Building Production-Ready AI Agents with Open-Source Tools

from langchain.agents import initialize_agent, AgentType
from langchain.llms import OpenAI
from langchain.tools import DuckDuckGoSearchRun, Calculator

# Initialize tools
search = DuckDuckGoSearchRun()
calculator = Calculator()
tools = [search, calculator]

# Initialize LLM
llm = OpenAI(temperature=0)

# Create agent
agent = initialize_agent(
    tools=tools,
    llm=llm,
    agent=AgentType.ZERO_SHOT_REACT_DESCRIPTION,
    verbose=True,
    max_iterations=3
)

# Execute task
result = agent.run("What's the population of Tokyo and calculate 10% of that number?")

from langgraph.graph import StateGraph, END
from typing import TypedDict, List

class AgentState(TypedDict):
    query: str
    research_results: List[str]
    analysis: str
    confidence: float

def research_node(state: AgentState) -> AgentState:
    # Perform web search
    results = search_tool.run(state["query"])
    return {**state, "research_results": results}

def analyze_node(state: AgentState) -> AgentState:
    # Analyze research results
    analysis = llm.predict(f"Analyze: {state['research_results']}")
    confidence = calculate_confidence(analysis)
    return {**state, "analysis": analysis, "confidence": confidence}

def should_continue(state: AgentState) -> str:
    return "end" if state["confidence"] > 0.8 else "research"

# Build graph
workflow = StateGraph(AgentState)
workflow.add_node("research", research_node)
workflow.add_node("analyze", analyze_node)
workflow.add_edge("research", "analyze")
workflow.add_conditional_edges("analyze", should_continue, {
    "research": "research",
    "end": END
})
workflow.set_entry_point("research")

app = workflow.compile()