Building Intelligent Agents with LangChain and OpenAI: A Developer&#s Guide

The rise of artificial intelligence empowers developers to integrate intelligent capabilities into daily workflows. A key approach involves creating autonomous agents that blend reasoning with action. This article demonstrates building such agents using LangChain, OpenAI's GPT-4, and LangChain's experimental features. These agents will execute Python code, interact with CSV files, and tackle complex queries. Let's begin!

---

Why Choose LangChain?

LangChain excels as a framework for developing applications leveraging language models. Its strength lies in creating modular, reusable components—like agents—capable of:

• Executing Python code.
• Analyzing and interacting with data files.
• Performing reasoning and decision-making using tools.

Combining LangChain with OpenAI's GPT-4 enables the creation of agents tailored to specific needs, including data analysis and code debugging.

---

Getting Started: Environment Setup

Before coding, ensure your environment is properly configured:

• Install Python Libraries:

<code class="language-bash">pip install langchain langchain-openai python-dotenv</code>

• Create a .env file: Store your OpenAI API key securely:

<code>OPENAI_API_KEY=your_api_key_here</code>

---

Building a Python Execution Agent

A crucial agent capability is executing Python code. This is achieved using LangChain's PythonREPLTool. Let's define the agent:

Instruction Design

The agent's operation relies on a set of instructions. Here's the prompt:

<code>instruction = """
You are an agent tasked with writing and executing Python code to answer questions.
You have access to a Python REPL for code execution.
Debug your code if errors occur and retry.
Use only the code's output to answer.
If code cannot answer the question, respond with 'I don't know'.
"""</code>

Agent Setup

LangChain's REACT framework will build this agent:

<code class="language-python">from langchain import hub
from langchain_openai import ChatOpenAI
from langchain_experimental.tools import PythonREPLTool
from langchain.agents import create_react_agent, AgentExecutor

base_prompt = hub.pull("langchain-ai/react-agent-template")
prompt = base_prompt.partial(instructions=instruction)

tools = [PythonREPLTool()]
python_agent = create_react_agent(
    prompt=prompt,
    llm=ChatOpenAI(temperature=0, model="gpt-4-turbo"),
    tools=tools,
)
python_executor = AgentExecutor(agent=python_agent, tools=tools, verbose=True)</code>

This agent executes Python code and returns the results.

---

Adding CSV Analysis to the Agent

Data analysis is a frequent AI agent task. Integrating LangChain's create_csv_agent allows our agent to query and process data from CSV files.

CSV Agent Setup

Here's how to add CSV capabilities:

<code class="language-python">from langchain_experimental.agents.agent_toolkits import create_csv_agent

csv_agent = create_csv_agent(
    llm=ChatOpenAI(temperature=0, model="gpt-4-turbo"),
    path="episode-info.csv",
    verbose=True,
    allow_dangerous_code=True,
)</code>

This agent answers questions about episode-info.csv, such as row/column counts and the season with the most episodes.

---

Combining Tools for a Unified Agent

For versatility, we combine Python execution and CSV analysis into a single agent, allowing seamless tool switching based on the task.

Unified Agent Definition

<code class="language-python">from langchain.agents import Tool

def python_executor_wrapper(prompt: str):
    python_executor.invoke({"input": prompt})

tools = [
    Tool(
        name="Python Agent",
        func=python_executor_wrapper,
        description="""
        Transforms natural language to Python code and executes it.
        Does not accept code as input.
        """
    ),
    Tool(
        name="CSV Agent",
        func=csv_agent.invoke,
        description="""
        Answers questions about episode-info.csv using pandas calculations.
        """
    ),
]

grant_agent = create_react_agent(
    prompt=base_prompt.partial(instructions=""),
    llm=ChatOpenAI(temperature=0, model="gpt-4-turbo"),
    tools=tools,
)
grant_agent_executor = AgentExecutor(agent=grant_agent, tools=tools, verbose=True)</code>

This agent handles both Python logic and CSV data analysis.

---

Practical Example: TV Show Episode Analysis

Let's test the unified agent with episode-info.csv:

<code class="language-bash">pip install langchain langchain-openai python-dotenv</code>

The agent analyzes the CSV and returns the season with the most episodes, utilizing pandas.

---

Next Steps and Conclusion

• Experiment with more tools and datasets.
• Explore LangChain's documentation for more advanced agent creation.

LangChain enables the creation of highly customized intelligent agents, simplifying complex workflows. With tools like the Python REPL and CSV agent, the possibilities are vast—from automating data analysis to code debugging and beyond. Start building your intelligent agent today!

The above is the detailed content of Building Intelligent Agents with LangChain and OpenAI: A Developer&#s Guide. For more information, please follow other related articles on the PHP Chinese website!