Initial Screen:

ProustGPT is a conversational web application inspired by Marcel Proust and powered by a LLM. It currently supports modes:

1. Explore Lost Time: Retrieve relevant passages from In Search of Lost Time using Retrieval-Augmented Generation (RAG).
2. Reflect on My Day: Engage in thoughtful and introspective conversation to reflect on daily experiences.

• Enter a query to explore related passages from Proust's novel, In Search of Lost Time.
• Passages are retrieved based on semantic similarity using Sentence Transformers and vector search.

• Engage in introspective conversation with prompts designed to help you reflect on your daily experiences.

• React
• Styled Components
• React Router

• Flask
• llama_cpp (for the LLaMA model)
• Sentence Transformers (for semantic embedding and retrieval)

• Node.js (version 16+ recommended)
• Python 3.x (recommended 3.8+)
• virtualenv (for managing Python virtual environments)
• A compatible LLaMA model installed locally (e.g., Meta-Llama-3-8B-Instruct-Q4_K_M.gguf)

git clone https://github.com/realtobyfu/proust-gpt.git
cd proust-gpt

Navigate to the backend directory:

cd backend
python3 -m venv venv
source venv/bin/activate

Once activated, your terminal prompt will include (venv).

Run the following command to install backend dependencies:

pip install -r requirements.txt

Ensure that the LLaMA model is placed at the correct location as specified in server.py:

model_path = "/path/to/your/llama/model/Meta-Llama-3-8B-Instruct-Q4_K_M.gguf"

In your activated environment, run:

flask run

The backend will start running at http://127.0.0.1:5000.

Open a new terminal and navigate to the frontend directory:

cd frontend
npm install
npm start

The frontend will start running at http://localhost:3000.

With both frontend and backend running, open your browser at http://localhost:3000.

Choose between:

• Explore Lost Time: Enter queries to explore Marcel Proust's text.
• Reflect on My Day: Converse with ProustGPT for reflective guidance.

• /api/explore_lost_time (POST)

  • Retrieves passages from In Search of Lost Time relevant to your query.

• /api/reflect (POST)

  • Provides reflective responses to your messages without retrieval from external sources.

├── backend
│   ├── app.py                 # Main Flask application
│   ├── parsed.json            # Parsed text data
│   ├── requirements.txt       # Python dependencies
├── frontend
│   ├── public
│   ├── src
│   │   ├── components
│   │   │   ├── ChatPage.tsx   # Main chat page
│   │   └── App.tsx            # React app entry point
│   ├── assets
│   ├── package.json           # Frontend dependencies
│   └── tsconfig.json          # TypeScript configuration
└── README.md                  # Project documentation

• Complete the 'Combray' gallery page
• Add conversation history support
• Add a French-language interface

• Improve semantic retrieval for better context in Explore Lost Time.
• Fine-tune the reflective prompts for a deeper and richer conversational experience.
• Expand corpus with additional relevant documents.

This project is licensed under the Apache 2.0 License.

• Meta (LLaMA): For providing the large language model.
• Sentence Transformers: For the powerful embeddings.
• The open-source community for helpful libraries and inspiration.

Enjoy your thoughtful conversations with ProustGPT! 🌸📖✨