Legacy Migration Engine

What it does

• Parse C into AST (pycparser)
• Store AST in Neo4j
• Vectorize AST nodes (Ollama embeddings)
• Provide AST context to LLaMA3.2 (Ollama) to convert code to OOP Python
• Streamlit UI for upload/paste, visualize AST, download generated code

Prerequisites

• Python 3.10+
• Neo4j running locally or remotely (bolt)
• Ollama installed locally (https://ollama.com)
• Models pulled to Ollama:
  • ollama pull nomic-embed-text
  • ollama pull llama3.2
• Groq API (If Ollama is not installed)

🛠️ Setup & Usage

1. Clone Repository

git clone https://github.com/your-username/your-repo.git
cd your-repo

2. Build & Start Containers

docker-compose up --build

This will:

• Start Neo4j at http://localhost:7474
• Start Streamlit app at http://localhost:8501

3. Stop Containers

docker-compose down

To also remove volumes (⚠️ deletes all Neo4j data):

docker-compose down -v

📊 Volumes

1. neo4j_data → Stores Neo4j database files (persistent)
2. ./outputs → Stores outputs generated by Streamlit app

🐳 Useful Docker Commands

1. Check running containers:

docker ps

2. Inspect logs for Streamlit app:

docker logs <app-container-id>

3. Open interactive shell inside app container:

docker exec -it <app-container-id> bash

🌐 Access

• Neo4j Browser: http://localhost:7474 (Login with neo4j / strongpass123)

• Streamlit App: http://localhost:8501

✅ Notes

1. Replace GROQ_API_KEY with your valid key (in Dockerfile, docker-compose.yml, etc).