5.4. Optimization Context Managers

  • Context Managers

In Python, context managers are implemented using the with statement. The with statement provides a way to wrap a block of code with methods defined by a context manager. When the block of code is entered, the __enter__() method of the context manager is called, and when the block of code is exited, the __exit__() method is called.

5.4.1. Contex Manager

We need to import time() function to get current timestamp (number of seconds from 1970-01-01 00:00:00 UTC):

>>> from time import time

Define our context manager:

>>> class Timeit:
...     def __enter__(self):
...         self.start = time()
...
...     def __exit__(self, *args):
...         self.stop = time()
...         duration = self.stop - self.start
...         print(f'Duration: {duration:.4f} seconds')

>>> #
... with Timeit():
...     result = [x**x for x in range(0, 10_000)]
...
Duration: 5.9882 seconds

5.4.2. Context Decorator Class

  • Inherit from contextlib.ContextDecorator

  • Class become context manager decorator

  • Mind the brackets in decorator @Timeit()

We need to import time() function to get current timestamp (number of seconds from 1970-01-01 00:00:00 UTC). Moreover, this time we need also contextlib.ContextDecorator for our class to inherit from:

>>> from time import time
>>> from contextlib import ContextDecorator

Define our context manager:

>>> class Timeit(ContextDecorator):
...     def __enter__(self):
...         self.start = time()
...
...     def __exit__(self, *args):
...         self.stop = time()
...         duration = self.stop - self.start
...         print(f'Duration: {duration:.4f} seconds')

Define the function which will be automatically wrapped by context manager. Mind the brackets in @Timeit():

>>> @Timeit()
... def run():
...     result = [x**x for x in range(0, 10_000)]

Calling function will result in executing context manager:

>>> run()
Duration: 5.9302 seconds

5.4.3. Context Decorator Function

  • Split function for parts before and after yield

  • Code before yield becomes __enter__()

  • Code after yield becomes __exit__()

We need to import time() function to get current timestamp (number of seconds from 1970-01-01 00:00:00 UTC):

>>> from time import time
>>> from contextlib import contextmanager

Define our context manager. Mind that Python will split our function for parts before and after yield. Code before yield becomes __enter__() and code after yield becomes __exit__():

>>> @contextmanager
... def timeit():
...     start = time()
...     yield
...     end = time()
...     duration = stop - start
...     print(f'Duration: {duration:.4f} seconds')

Now we can use our function as a context manager:

>>> with timeit():
...     result = [x**x for x in range(0, 10_000)]
...
Duration 4.0250 seconds