[python/en] Updated Decorator and wrapping explanation (#4749)

Now it includes motivation, an explanation of functools.wraps, and demonstrates the utility of wrapping.

python.html.markdown

@@ -1016,31 +1016,67 @@ gen_to_list = list(values)
 print(gen_to_list)  # => [-1, -2, -3, -4, -5]
 
 
-# Decorators
+# Decorators are a form of syntactic sugar.
-# In this example `beg` wraps `say`. If say_please is True then it
+# They make code easier to read while accomplishing clunky syntax.
-# will change the returned message.
-from functools import wraps
 
+# Wrappers are one type of decorator.
+# They're really useful for adding logging to existing functions without needing to modify them.
 
-def beg(target_function):
+def log_function(func):
-    @wraps(target_function)
     def wrapper(*args, **kwargs):
-        msg, say_please = target_function(*args, **kwargs)
+        print("Entering function", func.__name__)
-        if say_please:
+        result = func(*args, **kwargs)
-            return "{} {}".format(msg, "Please! I am poor :(")
+        print("Exiting function", func.__name__)
-        return msg
+        return result
-
     return wrapper
 
+@log_function               # equivalent:
+def my_function(x,y):       # def my_function(x,y):
+    return x+y              #   return x+y
+                            # my_function = log_function(my_function)
+# The decorator @log_function tells us as we begin reading the function definition
+# for my_function that this function will be wrapped with log_function.
+# When function definitions are long, it can be hard to parse the non-decorated
+# assignment at the end of the definition.
 
-@beg
+my_function(1,2) # => "Entering function my_function"
-def say(say_please=False):
+                 # => "3"
-    msg = "Can you buy me a beer?"
+                 # => "Exiting function my_function"
-    return msg, say_please
 
+# But there's a problem.
+# What happens if we try to get some information about my_function?
+
+print(my_function.__name__) # => 'wrapper'
+print(my_function.__code__.co_argcount) # => 0. The argcount is 0 because both arguments in wrapper()'s signature are optional.
+
+# Because our decorator is equivalent to my_function = log_function(my_function)
+# we've replaced information about my_function with information from wrapper
+
+# Fix this using functools
+
+from functools import wraps
+
+def log_function(func):
+    @wraps(func) # this ensures docstring, function name, arguments list, etc. are all copied
+                 # to the wrapped function - instead of being replaced with wrapper's info
+    def wrapper(*args, **kwargs):
+        print("Entering function", func.__name__)
+        result = func(*args, **kwargs)
+        print("Exiting function", func.__name__)
+        return result
+    return wrapper
+
+@log_function               
+def my_function(x,y):       
+    return x+y              
+                            
+my_function(1,2) # => "Entering function my_function"
+                 # => "3"
+                 # => "Exiting function my_function"
+
+print(my_function.__name__) # => 'my_function'
+print(my_function.__code__.co_argcount) # => 2
 
-print(say())                 # Can you buy me a beer?
-print(say(say_please=True))  # Can you buy me a beer? Please! I am poor :(
 ```
 
 ### Free Online