Build a Decorator That Logs Arguments

Decorators can add behavior to existing functions without modifying them. In this exercise you'll write a decorator named log_args that does two things: - Wraps any function and preserves its metadata (use functools.wraps). - On each call, appends a string describing the positional and keyword arguments to a list attached to the decorated function as .log, then returns the original function's result. The log entry format should be exactly: "args=<args_tuple>, kwargs=<kwargs_dict>" where <args_tuple> and <kwargs_dict> are the standard Python representations of the tuple of positional arguments and dictionary of keyword arguments respectively. Examples: - Calling add(2, 3) should produce the log entry: "args=(2, 3), kwargs={}". - Calling power(2, exp=5) should produce: "args=(2,), kwargs={'exp': 5}". You must support any signature (positional, keyword, defaults, varargs). The decorated function should expose the .log attribute (a list) that accumulates these strings across calls.

Implement a decorator @log_args that appends a readable representation of the called arguments (args and kwargs) to a .log list on the decorated function every time it's called.

Do not modify the decorated function's behavior other than adding logging. Use functools.wraps to preserve metadata. The .log attribute must be a list and persist across calls.