How priya-dwivedi/chat-with-code works
Key Components to Include:
Frontend Layer
• Streamlit Web Interface
• User Input Components (OpenAI API key, GitHub repository input)
• Chat Interface
Core Processing Layer
• Repository Handler (for cloning and managing GitHub repositories)
• Text Processing Engine (for chunking code)
• Embedding System (for converting code chunks to embeddings)
• QA Retrieval System (Langchain)
External Services
• OpenAI API (for LLM interactions)
• GitHub (for repository access)
• Deeplake (for storing embeddings)
Component Relationships
• User → Streamlit Interface
• Streamlit Interface → Repository Handler
• Repository Handler → GitHub
• Text Processing Engine → Embedding System
• Embedding System → Deeplake
• QA Retrieval System → OpenAI API
• QA Retrieval System → Chat Interface
Specific Implementation Details
• Environment variables configuration (.env)
• Main application file (chatbot.py)
• Utility functions (utils.py)
Asynchronous nature of the chat interface
Level of Detail:
• Keep high-level architecture visible
• Main processing steps
• External dependencies
• Maintain clarity on data flow
• Configuration components