Backend

Folder: src\App\backend

The backend is a Python Quart app that processes queries, generates insights, and communicates with databases and AI services.

Features

1. Azure OpenAI Integration

   • Handles natural language queries from users.
   • Calls Azure OpenAI for understanding and response generation.

2. Semantic Kernel Plugin:

   • Powers natural language interactions via custom kernel functions

3. Data Access

   • SQL for structured data.
   • Azure Cognitive Search for transcripts.

4. Chat History

   • Cosmos DB for storing user conversations.

5. Chart Processing

   • Converts results to chart-ready JSON that is then used to diplay chart on the frontend.

Semantic Kernel Plugin Breakdown

Located in ChatWithDataPlugin:

greeting()

- Responds to simple greetings or general questions

- Uses either Azure AI Project or direct OpenAI client

get_SQL_Response()

- Converts natural language questions into valid SQL queries

get_answers_from_calltranscripts()

- Performs Retrieval-Augmented Generation (RAG)

- Uses semantic + vector hybrid search with Azure AI Search

- Returns summarized or specific insights from indexed call data

Tools & Libraries

• Quart
• Azure OpenAI
• CosmosDB SDK
• SQLAlchemy
• Semantic Kernel
• Azure AI Search

---