microsoft/RAG_Hack

Project: Explore the world with Contoso Travel Agency

Opened this issue · 3 comments

Project Name

ContosoTravelAgency

Description

ContosoTravelAgency Logo
ContosoTravelAgency


Visit ContosoTravelAgency More Info

Architecture

Short description

Our app is an advanced travel planning system built using Azure Durable Functions and Semantic Kernel. The Blazor frontend, hosted on Azure App Service, interacts with a backend that leverages the Retrieval-Augmented Generation (RAG) pattern for efficient, scalable, and reliable travel planning. By incorporating Azure CosmosDB, we enhance our ability to process and respond to natural language queries, making it a powerful reference application for those interested in exploring the potential of Azure Durable Functions and advanced semantic search.

Data sources used

We utilized Azure Cosmos DB for its scalability, low latency, and vector search capabilities. Our data sources include comprehensive travel information, with real-time updates managed through Cosmos DB’s Change Feed. Additionally, we integrate with external services for flight bookings and weather information, and use Redis for real-time notifications and Postmark for sending confirmation emails.

Additional modifications to the data ingestion or prompting flows

We enhanced the data ingestion process by incorporating the Change Feed in Cosmos DB for real-time updates. For prompting flows, we implemented natural language to SQL conversion using Azure Cosmos DB's SQL Copilot REST API. We also optimized query processing with the fan-out and fan-in pattern to balance efficiency and performance.

Who can benefit from your app?

The app is ideal for travel agencies looking to streamline travel planning and booking processes, frequent travelers seeking personalized flight options and efficient booking management, and general users who need accurate retrieval of past bookings and travel deals. Additionally, it serves as a valuable reference for developers and IT professionals interested in leveraging Semantic Kernel and Azure Durable Functions in real-world applications.

Technology & Languages

  • JavaScript
  • Java
  • .NET
  • Python
  • AI Studio
  • AI Search
  • PostgreSQL
  • Cosmos DB
  • Azure SQL

Project Repository URL

https://github.com/Cloud-Jas/DurableMultiAgents

Deployed Endpoint URL

https://travelagency.iamdivakarkumar.com/

Project Video

https://www.youtube.com/watch?v=KzqwhnGl7ws

Team Members

Divakar-kumar [Linkedin]

Amazing work
@Divakar-kumar
We Can use Skyscanner or any third-party API and create SaaS website

Amazing work @Divakar-kumar We Can use Skyscanner or any third-party API and create SaaS website

@Yash-Kavaiya ,Thank you for your suggestion! While using third-party APIs like Skyscanner is certainly a great option, my goal for this project was to create a reference implementation that showcases the power of Semantic Kernel and Azure Durable Functions.

Hello @Divakar-kumar, thank you for participating in RAG Hack!

The team is working hard to distribute badges. Please have each team member fill out this form:
aka.ms/raghack/badge-dist

Thank you!