import requests import re import time from api.config import BASE_URL, headers # Cache variables cached_hid = None cache_time = 0 CACHE_DURATION = 36000 # Cache duration in seconds (10 hours) def getHid(force_refresh=False): global cached_hid, cache_time current_time = time.time() # Check if cache is valid and not forced to refresh if not force_refresh and cached_hid and (current_time - cache_time) < CACHE_DURATION: print("using cached_hid:", cached_hid) return cached_hid try: # Fetch initial HTML content response = requests.get(BASE_URL, headers=headers) response.raise_for_status() content = response.text # Use regex to find the specific static/chunks path 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 JS file content js_response = requests.get(full_url, headers=headers) js_response.raise_for_status() # Search for h-value in the JS content using regex 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 h-value:", h_value) # Update cache cached_hid = h_value cache_time = current_time return h_value else: print("h-value not found in JS content") return None else: print("JS file path not found in HTML content") return None except requests.exceptions.RequestException as e: print(f"Error occurred: {e}") return None