main_page_downloader_langgraph_tool_react_agent_oop.py

import os
from typing import Type

from langchain_core.messages import HumanMessage
from langchain_core.output_parsers import StrOutputParser
from langchain_core.prompts import PromptTemplate
from langchain_core.runnables import RunnablePassthrough
from langchain_core.tools import BaseTool
# from langchain.pydantic_v1 import BaseModel, Field
from pydantic import BaseModel, Field
import requests
from bs4 import BeautifulSoup
from langchain_ollama.chat_models import ChatOllama
from langgraph.prebuilt import create_react_agent
from langchain_openai import ChatOpenAI
from typing import TypedDict, Annotated
import operator
from urllib3.exceptions import InsecureRequestWarning
from urllib3 import disable_warnings

# =========================
# Global LLM (shared)
# =========================
llm = ChatOllama(model="llama3.2", temperature=0.3)

disable_warnings(InsecureRequestWarning)

# --- Tool Definition ---
class WebScraperInput(BaseModel):
    """Input for the WebScraperTool."""
    url: str = Field(..., description="The URL of the webpage to scrape.")


class WebScraperTool(BaseTool):
    """A tool to scrape text content from a webpage."""
    name: str = "web_scraper"
    description: str = "Scrapes the text content from a given URL."
    args_schema: Type[BaseModel] = WebScraperInput

    def _run(self, url: str) -> str:
        """
        Scrapes the text content from the given URL.
        """
        print("========================")
        print("Fetching web content...", url)
        print("========================")
        try:
            response = requests.get(url, verify=False, timeout=120)
            response.raise_for_status()

            # Use BeautifulSoup to parse the HTML and extract text
            soup = BeautifulSoup(response.text, 'html.parser')
            text = soup.get_text(separator=' ', strip=True)

            print("Content fetched and parsed successfully.")
            return text
        except requests.exceptions.RequestException as e:
            return f"Error fetching content: {e}"
        except Exception as e:
            return f"Error parsing HTML: {e}"


class SummarizeContentInput(BaseModel):
    """Input for the SummarizeContentTool."""
    content: str = Field(..., description="The text content to be summarized.")


class SummarizeContent(BaseTool):
    """Tool to Read any text and provide a simple English summary of its contents using LLM."""
    name: str = "summarize_text"
    description: str = "Read any text and provide a simple English summary of its contents"
    args_schema: Type[BaseModel] = SummarizeContentInput

    def _run(content: str) -> str:
        """Read and summarize any content using LLM and return simple text in not more than 20 to 30 lines."""
        print("========================")
        print("Summarizing content...")
        print("========================")

        summary_prompt = PromptTemplate.from_template(
            "Please provide a concise summary of the following text in not more than 10 lines as plain continuous text only:\n\n{text}"
        )

        summarization_chain = (
            {"text": RunnablePassthrough()} | summary_prompt | llm | StrOutputParser()
        )
        summary = summarization_chain.invoke(content)
        return summary


# --- Agent State ---
class AgentState(TypedDict):
    """
    Represents the state of our agent.
    """
    messages: Annotated[list, operator.add]


# --- Main Application Logic ---
def main():
    """
    Main function to run the LangGraph agent.
    """
    # Initialize the LLM
    # Make sure you have Ollama running with a model like 'llama3'
    # You can pull a model with: ollama pull llama3
    # llm = ChatOllama(model="llama3.2")

    # Initialize the tool
    tools = [WebScraperTool(), SummarizeContent()]

    # Create the ReAct agent
    agent_executor = create_react_agent(llm, tools)

    # --- Run the agent ---
    print("LangGraph Web Scraper Agent")
    print("---------------------------")

    # Example: Scrape a blog post
    url_to_scrape = "https://www.usbank.com/credit-cards/cash-plus-visa-signature-credit-card.html"

    task = f"""
        Please help me with the following workflow:
        1. Fetch content from this URL: {url_to_scrape} using WebScraperTool
        2. Summarize the content using SummarizeContent
        3. Finally present the summary as your final answer

        Execute above steps in order.
        """

    print(f"\nAttempting to scrape: {url_to_scrape}\n")

    # Use invoke to get the final result directly
    final_state = agent_executor.invoke({"messages": [HumanMessage(content=task)]})

    # The result is the final state of the graph
    final_message = final_state['messages'][-1]

    print("--- Final AI Response ---")
    print(final_message.content)

    # If you want to see the tool output, you can inspect the messages list
    # The tool output will likely be the second-to-last message
    tool_output_message = final_state['messages'][-2]
    if tool_output_message.type == 'tool':
        print("\n--- Scraped Content (from final state) ---")
        print(tool_output_message.content[:2000] + "...")


if __name__ == "__main__":
    # To run this, you need to have the following packages installed:
    # pip install langchain langgraph langchain_ollama beautifulsoup4 requests
    main()