🚀 Vertex AI RAG Agent with ADK

❓ Sample Questions for the F1 RAG Agent

Here are some example prompts you can use to test and demonstrate the Formula 1 RAG Agent's retrieval capabilities:

• "How did Ayrton Senna perform in the 1991 Monaco Grand Prix?"
• "Which constructor dominated the 2004 Formula 1 season and why?"
• "Give me a summary of Lewis Hamilton’s 2018 season, including key races he won."

📘 What I Did in This Project

This project is a fully functional Retrieval Augmented Generation (RAG) agent built using Google Cloud Vertex AI and the Agent Development Kit (ADK).
The agent is powered by a custom Formula 1 corpus containing 1,800+ documents.

Summary of Accomplishments

• Set up a Google Cloud project, fixed SDK conflicts, and authenticated using Application Default Credentials.
• Installed Google Cloud SDK cleanly and configured the correct project, region, and zone.
• Creatd a corpus named Formula1
• Created a GCS bucket (f1-corpus-8474-eu) to store all Formula 1 documents.
• Extracted data from the public F1 dataset (1950-2020): https://www.kaggle.com/datasets/rohanrao/formula-1-world-championship-1950-2024
• Generated a comprehensive f1_manifest.txt file containing 1,825+ GCS file paths.
• Built a bootstrap ingestion script that imports files in safe batches of 200.
• Successfully ingested 1,825+ Formula 1 documents into a RAG corpus named Formula1.
• Integrated RAG tools into a single ADK agent.
• Connected the agent to the ADK Web UI for live querying.
• Verified ingestion using verify_corpus and executed retrieval-based F1 Q&A.
• Cleaned the project structure (removed unused folders) and prepared it for GitHub.

---

📄 Overview

This repository contains a Vertex AI-powered Retrieval Augmented Generation (RAG) Agent implemented using the Google Agent Development Kit (ADK).

---

🧩 Prerequisites

• Google Cloud account with billing enabled
• Google Cloud project with Vertex AI API enabled
• Python 3.9+
• Google Cloud SDK installed
• Permissions for Vertex AI RAG
• ADC configured locally

---

🔐 Setting Up Google Cloud Authentication

1. Install Google Cloud SDK

Follow: https://cloud.google.com/sdk/docs/install

2. Initialize

gcloud init

3. Authenticate (ADC)

gcloud auth application-default login

4. Verify

gcloud auth list
gcloud config list

5. Enable Vertex AI API

gcloud services enable aiplatform.googleapis.com

---

🤖 Using the Agent

🔍 Query Documents — rag_query

• Ask natural-language questions
• Retrieves relevant text chunks
• Generates grounded answers

---

📚 List Corpora — list_corpora

• Lists all RAG corpora in your Google Cloud project

---

🏗️ Create Corpus — create_corpus

• Creates an empty RAG corpus

---

📥 Add Data — add_data

• Supports GCS URLs, Google Drive links, Google Docs URLs

---

📊 Get Corpus Information — get_corpus_info

• Shows file count, metadata, timestamps

---

🗑️ Delete Corpus — delete_corpus

• Permanently deletes a corpus