How 050903/vertex-ai-reasoning-agent-demo works
Project type and purpose
• This is a small demo application that combines a desktop GUI (built with PyQt6) and a cloud-based AI agent (deployed on Google Vertex AI Reasoning Engines).
• The purpose is to show how to deploy a custom “addition agent” to Vertex AI and interact with it via a local PyQt6 interface.
File structure
• Top-level files only, no subdirectories:
• addition_agent.py – Python script to define and deploy the add-two-numbers agent.
• reasoning_ui.py – PyQt6 GUI application that prompts the user for input and queries the deployed agent.
• cli.txt – Sample command to launch the GUI.
• requirements.txt – Lists Python dependencies (PyQt6, google-cloud-aiplatform, etc.).
• GEMINI.md, LICENSE, README.md – documentation and license.
README for additional insights
• Core workflow:
a) Run addition_agent.py → uses google-cloud-aiplatform client library → deploys an agent to Vertex AI.
b) Run reasoning_ui.py → displays a window where the user enters “x y” → sends a query to the deployed Vertex AI agent → displays the result.
• Hard-coded project/region/engine IDs in addition_agent.py demonstrate a single-engine deployment.
Elements
• “PyQt6 GUI (reasoning_ui.py)”
• “Addition Agent Deployer (addition_agent.py)”
• “Vertex AI Reasoning Engine (AdditionAgent)”
• “Google Cloud IAM / Auth”
• “Deploy Agent API Call”
• “Query: add x y”
• “Response: sum”
• Technologies next to each component