api/validate.py

# api/validate.py

import re
import time
import asyncio
import httpx
from typing import Optional  # Added import for Optional

base_url = "https://www.blackbox.ai"
headers = {
    "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/91.0.4472.124 Safari/537.36"
}

# Cache variables
cached_hid = None
cache_time = 0
CACHE_DURATION = 36000  # Cache duration in seconds (10 hours)

async def getHid(force_refresh: bool = False) -> Optional[str]:
    global cached_hid, cache_time
    current_time = time.time()

    # Check if a forced refresh is needed or if the cached values are still valid.
    if not force_refresh and cached_hid and (current_time - cache_time) < CACHE_DURATION:
        print("Using cached_hid:", cached_hid)
        return cached_hid

    try:
        async with httpx.AsyncClient() as client:
            # Retrieve the initial HTML content.
            response = await client.get(base_url, headers=headers)
            response.raise_for_status()
            content = response.text

            # Use a regular expression to find specific `static/chunks` paths.
            pattern = r"static/chunks/app/layout-[a-zA-Z0-9]+\.js"
            match = re.search(pattern, content)

            if match:
                # Construct the full URL of the JS file.
                js_path = match.group()
                full_url = f"{base_url}/_next/{js_path}"

                # Request the content of the JS file.
                js_response = await client.get(full_url, headers=headers)
                js_response.raise_for_status()

                # In JS content, use a regular expression to search for h-value.
                h_pattern = r'h="([0-9a-f-]+)"'
                h_match = re.search(h_pattern, js_response.text)

                if h_match:
                    h_value = h_match.group(1)
                    print("Found the h-value:", h_value)
                    # Update the cache.
                    cached_hid = h_value
                    cache_time = current_time
                    return h_value
                else:
                    print("The h-value was not found in the JS content.")
                    return None
            else:
                print("The specified JS file path was not found in the HTML content.")
                return None
    except httpx.RequestError as e:
        print(f"An error occurred during the request: {e}")
        return None
    except httpx.HTTPStatusError as e:
        print(f"HTTP error occurred: {e}")
        return None