by-agent-for-agent/scripts/twitter_browser_dl.py

# /// script
# requires-python = ">=3.12"
# dependencies = [
#     "playwright",
# ]
# ///
import asyncio
import os
import sys
from playwright.async_api import async_playwright

async def download_twitter_photo(url, output_dir="downloads"):
    if not os.path.exists(output_dir):
        os.makedirs(output_dir)

    async with async_playwright() as p:
        print("💡 正在启动浏览器补丁并安装内核 (如果是第一次运行可能会慢一点)...")
        # 确保安装浏览器内核
        os.system("uv run playwright install chromium")
        
        browser = await p.chromium.launch(headless=True)
        context = await browser.new_context(
            viewport={'width': 1920, 'height': 1080},
            user_agent="Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36"
        )
        page = await context.new_page()

        print(f"🚀 正在访问推文地址: {url}")
        try:
            # 访问页面
            await page.goto(url, wait_until="networkidle", timeout=60000)
            
            # 尝试定位图片
            # Twitter 的图片通常在 article 里的 img 标签
            print("📸 正在寻找图片元素...")
            images = await page.query_selector_all("img[src*='format=jpg'], img[src*='format=png']")
            
            if not images:
                print("❌ 未能找到有效的图片元素，推文可能涉及权限或加载失败。")
                await browser.close()
                return

            # 获取当前 URL 中指定的 photo/X 对应的图片，或者全部下载
            for i, img in enumerate(images):
                src = await img.get_attribute("src")
                if "profile_images" in src or "emoji" in src:
                    continue
                
                # 转换成原图链接
                if "name=" in src:
                    large_url = src.split("name=")[0] + "name=large"
                else:
                    large_url = src + "&name=large"

                filename = f"twitter_image_{i}.jpg"
                save_path = os.path.join(output_dir, filename)
                
                print(f"📥 正在下载原图: {large_url}")
                img_response = await page.request.get(large_url)
                if img_response.status == 200:
                    with open(save_path, "wb") as f:
                        f.write(await img_response.body())
                    print(f"✅ 已保存至: {save_path}")
                
        except Exception as e:
            print(f"💥 运行出错: {e}")
        finally:
            await browser.close()

if __name__ == "__main__":
    if len(sys.argv) < 2:
        print("用法: uv run scripts/twitter_browser_dl.py <url>")
    else:
        asyncio.run(download_twitter_photo(sys.argv[1]))