Skip to content

Setup & Prerequisites

Before starting the hands-on exercises, please ensure you have the following setup completed.

1. Clone the Repository

All code examples are sourced from the agentic-playground repository. Clone it to your local machine:

git clone https://github.com/denniszielke/agentic-playground.git
cd agentic-playground

2. Python Environment

We recommend using a virtual environment to manage dependencies.

  • Ensure Python is installed: Python 3.9 or higher is recommended.
  • Create a virtual environment: 
    # Windows
python -m venv .venv
.venv\Scripts\activate

# macOS/Linux
python3 -m venv .venv
source .venv/bin/activate
  • Install Dependencies: The repository includes a requirements.txt file listing the necessary Python packages. 
    pip install -r requirements.txt
    Note: Some examples require additional dependencies (e.g., pytz, graphviz, requests, pyaudio, sounddevice, pydub, pyperclip, specific langchain, llama-index, semantic-kernel, or autogen components). Install these as needed when running specific examples, following instructions in the module content or error messages. For Graphviz visualization (knowledge-graphs.py), you also need to install the Graphviz binaries separately: * macOS: brew install graphviz * Ubuntu/Debian: sudo apt-get update && sudo apt-get install -y graphviz * Windows: Download from the official Graphviz website and add to PATH.

3. GitHub Personal Access Token (PAT)

Many examples in this repository interact with LLMs hosted via GitHub Models inference endpoints. Accessing these requires a GitHub Personal Access Token (PAT).

  • Generate a PAT:
    1. Go to your GitHub Settings > Developer settings > Personal access tokens > Tokens (classic).
    2. Click "Generate new token" (or "Generate new token (classic)").
    3. Give your token a descriptive name (e.g., "Agentic Playground Workshop").
    4. Set an expiration date.
    5. Crucially, for GitHub Models inference, a token with no specific scopes/permissions is sufficient.
    6. Click "Generate token".
    7. Copy the generated token immediately. You won't be able to see it again.
  • Configure the PAT: The scripts typically use the python-dotenv library to load environment variables from a .env file in the repository's root directory.
    1. Create a file named .env in the root of the cloned agentic-playground directory.
    2. Add the following line, replacing your_github_pat_here with the token you just copied: 
      GITHUB_TOKEN="your_github_pat_here"

4. Key Libraries Overview (High Level)

The repository utilizes several popular libraries for building AI agents:

  • OpenAI Python Library (openai): Used for interacting with OpenAI-compatible APIs, including the GitHub Models endpoint. Provides methods for chat completions, streaming, tool calling, and multimodal inputs.
  • LangChain (langchain, langchain-openai, etc.): A framework for developing applications powered by language models. Provides tools for managing prompts, chains, agents (like ReAct), memory, and document loading.
  • LlamaIndex (llama-index, llama-index-llms-openai, etc.): A data framework for LLM applications, focusing on connecting LLMs with external data. Also provides agent implementations (like ReAct).
  • Semantic Kernel (semantic-kernel): A Microsoft-developed SDK for integrating LLMs into applications. Offers features like plugins (tools), planners, memory, and agent abstractions (including process frameworks and agent chats).
  • AutoGen (autogen-agentchat, autogen-core, autogen-ext): A framework for enabling multi-agent conversations and workflows, often using an orchestrator model to manage dynamic interactions.
  • Pydantic: Used for data validation and defining structured data models (like the KnowledgeGraph or state objects).

We will explore specific features of these libraries as we encounter them in the code examples throughout the workshop.