feat: Create LocalRAG project with local models

[Awouzz]

This commit introduces a new project, LocalRAG, which implements a Retrieval Augmented Generation pipeline using entirely local models for embeddings and language generation.

Key features:

• Project structure based on common Python application layouts.
• Local sentence transformer embeddings (e.g., all-MiniLM-L6-v2) via sentence-transformers.
• Local Large Language Model (e.g., distilgpt2) via Hugging Face transformers.
• Configurable model choices through src/config.py.
• Basic RAG pipeline in src/core.py and src/main.py.
• Data loading and processing for text files.
• Placeholder for vector database functionality.
• Comprehensive test suite (tests/) for all components.
• Detailed README.md with setup and usage instructions.

Note: Due to environmental constraints (insufficient disk space for PyTorch installation), I could not successfully run the automated tests during development. The code and test suite are otherwise complete.

[Awouzz]

。