agent.py

import asyncio
from playwright.async_api import async_playwright
import json
from vision_engine import get_elements_from_screenshot
from browser_engine import take_screenshot, click_element

async def main():
    async with async_playwright() as p:
        browser = await p.chromium.launch(headless=True)
        page = await browser.new_page()
        await page.goto("https://www.google.com") # Replace with ur target
        
        # 1. Take a look
        img_path = await take_screenshot(page)
        
        # 2. Ask Vision model where things are
        json_response = get_elements_from_screenshot(img_path)
        elements = json.loads(json_response)
        
        # 3. Decision Logic (Hardcoded for testing)
        target = next((e for e in elements if e["name"] == "search_magnifying_glass"), None)
        if target:
            x = (target["box"][0] + target["box"][2]) / 2
            y = (target["box"][1] + target["box"][3]) / 2
            await click_element(page, x, y)
        
        await browser.close()

asyncio.run(main())