vnai 2.1.7__py3-none-any.whl → 2.1.8__py3-none-any.whl

vnai/beam/quota.py

@@ -1,333 +1,486 @@
-import time
-import functools
-import threading
-from collections import defaultdict
-from datetime import datetime
-
-class RateLimitExceeded(Exception):
-    def __init__(self, resource_type, limit_type="min", current_usage=None, limit_value=None, retry_after=None):
-        self.resource_type = resource_type
-        self.limit_type = limit_type
-        self.current_usage = current_usage
-        self.limit_value = limit_value
-        self.retry_after = retry_after
-        message =f"Bạn đã gửi quá nhiều request tới {resource_type}. "
-        if retry_after:
-            message +=f"Vui lòng thử lại sau {round(retry_after)} giây."
-        else:
-            message +="Vui lòng thêm thời gian chờ giữa các lần gửi request."
-        super().__init__(message)
-
-class Guardian:
-    _instance = None
-    _lock = threading.Lock()
-
-    def __new__(cls):
-        with cls._lock:
-            if cls._instance is None:
-                cls._instance = super(Guardian, cls).__new__(cls)
-                cls._instance._initialize()
-            return cls._instance
-
-    def _initialize(self):
-        self.resource_limits = defaultdict(lambda: defaultdict(int))
-        self.usage_counters = defaultdict(lambda: defaultdict(list))
-        self.resource_limits["default"] = {"min": 60,"hour": 3000}
-        self.resource_limits["TCBS"] = {"min": 60,"hour": 3000}
-        self.resource_limits["VCI"] = {"min": 60,"hour": 3000}
-        self.resource_limits["MBK"] = {"min": 600,"hour": 36000}
-        self.resource_limits["MAS.ext"] = {"min": 600,"hour": 36000}
-        self.resource_limits["VCI.ext"] = {"min": 600,"hour": 36000}
-        self.resource_limits["FMK.ext"] = {"min": 600,"hour": 36000}
-        self.resource_limits["VND.ext"] = {"min": 600,"hour": 36000}
-        self.resource_limits["CAF.ext"] = {"min": 600,"hour": 36000}
-        self.resource_limits["SPL.ext"] = {"min": 600,"hour": 36000}
-        self.resource_limits["VDS.ext"] = {"min": 600,"hour": 36000}
-        self.resource_limits["FAD.ext"] = {"min": 600,"hour": 36000}
-
-    def verify(self, operation_id, resource_type="default"):
-        current_time = time.time()
-        limits = self.resource_limits.get(resource_type, self.resource_limits["default"])
-        minute_cutoff = current_time - 60
-        self.usage_counters[resource_type]["min"] = [
-            t for t in self.usage_counters[resource_type]["min"]
-            if t > minute_cutoff
-        ]
-        minute_usage = len(self.usage_counters[resource_type]["min"])
-        minute_exceeded = minute_usage >= limits["min"]
-        if minute_exceeded:
-            from vnai.beam.metrics import collector
-            collector.record(
-"rate_limit",
-                {
-"resource_type": resource_type,
-"limit_type":"min",
-"limit_value": limits["min"],
-"current_usage": minute_usage,
-"is_exceeded": True
-                },
-                priority="high"
-            )
-            raise RateLimitExceeded(
-                resource_type=resource_type,
-                limit_type="min",
-                current_usage=minute_usage,
-                limit_value=limits["min"],
-                retry_after=60 - (current_time % 60)
-            )
-        hour_cutoff = current_time - 3600
-        self.usage_counters[resource_type]["hour"] = [
-            t for t in self.usage_counters[resource_type]["hour"]
-            if t > hour_cutoff
-        ]
-        hour_usage = len(self.usage_counters[resource_type]["hour"])
-        hour_exceeded = hour_usage >= limits["hour"]
-        from vnai.beam.metrics import collector
-        collector.record(
-"rate_limit",
-            {
-"resource_type": resource_type,
-"limit_type":"hour" if hour_exceeded else"min",
-"limit_value": limits["hour"] if hour_exceeded else limits["min"],
-"current_usage": hour_usage if hour_exceeded else minute_usage,
-"is_exceeded": hour_exceeded
-            }
-        )
-        if hour_exceeded:
-            raise RateLimitExceeded(
-                resource_type=resource_type,
-                limit_type="hour",
-                current_usage=hour_usage,
-                limit_value=limits["hour"],
-                retry_after=3600 - (current_time % 3600)
-            )
-        self.usage_counters[resource_type]["min"].append(current_time)
-        self.usage_counters[resource_type]["hour"].append(current_time)
-        return True
-
-    def usage(self, resource_type="default"):
-        current_time = time.time()
-        limits = self.resource_limits.get(resource_type, self.resource_limits["default"])
-        minute_cutoff = current_time - 60
-        hour_cutoff = current_time - 3600
-        self.usage_counters[resource_type]["min"] = [
-            t for t in self.usage_counters[resource_type]["min"]
-            if t > minute_cutoff
-        ]
-        self.usage_counters[resource_type]["hour"] = [
-            t for t in self.usage_counters[resource_type]["hour"]
-            if t > hour_cutoff
-        ]
-        minute_usage = len(self.usage_counters[resource_type]["min"])
-        hour_usage = len(self.usage_counters[resource_type]["hour"])
-        minute_percentage = (minute_usage / limits["min"]) * 100 if limits["min"] > 0 else 0
-        hour_percentage = (hour_usage / limits["hour"]) * 100 if limits["hour"] > 0 else 0
-        return max(minute_percentage, hour_percentage)
-
-    def get_limit_status(self, resource_type="default"):
-        current_time = time.time()
-        limits = self.resource_limits.get(resource_type, self.resource_limits["default"])
-        minute_cutoff = current_time - 60
-        hour_cutoff = current_time - 3600
-        minute_usage = len([t for t in self.usage_counters[resource_type]["min"] if t > minute_cutoff])
-        hour_usage = len([t for t in self.usage_counters[resource_type]["hour"] if t > hour_cutoff])
-        return {
-"resource_type": resource_type,
-"minute_limit": {
-"usage": minute_usage,
-"limit": limits["min"],
-"percentage": (minute_usage / limits["min"]) * 100 if limits["min"] > 0 else 0,
-"remaining": max(0, limits["min"] - minute_usage),
-"reset_in_seconds": 60 - (current_time % 60)
-            },
-"hour_limit": {
-"usage": hour_usage,
-"limit": limits["hour"],
-"percentage": (hour_usage / limits["hour"]) * 100 if limits["hour"] > 0 else 0,
-"remaining": max(0, limits["hour"] - hour_usage),
-"reset_in_seconds": 3600 - (current_time % 3600)
-            }
-        }
-guardian = Guardian()
-
-class CleanErrorContext:
-    _last_message_time = 0
-    _message_cooldown = 5
-
-    def __enter__(self):
-        return self
-
-    def __exit__(self, exc_type, exc_val, exc_tb):
-        if exc_type is RateLimitExceeded:
-            current_time = time.time()
-            if current_time - CleanErrorContext._last_message_time >= CleanErrorContext._message_cooldown:
-                print(f"\n⚠️ {str(exc_val)}\n")
-                CleanErrorContext._last_message_time = current_time
-            import sys
-            sys.exit(f"Rate limit exceeded. {str(exc_val)} Process terminated.")
-            return False
-        return False
-
-def optimize(resource_type='default', loop_threshold=10, time_window=5, ad_cooldown=150, content_trigger_threshold=3,
-            max_retries=2, backoff_factor=2, debug=False):
-    if callable(resource_type):
-        func = resource_type
-        return _create_wrapper(func,'default', loop_threshold, time_window, ad_cooldown, content_trigger_threshold,
-                             max_retries, backoff_factor, debug)
-    if loop_threshold < 2:
-        raise ValueError(f"loop_threshold must be at least 2, got {loop_threshold}")
-    if time_window <= 0:
-        raise ValueError(f"time_window must be positive, got {time_window}")
-    if content_trigger_threshold < 1:
-        raise ValueError(f"content_trigger_threshold must be at least 1, got {content_trigger_threshold}")
-    if max_retries < 0:
-        raise ValueError(f"max_retries must be non-negative, got {max_retries}")
-    if backoff_factor <= 0:
-        raise ValueError(f"backoff_factor must be positive, got {backoff_factor}")
-
-    def decorator(func):
-        return _create_wrapper(func, resource_type, loop_threshold, time_window, ad_cooldown, content_trigger_threshold,
-                             max_retries, backoff_factor, debug)
-    return decorator
-
-def _create_wrapper(func, resource_type, loop_threshold, time_window, ad_cooldown, content_trigger_threshold,
-                   max_retries, backoff_factor, debug):
-    call_history = []
-    last_ad_time = 0
-    consecutive_loop_detections = 0
-    session_displayed = False
-    session_start_time = time.time()
-    session_timeout = 1800
-    @functools.wraps(func)
-
-    def wrapper(*args, **kwargs):
-        nonlocal last_ad_time, consecutive_loop_detections, session_displayed, session_start_time
-        current_time = time.time()
-        content_triggered = False
-        if current_time - session_start_time > session_timeout:
-            session_displayed = False
-            session_start_time = current_time
-        retries = 0
-        while True:
-            call_history.append(current_time)
-            while call_history and current_time - call_history[0] > time_window:
-                call_history.pop(0)
-            loop_detected = len(call_history) >= loop_threshold
-            if debug and loop_detected:
-                print(f"[OPTIMIZE] Đã phát hiện vòng lặp cho {func.__name__}: {len(call_history)} lần gọi trong {time_window}s")
-            if loop_detected:
-                consecutive_loop_detections += 1
-                if debug:
-                    print(f"[OPTIMIZE] Số lần phát hiện vòng lặp liên tiếp: {consecutive_loop_detections}/{content_trigger_threshold}")
-            else:
-                consecutive_loop_detections = 0
-            should_show_content = (consecutive_loop_detections >= content_trigger_threshold) and                                 (current_time - last_ad_time >= ad_cooldown) and                                 not session_displayed
-            if should_show_content:
-                last_ad_time = current_time
-                consecutive_loop_detections = 0
-                content_triggered = True
-                session_displayed = True
-                if debug:
-                    print(f"[OPTIMIZE] Đã kích hoạt nội dung cho {func.__name__}")
-                try:
-                    from vnai.scope.promo import manager
-                    try:
-                        from vnai.scope.profile import inspector
-                        environment = inspector.examine().get("environment", None)
-                        manager.present_content(environment=environment, context="loop")
-                    except ImportError:
-                        manager.present_content(context="loop")
-                except ImportError:
-                    print(f"Phát hiện vòng lặp: Hàm '{func.__name__}' đang được gọi trong một vòng lặp")
-                except Exception as e:
-                    if debug:
-                        print(f"[OPTIMIZE] Lỗi khi hiển thị nội dung: {str(e)}")
-            try:
-                with CleanErrorContext():
-                    guardian.verify(func.__name__, resource_type)
-            except RateLimitExceeded as e:
-                from vnai.beam.metrics import collector
-                collector.record(
-"error",
-                    {
-"function": func.__name__,
-"error": str(e),
-"context":"resource_verification",
-"resource_type": resource_type,
-"retry_attempt": retries
-                    },
-                    priority="high"
-                )
-                if not session_displayed:
-                    try:
-                        from vnai.scope.promo import manager
-                        try:
-                            from vnai.scope.profile import inspector
-                            environment = inspector.examine().get("environment", None)
-                            manager.present_content(environment=environment, context="loop")
-                            session_displayed = True
-                            last_ad_time = current_time
-                        except ImportError:
-                            manager.present_content(context="loop")
-                            session_displayed = True
-                            last_ad_time = current_time
-                    except Exception:
-                        pass
-                if retries < max_retries:
-                    wait_time = backoff_factor ** retries
-                    retries += 1
-                    if hasattr(e,"retry_after") and e.retry_after:
-                        wait_time = min(wait_time, e.retry_after)
-                    if debug:
-                        print(f"[OPTIMIZE] Đã đạt giới hạn tốc độ cho {func.__name__}, thử lại sau {wait_time} giây (lần thử {retries}/{max_retries})")
-                    time.sleep(wait_time)
-                    continue
-                else:
-                    raise
-            start_time = time.time()
-            success = False
-            error = None
-            try:
-                result = func(*args, **kwargs)
-                success = True
-                return result
-            except Exception as e:
-                error = str(e)
-                raise
-            finally:
-                execution_time = time.time() - start_time
-                try:
-                    from vnai.beam.metrics import collector
-                    collector.record(
-"function",
-                        {
-"function": func.__name__,
-"resource_type": resource_type,
-"execution_time": execution_time,
-"success": success,
-"error": error,
-"in_loop": loop_detected,
-"loop_depth": len(call_history),
-"content_triggered": content_triggered,
-"timestamp": datetime.now().isoformat(),
-"retry_count": retries if retries > 0 else None
-                        }
-                    )
-                    if content_triggered:
-                        collector.record(
-"ad_opportunity",
-                            {
-"function": func.__name__,
-"resource_type": resource_type,
-"call_frequency": len(call_history) / time_window,
-"consecutive_loops": consecutive_loop_detections,
-"timestamp": datetime.now().isoformat()
-                            }
-                        )
-                except ImportError:
-                    pass
-            break
-    return wrapper
-
-def rate_limit_status(resource_type="default"):
-    return guardian.get_limit_status(resource_type)
+# vnai/beam/quota.py
+
+import time
+import functools
+import threading
+from collections import defaultdict
+from datetime import datetime
+
+class RateLimitExceeded(Exception):
+    """Custom exception for rate limit violations."""
+    def __init__(self, resource_type, limit_type="min", current_usage=None, limit_value=None, retry_after=None):
+        self.resource_type = resource_type
+        self.limit_type = limit_type
+        self.current_usage = current_usage
+        self.limit_value = limit_value
+        self.retry_after = retry_after
+        
+        # Create a user-friendly message
+        message = f"Bạn đã gửi quá nhiều request tới {resource_type}. "
+        if retry_after:
+            message += f"Vui lòng thử lại sau {round(retry_after)} giây."
+        else:
+            message += "Vui lòng thêm thời gian chờ giữa các lần gửi request."
+            
+        super().__init__(message)
+
+class Guardian:
+    """Ensures optimal resource allocation"""
+    
+    _instance = None
+    _lock = threading.Lock()
+    
+    def __new__(cls):
+        with cls._lock:
+            if cls._instance is None:
+                cls._instance = super(Guardian, cls).__new__(cls)
+                cls._instance._initialize()
+            return cls._instance
+    
+    def _initialize(self):
+        """Initialize guardian"""
+        self.resource_limits = defaultdict(lambda: defaultdict(int))
+        self.usage_counters = defaultdict(lambda: defaultdict(list))
+        
+        # Define resource limits
+        self.resource_limits["default"] = {"min": 60, "hour": 3000}
+        self.resource_limits["TCBS"] = {"min": 60, "hour": 3000}
+        self.resource_limits["VCI"] = {"min": 60, "hour": 3000}
+        self.resource_limits["MBK"] = {"min": 600, "hour": 36000}
+        self.resource_limits["MAS.ext"] = {"min": 600, "hour": 36000}
+        self.resource_limits["VCI.ext"] = {"min": 600, "hour": 36000}
+        self.resource_limits["FMK.ext"] = {"min": 600, "hour": 36000}
+        self.resource_limits["VND.ext"] = {"min": 600, "hour": 36000}
+        self.resource_limits["CAF.ext"] = {"min": 600, "hour": 36000}
+        self.resource_limits["SPL.ext"] = {"min": 600, "hour": 36000}
+        self.resource_limits["VDS.ext"] = {"min": 600, "hour": 36000}
+        self.resource_limits["FAD.ext"] = {"min": 600, "hour": 36000}
+    
+    def verify(self, operation_id, resource_type="default"):
+        """Verify resource availability before operation"""
+        current_time = time.time()
+        
+        # Get limits for this resource type (or use default)
+        limits = self.resource_limits.get(resource_type, self.resource_limits["default"])
+        
+        # Check minute limit
+        minute_cutoff = current_time - 60
+        self.usage_counters[resource_type]["min"] = [
+            t for t in self.usage_counters[resource_type]["min"] 
+            if t > minute_cutoff
+        ]
+        
+        minute_usage = len(self.usage_counters[resource_type]["min"])
+        minute_exceeded = minute_usage >= limits["min"]
+        
+        if minute_exceeded:
+            # Track limit check through metrics module
+            from vnai.beam.metrics import collector
+            collector.record(
+                "rate_limit",
+                {
+                    "resource_type": resource_type,
+                    "limit_type": "min",
+                    "limit_value": limits["min"],
+                    "current_usage": minute_usage,
+                    "is_exceeded": True
+                },
+                priority="high"
+            )
+            # Raise custom exception with retry information
+            raise RateLimitExceeded(
+                resource_type=resource_type, 
+                limit_type="min",
+                current_usage=minute_usage,
+                limit_value=limits["min"],
+                retry_after=60 - (current_time % 60)  # Seconds until the minute rolls over
+            )
+            
+        # Check hour limit
+        hour_cutoff = current_time - 3600
+        self.usage_counters[resource_type]["hour"] = [
+            t for t in self.usage_counters[resource_type]["hour"] 
+            if t > hour_cutoff
+        ]
+        
+        hour_usage = len(self.usage_counters[resource_type]["hour"])
+        hour_exceeded = hour_usage >= limits["hour"]
+        
+        # Track rate limit check
+        from vnai.beam.metrics import collector
+        collector.record(
+            "rate_limit",
+            {
+                "resource_type": resource_type,
+                "limit_type": "hour" if hour_exceeded else "min",
+                "limit_value": limits["hour"] if hour_exceeded else limits["min"],
+                "current_usage": hour_usage if hour_exceeded else minute_usage,
+                "is_exceeded": hour_exceeded
+            }
+        )
+        
+        if hour_exceeded:
+            # Raise custom exception with retry information
+            raise RateLimitExceeded(
+                resource_type=resource_type, 
+                limit_type="hour",
+                current_usage=hour_usage,
+                limit_value=limits["hour"],
+                retry_after=3600 - (current_time % 3600)  # Seconds until the hour rolls over
+            )
+            
+        # Record this request
+        self.usage_counters[resource_type]["min"].append(current_time)
+        self.usage_counters[resource_type]["hour"].append(current_time)
+        return True
+    
+    def usage(self, resource_type="default"):
+        """Get current usage percentage for resource limits"""
+        current_time = time.time()
+        limits = self.resource_limits.get(resource_type, self.resource_limits["default"])
+        
+        # Clean old timestamps
+        minute_cutoff = current_time - 60
+        hour_cutoff = current_time - 3600
+        
+        self.usage_counters[resource_type]["min"] = [
+            t for t in self.usage_counters[resource_type]["min"] 
+            if t > minute_cutoff
+        ]
+        
+        self.usage_counters[resource_type]["hour"] = [
+            t for t in self.usage_counters[resource_type]["hour"] 
+            if t > hour_cutoff
+        ]
+        
+        # Calculate percentages
+        minute_usage = len(self.usage_counters[resource_type]["min"])
+        hour_usage = len(self.usage_counters[resource_type]["hour"])
+        
+        minute_percentage = (minute_usage / limits["min"]) * 100 if limits["min"] > 0 else 0
+        hour_percentage = (hour_usage / limits["hour"]) * 100 if limits["hour"] > 0 else 0
+        
+        # Return the higher percentage
+        return max(minute_percentage, hour_percentage)
+    
+    def get_limit_status(self, resource_type="default"):
+        """Get detailed information about current limit status"""
+        current_time = time.time()
+        limits = self.resource_limits.get(resource_type, self.resource_limits["default"])
+        
+        # Clean old timestamps
+        minute_cutoff = current_time - 60
+        hour_cutoff = current_time - 3600
+        
+        minute_usage = len([t for t in self.usage_counters[resource_type]["min"] if t > minute_cutoff])
+        hour_usage = len([t for t in self.usage_counters[resource_type]["hour"] if t > hour_cutoff])
+        
+        return {
+            "resource_type": resource_type,
+            "minute_limit": {
+                "usage": minute_usage,
+                "limit": limits["min"],
+                "percentage": (minute_usage / limits["min"]) * 100 if limits["min"] > 0 else 0,
+                "remaining": max(0, limits["min"] - minute_usage),
+                "reset_in_seconds": 60 - (current_time % 60)
+            },
+            "hour_limit": {
+                "usage": hour_usage,
+                "limit": limits["hour"],
+                "percentage": (hour_usage / limits["hour"]) * 100 if limits["hour"] > 0 else 0,
+                "remaining": max(0, limits["hour"] - hour_usage),
+                "reset_in_seconds": 3600 - (current_time % 3600)
+            }
+        }
+
+# Create singleton instance
+guardian = Guardian()
+
+class CleanErrorContext:
+    """Context manager to clean up tracebacks for rate limits"""
+    # Class variable to track if a message has been displayed recently
+    _last_message_time = 0
+    _message_cooldown = 5  # Only show a message every 5 seconds
+    
+    def __enter__(self):
+        return self
+        
+    def __exit__(self, exc_type, exc_val, exc_tb):
+        if exc_type is RateLimitExceeded:
+            current_time = time.time()
+            
+            # Only print the message if enough time has passed since the last one
+            if current_time - CleanErrorContext._last_message_time >= CleanErrorContext._message_cooldown:
+                print(f"\n⚠️ {str(exc_val)}\n")
+                CleanErrorContext._last_message_time = current_time
+            
+            # Re-raise the exception more forcefully to ensure it propagates
+            # This will bypass any try/except blocks that might be catching RateLimitExceeded
+            import sys
+            sys.exit(f"Rate limit exceeded. {str(exc_val)} Process terminated.")
+            
+            # The line below won't be reached, but we keep it for clarity
+            return False
+        return False
+
+
+def optimize(resource_type='default', loop_threshold=10, time_window=5, ad_cooldown=150, content_trigger_threshold=3, 
+            max_retries=2, backoff_factor=2, debug=False):
+    """
+    Decorator that optimizes function execution, tracks metrics, and detects loop patterns for ad opportunities.
+    
+    Features:
+    - Resource verification
+    - Performance metrics collection
+    - Loop detection for ad/content opportunities
+    - Automatic retry with exponential backoff for rate limit errors
+    
+    Args:
+        resource_type: Type of resource used by function ("network", "database", "cpu", "memory", "io", "default")
+        loop_threshold: Number of calls within time_window to consider as a loop (min: 2)
+        time_window: Time period in seconds to consider for loop detection
+        ad_cooldown: Minimum seconds between showing ads for the same function
+        content_trigger_threshold: Number of consecutive loop detections before triggering content (min: 1)
+        max_retries: Maximum number of times to retry when rate limits are hit
+        backoff_factor: Base factor for exponential backoff (wait time = backoff_factor^retry_count)
+        debug: When True, prints diagnostic information about loop detection
+    
+    Examples:
+        @optimize
+        def simple_function():
+            return "result"
+            
+        @optimize("network")
+        def fetch_stock_data(symbol):
+            # Makes network calls
+            return data
+            
+        @optimize("database", loop_threshold=4, time_window=10)
+        def query_financial_data(params):
+            # Database queries
+            return results
+    """
+    # Handle case where decorator is used without arguments: @optimize
+    if callable(resource_type):
+        func = resource_type
+        return _create_wrapper(func, 'default', loop_threshold, time_window, ad_cooldown, content_trigger_threshold, 
+                             max_retries, backoff_factor, debug)
+    
+    # Basic validation
+    if loop_threshold < 2:
+        raise ValueError(f"loop_threshold must be at least 2, got {loop_threshold}")
+    if time_window <= 0:
+        raise ValueError(f"time_window must be positive, got {time_window}")
+    if content_trigger_threshold < 1:
+        raise ValueError(f"content_trigger_threshold must be at least 1, got {content_trigger_threshold}")
+    if max_retries < 0:
+        raise ValueError(f"max_retries must be non-negative, got {max_retries}")
+    if backoff_factor <= 0:
+        raise ValueError(f"backoff_factor must be positive, got {backoff_factor}")
+    
+    # Return the actual decorator
+    def decorator(func):
+        return _create_wrapper(func, resource_type, loop_threshold, time_window, ad_cooldown, content_trigger_threshold, 
+                             max_retries, backoff_factor, debug)
+    return decorator
+
+def _create_wrapper(func, resource_type, loop_threshold, time_window, ad_cooldown, content_trigger_threshold, 
+                   max_retries, backoff_factor, debug):
+    """Creates the function wrapper with call tracking for loop detection"""
+    # Static storage for each decorated function instance
+    call_history = []
+    last_ad_time = 0
+    consecutive_loop_detections = 0
+    session_displayed = False  # Track if we've displayed an ad in this session
+    session_start_time = time.time()
+    session_timeout = 1800  # 30 minutes for session expiration
+    
+    @functools.wraps(func)
+    def wrapper(*args, **kwargs):
+        nonlocal last_ad_time, consecutive_loop_detections, session_displayed, session_start_time
+        current_time = time.time()
+        content_triggered = False
+        
+        # Reset session if it has expired
+        if current_time - session_start_time > session_timeout:
+            session_displayed = False
+            session_start_time = current_time
+            
+        # For automatic retries with rate limits
+        retries = 0
+        while True:
+            # ===== LOOP DETECTION LOGIC =====
+            # Add current call to history
+            call_history.append(current_time)
+            
+            # Prune old calls outside the time window
+            while call_history and current_time - call_history[0] > time_window:
+                call_history.pop(0)
+            
+            # Check if we're in a loop pattern
+            loop_detected = len(call_history) >= loop_threshold
+            
+            if debug and loop_detected:
+                print(f"[OPTIMIZE] Đã phát hiện vòng lặp cho {func.__name__}: {len(call_history)} lần gọi trong {time_window}s")
+            
+            # Handle loop detection
+            if loop_detected:
+                consecutive_loop_detections += 1
+                if debug:
+                    print(f"[OPTIMIZE] Số lần phát hiện vòng lặp liên tiếp: {consecutive_loop_detections}/{content_trigger_threshold}")
+            else:
+                consecutive_loop_detections = 0
+            
+            # Determine if we should show content - add session_displayed check
+            should_show_content = (consecutive_loop_detections >= content_trigger_threshold) and \
+                                 (current_time - last_ad_time >= ad_cooldown) and \
+                                 not session_displayed
+            
+            # Handle content opportunity
+            if should_show_content:
+                last_ad_time = current_time
+                consecutive_loop_detections = 0
+                content_triggered = True
+                session_displayed = True  # Mark that we've displayed in this session
+                
+                if debug:
+                    print(f"[OPTIMIZE] Đã kích hoạt nội dung cho {func.__name__}")
+                
+                # Trigger content display using promo manager with "loop" context
+                try:
+                    from vnai.scope.promo import manager
+                    
+                    # Get environment if available
+                    try:
+                        from vnai.scope.profile import inspector
+                        environment = inspector.examine().get("environment", None)
+                        manager.present_content(environment=environment, context="loop")
+                    except ImportError:
+                        manager.present_content(context="loop")
+                        
+                except ImportError:
+                    # Fallback if content manager is not available
+                    print(f"Phát hiện vòng lặp: Hàm '{func.__name__}' đang được gọi trong một vòng lặp")
+                except Exception as e:
+                    # Don't let content errors affect the main function
+                    if debug:
+                        print(f"[OPTIMIZE] Lỗi khi hiển thị nội dung: {str(e)}")
+            
+            # ===== RESOURCE VERIFICATION =====
+            try:
+                # Use a context manager to clean up the traceback
+                with CleanErrorContext():
+                    guardian.verify(func.__name__, resource_type)
+                    
+            except RateLimitExceeded as e:
+                # Record the rate limit error
+                from vnai.beam.metrics import collector
+                collector.record(
+                    "error",
+                    {
+                        "function": func.__name__,
+                        "error": str(e),
+                        "context": "resource_verification",
+                        "resource_type": resource_type,
+                        "retry_attempt": retries
+                    },
+                    priority="high"
+                )
+                
+                # Display rate limit content ONLY if we haven't shown any content this session
+                if not session_displayed:
+                    try:
+                        from vnai.scope.promo import manager
+                        try:
+                            from vnai.scope.profile import inspector
+                            environment = inspector.examine().get("environment", None)
+                            manager.present_content(environment=environment, context="loop")
+                            session_displayed = True  # Mark that we've displayed
+                            last_ad_time = current_time
+                        except ImportError:
+                            manager.present_content(context="loop")
+                            session_displayed = True
+                            last_ad_time = current_time
+                    except Exception:
+                        pass  # Don't let content errors affect the retry logic
+                
+                # Continue with retry logic
+                if retries < max_retries:
+                    wait_time = backoff_factor ** retries
+                    retries += 1
+                    
+                    # If the exception has a retry_after value, use that instead
+                    if hasattr(e, "retry_after") and e.retry_after:
+                        wait_time = min(wait_time, e.retry_after)
+                    
+                    if debug:
+                        print(f"[OPTIMIZE] Đã đạt giới hạn tốc độ cho {func.__name__}, thử lại sau {wait_time} giây (lần thử {retries}/{max_retries})")
+                    
+                    time.sleep(wait_time)
+                    continue  # Retry the call
+                else:
+                    # No more retries, re-raise the exception
+                    raise
+            
+            # ===== FUNCTION EXECUTION & METRICS =====
+            start_time = time.time()
+            success = False
+            error = None
+            
+            try:
+                # Execute the original function
+                result = func(*args, **kwargs)
+                success = True
+                return result
+            except Exception as e:
+                error = str(e)
+                raise
+            finally:
+                # Calculate execution metrics
+                execution_time = time.time() - start_time
+                
+                # Record metrics
+                try:
+                    from vnai.beam.metrics import collector
+                    collector.record(
+                        "function",
+                        {
+                            "function": func.__name__,
+                            "resource_type": resource_type,
+                            "execution_time": execution_time,
+                            "success": success,
+                            "error": error,
+                            "in_loop": loop_detected,
+                            "loop_depth": len(call_history),
+                            "content_triggered": content_triggered,
+                            "timestamp": datetime.now().isoformat(),
+                            "retry_count": retries if retries > 0 else None
+                        }
+                    )
+                    
+                    # Record content opportunity metrics if detected
+                    if content_triggered:
+                        collector.record(
+                            "ad_opportunity",
+                            {
+                                "function": func.__name__,
+                                "resource_type": resource_type,
+                                "call_frequency": len(call_history) / time_window,
+                                "consecutive_loops": consecutive_loop_detections,
+                                "timestamp": datetime.now().isoformat()
+                            }
+                        )
+                except ImportError:
+                    # Metrics module not available, just continue
+                    pass
+                
+            # If we got here, the function executed successfully, so break the retry loop
+            break
+                
+    return wrapper
+
+
+# Helper function for getting the current rate limit status
+def rate_limit_status(resource_type="default"):
+    """Get the current rate limit status for a resource type"""
+    return guardian.get_limit_status(resource_type)