GHCP RAG PromptFlow Extension

This repository contains a GitHub Copilot (GHCP) extension that interacts with an existing RAG (Retrieval-Augmented Generation) endpoint using Azure ML PromptFlow. The code here demonstrates how to set up a minimal Node.js server, parse incoming requests, and forward them to a PromptFlow endpoint for processing.

---

Table of Contents

• GHCP RAG PromptFlow Extension
  • Table of Contents
  • Overview
  • Features

---

Overview

This project serves as a starting point (skeleton) for creating further GitHub Copilot extensions that call existing RAG endpoints. It demonstrates:

• How to receive and parse requests from GHCP.
• How to call a remote PromptFlow endpoint.
• How to parse the response and structure output events.

---

Features

• Express.js Server: Launches REST endpoints (GET and POST) to handle requests.
• PromptFlow Integration: Makes POST calls to a remote RAG endpoint using fetch.
• Event Streaming: Uses utility methods (createAckEvent, createTextEvent, createDoneEvent) to stream events back to clients.
• GitHub Authentication: Retrieves the GitHub user login via X-GitHub-Token header.

---