tutorial-records/workshop

UCAgent Workshop Quick Start Guide

Welcome to the UCAgent Workshop! The code and environment for this workshop demonstration are available at the official repository: https://github.com/XS-MLVP/tutorial-records.

This guide will walk you through pulling the pre-built environment, launching the Web UI, and exploring the pre-configured digital circuit verification cases.

📦 Repository Overview

In this repository, you will find:

• Pre-prepared Cases: Classic verification examples along with their historical verification records.
• Docker Scripts: The Dockerfile and start_demo.sh scripts, which our CI uses to automatically build and publish the workshop image.
• Environment Template: A env.template file for optional API configuration.

---

🚀 Quick Start (Viewing the Demo)

You do not need to build the Docker image yourself. The tutorial repository's CI has automatically built and published the image for you.

1. Pull the Docker Image

First, pull the latest workshop image to your local machine:

docker pull ghcr.io/xs-mlvp/workshop:latest

2. Launch the Demo Environment

Start the container using the following command. This will spin up the Master API server and load the pre-configured background agents:

docker run -it --rm -p 8800:8800 -p 8765:8765 -p 8766:8766 ghcr.io/xs-mlvp/workshop:latest

3. Explore the Web UI

Once the terminal indicates that all services have started, open your web browser and navigate to: 👉 http://localhost:8800

In this unified Master Web Interface, you can:

• View Pre-loaded Tasks: On the left-hand task panel, you will see the cases. Both should have a green Online status indicator.
• Review History: Click the API connection button next to any task to jump into its real-time control center. Here, you can review previously generated code, stage verification logs, and Diff files without needing any LLM API keys.

---

🛠️ Advanced: Running Verifications Yourself (Optional)

The quick start above is perfect for exploring the UI and viewing the pre-generated results. However, if you wish to write custom prompts and have UCAgent actively generate new code and run verifications in real-time, you must configure your Large Language Model (LLM) credentials.

1. Configure your credentials
   Edit the provided env.template file and fill in your API credentials (e.g., OpenAI or a compatible interface):

   # OPENAI_API_KEY=your_api_key_here
   # OPENAI_API_BASE=your_api_base_here
   # OPENAI_MODEL=your_model_name_here
   

2. Launch with API Access
   Run the container, attaching the env.template file so the agents can connect to the LLM:

   docker run -it --rm --network host --env-file env.template ghcr.io/xs-mlvp/workshop:latest
   

With your API keys successfully loaded, the Agents are now fully empowered to execute new automated verification workflows based on your commands!