profiling-helpers

Profiling Helpers

A small python package that wraps Python's own cProfile library to make it more user friendly.

When developing Python programs, you'll sometimes have functions that take a long time to execute and you are really not sure why. Profiling helps to find and analyze these bottlenecks and guides you into fixing performance problems. Uses snakeviz for interactive visualizations.

Install it with pip install profiling-helpers. Visualize profile files with snakeviz profile_xyz.prof.

There are two decorators, time_it and profile_it. Use them anywhere in your code, like this:

from profiling_helpers import time_it, profile_it
from time import sleep

@time_it
def my_slow_function(x):
    sleep(10)
    return x

my_slow_function(42)  # Prints: Function "my_slow_function" took 10.01061 s to run

@profile_it("my/profile/save/dir", open_visualization=True)
def my_slow_function(x):
    sleep(10)
    return x

my_slow_function(42)  # Opens snakeviz after this function is completed

Profiles are normally saved on the local file system. If you have other save targets, you can either use included FileSavers (currently only for AWS S3) or implement your own one by inheriting from the BaseFileSaver class. Here is a variant with S3:

from profiling_helpers import profile_it, S3FileSaver

@profile_it(S3FileSaver("s3://my-bucket/my/path/to/profiles/", kms_key_id="..."))
def my_slow_function(x):
    sleep(10)
    return x

my_slow_function(42)