠 Model training script └── logs/ # 📋 Processing logs & sessions ``` ### 🎯 **Main Entry Points** | File | Purpose | Best For | |------|---------|----------| | `launch.py` | Simple launcher menu | First-time users | | `collect_dataset.py` | Full interactive interface | Regular use | | `scripts/enhanced_*.py` | Individual components | Advanced users | | `scripts/run_pipeline.py` | Automated processing | Batch jobs …