FunctionProfiler Usage

The pydecorium.decorators.FunctionProfiler decorator is used to profile some data about the function execution as the runtime, the memory usage, etc. FunctionProfiler is a sub-class of the class pydecorium.Decorator.

First we need to import the FunctionProfiler decorator with the following command:

from pydecorium.decorators import FunctionProfiler

Then we can connect some pydecorium.decorators.ProfilerUtils decorators to the FunctionProfiler decorator. By connecting a ProfilerUtils decorator to the FunctionProfiler decorator, the ProfilerUtils decorator will be use by the FunctionProfiler decorator to profile a given data about the function execution. If several ProfilerUtils decorators are connected to the FunctionProfiler decorator, all the connected ProfilerUtils decorators will be used to profile the function execution.

In this example, we will connect the pydecorium.decorators.Timer and pydecorium.decorators.Memory decorators to the FunctionLogger decorator. To do this, we use the profiler_utils argument of the FunctionProfiler decorator. We can also use the method connect_profiler_utils of the FunctionProfiler decorator as follows:

from pydecorium.decorators import Timer, Memory

function_profiler = FunctionProfiler(
    activated=True,
    signature_name_format='{name}',
    profiler_utils=[Timer, Memory],
    report_format='datetime')

# OR
fonction_profiler.connect_profiler_utils([Timer, Memory])

If we try to connect a ProfilerUtils decorator already connected to the FunctionProfiler decorator, the ProfilerUtils decorator will be connected only once.

Then we can use the FunctionProfiler as describe in the documentation How to use a decorator.

@function_profiler
def example_function():
    import time
    time.sleep(10)
    return [i for i in range(1000000)]

@function_profiler
def other_example_function():
    import time
    time.sleep(5)
    return [i for i in range(5000)]

example_function()
example_function()
other_example_function()
example_function()

To see the report of the FunctionProfiler decorator, just print function_profiler.

print(function_profiler)

The output when printing the function_profiler object depends on the report_format attribute of the FunctionProfiler object.

Selecting the report format

To change the format of the string reporting the function execution, use the method pydecorium.decorators.FunctionProfiler.set_report_format().

If the report_format is set to “datetime”, the output will be:

[2025-01-20 18:23:08.696188] - [example_function] - runtime : 0h 0m 10.0217s - memory usage : 38MB 320KB 0B
[2025-01-20 18:23:18.723690] - [example_function] - runtime : 0h 0m 10.0239s - memory usage : 36MB 256KB 0B
[2025-01-20 18:23:28.752144] - [other_example_function] - runtime : 0h 0m 5.0006s - memory usage : 0MB 0KB 0B
[2025-01-20 18:23:33.753020] - [example_function] - runtime : 0h 0m 10.0304s - memory usage : 28MB 512KB 0B

If the report_format is set to “function”, the output will be:

[example_function]
    [2025-01-20 18:36:44.569795] - runtime : 0h 0m 10.0225s - memory usage : 38MB 380KB 0B
    [2025-01-20 18:36:54.597933] - runtime : 0h 0m 10.0218s - memory usage : 37MB 256KB 0B
    [2025-01-20 18:37:09.626250] - runtime : 0h 0m 10.0193s - memory usage : 28MB 0KB 0B
[other_example_function]
        [2025-01-20 18:37:04.625249] - runtime : 0h 0m 5.0006s - memory : 0MB 0KB 0B

If the report_format is set to “cumulative”, the output will be:

[example_function] - 3 calls - runtime : 0h 0m 30.0666s - memory usage : 103MB 232KB 0B
[other_example_function] - 1 calls - runtime : 0h 0m 5.0008s - memory usage : 0MB 0KB 0B

Add new profiler utils

To add a new profiler utils to the FunctionProfiler decorator, refer to the documentation How to create a profiler utils decorator.