Minimal RAG with Azure AI Search and Azure OpenAI 🔎🤖🗄️

Index FAQ data from Azure SQL into Azure AI Search with vector embeddings.

• 📤 Push: Generate embeddings and upload docs.
• 📥 Pull: Use Indexer with Azure OpenAI skills.

Prerequisites 📋

• 🔎 Azure AI Search (vector enabled), 🧠 Azure OpenAI (embeddings), 🗄️Azure SQL, 🐍 Python 3.12+

Setup ⚙️

1. 📝 Copy .env.example → .env
2. 📦 Install dependencies: poetry install or pip install
3. 📄 Create a SQL database, then run faq_tbl.sql to create the CopilotFAQ table with sample data.

Ingest 🚚

• 📤 Build Azure AI Search Index: Use aisearch_faq_index.py to create the index.
• 📤 Add documents to the index
  1. Push: Run push_sql_indexer.py to upload documents.
  2. Pull: Run pull_sql_indexer.py to upload documents.

Environment variables (see .env.example) 🔑

• AZURE_SQL_CONNECTION_STRING (for indexer pull)
• AZURE_SQL_CONNECTION_STRING_ODBC (for push pyodbc client, includes Driver={ODBC Driver 18 for SQL Server})