How to create a profiler utils decorator

To create a new profiler utils decorator for the pydecorium.decorators.FunctionProfiler decorator, you need to create a class that inherits from the base class pydecorium.decorators.ProfilerUtils.

First import the class ProfilerUtils from the module pydecorium.decorators:

from pydecorium.decorators import ProfilerUtils

Then create a new class that inherits from the class ProfilerUtils:

class MyProfilerUtils(ProfilerUtils):

    def __init__(self, *args, **kwargs):
        super().__init__(*args, **kwargs)

Then you need to add the following methods and attributes:

• data_name: This attribute is a string that represents the name of the data that will be profiled.

• pre_execute: This method is called before the execution of the decorated function. It arguments are the function and the arguments of the function.

• post_execute: This method is called after the execution of the decorated function. It arguments are the function and the outputs of the function.

• handle_result: This method is called after the execution of the decorated function and has not arguments. It should return the raw data of the profiler utils.

• string_value: This method takes the raw data of the profiler utils as argument and returns a string representation of the profiler utils.

Example

The following example shows how to create a simple profiler utils decorator that measures the runtime of a function. This example is similar than pydecorium.decorators.Timer decorator.

import time

class MyProfilerUtils(ProfilerUtils):

    data_name = "runtime"

    def __init__(self, *args, **kwargs):
        super().__init__(*args, **kwargs)

    def pre_execute(self, func, *args, **kwargs):
        self.tic = time.time()

    def post_execute(self, func, outputs):
        self.toc = time.time()

    def handle_result(self):
        return self.toc - self.tic

    def string_value(self, raw_data):
        return f"{raw_data:.2f} seconds"