swsthik/Auralis_Schema

Customer Support Copilot

A Streamlit-based app for customer support ticket classification and retrieval-augmented generation (RAG).

Structure

• —app.py: Main Streamlit app
• —src/: Source code modules
• —data/: Data and knowledge base
• —tests/: Unit tests
• —docs/: Documentation

Configuration

You can control the RAG agent's retrieval and escalation logic dynamically using environment variables in your .env file:

• —RAG_MAX_DOCS: Number of top documents to retrieve for each query (default: 5)
• —ESCALATION_THRESHOLD: Escalation score threshold for routing to a human agent (default: 0.6)

Example .env entries:

RAG_MAX_DOCS=5
ESCALATION_THRESHOLD=0.6

You can also override these values at runtime by passing max_docs and escalation_threshold as arguments to the RAGAgent or its process_query method.

Setup

1. 1.Install dependencies: pip install -r requirements.txt
2. 2.Run the app: streamlit run app.py