🤖 LangChain Labs 101: Your Gateway to Agentic AI Tools

Welcome to the LangChain 101 repository! This project serves as an accessible entry point for beginners eager to explore the world of agentic AI, focusing on the crucial concept of tools. LangChain is a powerful framework for building applications with large language models (LLMs), and this tutorial will guide you through your first steps in creating AI-powered tools. You can run all of this in VSCode, or your favorite IDE if it supports Jupyter notebooks and Python. The notebooks also optionally run on Google Colab (lessons 001 and 002), however the rest are built as Python scripts.

🎯 What You'll Learn

• Understand the basics of agentic AI and the importance of tools
• Create and implement custom AI tools using LangChain
• Explore how these tools can be used in more complex AI systems

🌟 Who This Is For

• Beginners with basic Python knowledge: You should have Python and either conda or venv installed. We'll provide conda instructions to get you started.
• Experienced developers looking to dive into agentic AI development

🔑 What You'll Need

• Python installed on your system
• Conda or venv for creating a new environment (we'll use conda in our instructions)
• An OpenAI API key (We'll use gpt4-mini, which is cost-effective for learning)
• Basic familiarity with command line interfaces

🚀 What to Expect

This tutorial is Part One of a broader series on agentic AI. We'll focus on tools as a fundamental component of agentic systems. You'll learn how to create, implement, and use AI tools, setting the foundation for more advanced agentic AI applications in future tutorials.

Get ready to embark on your journey into the world of agentic AI development!

🎓 Tutorials - Part 1

Basics in Jupyter Notebooks / Google Colab

1. 001-tutorial: Introduction to LangChain basics -

   • Set up your first LangChain environment
   • Learn how to initialize and use a language model
   • Explore basic interactions with the AI

2. 002-tutorial: Working with prompt templates and RunnableSequence -

   • Learn how to load prompts from external files
   • Understand and implement ChatPromptTemplates
   • Explore the concept of RunnableSequence for creating complex workflows

Tool Usage examples

3. 003-tutorial: Creating custom tools and agents

   • Implement a simple custom tool (Mock Weather Checker)
   • Create a React agent using the custom tool
   • Understand how to combine prompts, tools, and agents

4. 004-tutorial: Using DuckDuckGo Search Tool

   • Integrate the DuckDuckGo search tool into your LangChain application
   • Learn how to create an agent that can perform web searches
   • Explore more complex prompt structures and agent interactions

5. 005-tutorial-streamlit: Integrating Streamlit (With Error Handling)

   • Build a user-friendly web interface for your LangChain application using Streamlit
   • Implement error handling for more robust applications
   • Learn how to display intermediate steps and search processes to users

(Part 2 / More tutorials will be added as they are developed)

🚀 Getting Started

Prerequisites

Before you begin, ensure you have the following installed:

• Python 3.10+
• pip (Python package manager)

Installation

1. Clone this repository:

   git clone https://github.com/Agentic-Insights/langchain-labs
   cd langchain-labs
   

2. Create a conda environment (optional but recommended):

   conda create -n langchain-labs python=3.10
   conda activate langchain-labs
   

3. Install the required packages:

   pip install -r requirements.txt
   

🗂️ Repository Structure

• labs/: Contains individual tutorial scripts and related files
• README.md: This file, providing an overview of the repository

🛠️ Usage

For Jupyter notebooks, we recommend using VSCode or the Colab support.

Each tutorial in the labs/ directory is a self-contained script. To run a tutorial:

1. Navigate to the labs/ directory:

   cd labs
   

2. Run the desired tutorial script. For .py file tutorials:

   python 003-tutorial.py
   

   For the 005-tutorial which uses Streamlit:

   streamlit run 005-tutorial-streamlit.py
   

3. Follow the prompts or instructions provided by the tutorial script.

4. For tutorials that require API keys or other sensitive information, make sure to set up your .env file with the necessary credentials.

5. Some tutorials may require additional setup or resources. Always check the comments at the beginning of each script for any specific instructions.

6. To explore the code and understand how it works, open the tutorial scripts in your preferred code editor.

7. Feel free to modify the scripts and experiment with different parameters or inputs to deepen your understanding of LangChain concepts.

Remember to activate your virtual environment before running the tutorials if you're using one.