app.py

import os
import tempfile
import socket
import traceback
from urllib.parse import urlparse
from playwright.sync_api import sync_playwright
import gradio as gr

BROWSER_ARGS = [
    "--no-sandbox",
    "--disable-setuid-sandbox",
    "--disable-dev-shm-usage",
    "--disable-gpu",
    "--single-process",
    "--no-zygote",
]

def open_and_screenshot(url="https://example.com"):
    diag = {}
    screenshot_path = None
    title = ""

    # quick env/proxy info
    diag['HTTP_PROXY'] = os.environ.get('HTTP_PROXY') or os.environ.get('http_proxy')
    diag['HTTPS_PROXY'] = os.environ.get('HTTPS_PROXY') or os.environ.get('https_proxy')
    diag['NO_PROXY'] = os.environ.get('NO_PROXY') or os.environ.get('no_proxy')

    # parse host for DNS test
    try:
        parsed = urlparse(url)
        host = parsed.netloc or parsed.path
        # strip possible port
        host = host.split(':')[0]
        diag['url_host'] = host
    except Exception as e:
        host = None
        diag['url_host_parse_error'] = str(e)

    # DNS resolution
    if host:
        try:
            resolved_ip = socket.gethostbyname(host)
            diag['resolved_ip'] = resolved_ip
        except Exception as e:
            diag['resolved_ip_error'] = repr(e)
            resolved_ip = None
    else:
        resolved_ip = None

    # optional pre-check with requests (safe, wrapped)
    try:
        import requests
        try:
            resp = requests.get(url, timeout=6)
            diag['requests_status_code'] = resp.status_code
        except Exception as e:
            diag['requests_error'] = repr(e)
    except Exception:
        diag['requests_not_installed'] = True

    # Playwright diagnostics + navigation attempts
    try:
        with sync_playwright() as pw:
            # executable path (property, not callable)
            exec_path = getattr(pw.chromium, "executable_path", None)
            diag['playwright_chromium_executable'] = exec_path

            browser = pw.chromium.launch(headless=True, args=BROWSER_ARGS)
            ctx = browser.new_context()
            page = ctx.new_page()

            # increase navigation timeout to 60s (60000 ms)
            page.set_default_navigation_timeout(60000)

            try:
                page.goto(url, wait_until="domcontentloaded", timeout=60000)
                title = page.title()
                # take screenshot
                tmp = tempfile.NamedTemporaryFile(suffix=".png", delete=False)
                screenshot_path = tmp.name
                page.screenshot(path=screenshot_path, full_page=True)
                diag['playwright_result'] = "success"
                diag['navigator_user_agent'] = page.evaluate("navigator.userAgent")
            except Exception as e_goto:
                diag['playwright_goto_trace'] = traceback.format_exc()

                # fallback 1: try accessing by resolved IP (bypass DNS), if we have one
                if resolved_ip:
                    try:
                        # try plain http to the resolved IP (some hosts block https by IP)
                        alt_url = f"http://{resolved_ip}/"
                        page.goto(alt_url, wait_until="domcontentloaded", timeout=15000)
                        title = page.title()
                        tmp = tempfile.NamedTemporaryFile(suffix=".png", delete=False)
                        screenshot_path = tmp.name
                        page.screenshot(path=screenshot_path, full_page=True)
                        diag['playwright_fallback_ip'] = f"success via {alt_url}"
                    except Exception:
                        diag['playwright_fallback_ip_trace'] = traceback.format_exc()

                # fallback 2: try switching https <-> http on original hostname
                try:
                    if url.startswith("https://"):
                        alt_url2 = url.replace("https://", "http://", 1)
                    else:
                        alt_url2 = url.replace("http://", "https://", 1)
                    page.goto(alt_url2, wait_until="domcontentloaded", timeout=15000)
                    title = page.title()
                    tmp = tempfile.NamedTemporaryFile(suffix=".png", delete=False)
                    screenshot_path = tmp.name
                    page.screenshot(path=screenshot_path, full_page=True)
                    diag['playwright_fallback_proto'] = f"success via {alt_url2}"
                except Exception:
                    diag['playwright_fallback_proto_trace'] = traceback.format_exc()

            try:
                browser.close()
            except Exception:
                pass

    except Exception:
        diag['playwright_launch_trace'] = traceback.format_exc()

    # flatten diag for return
    diag_lines = []
    for k, v in diag.items():
        diag_lines.append(f"{k}: {v}")
    diag_text = "\n".join(diag_lines)

    return title, screenshot_path, diag_text

# Example Gradio UI (replace your existing Blocks UI's call)
with gr.Blocks(title="Chromium-launch Space (Playwright)") as demo:
    url_input = gr.Textbox(value="https://example.com", label="URL")
    run_btn = gr.Button("Open URL in Chromium")
    output_title = gr.Textbox(label="Page title")
    output_img = gr.Image(label="Screenshot")
    output_diag = gr.Textbox(label="Diagnostics", lines=12)

    run_btn.click(fn=open_and_screenshot, inputs=url_input, outputs=[output_title, output_img, output_diag])

if __name__ == "__main__":
    demo.launch(server_name="0.0.0.0", server_port=int(os.environ.get("PORT", 7860)))