SNN2.src.contextManagers.contextManagers module

Context Managers Module

This module provides context managers and utilities for timing code execution and handling augmented class names within the SNN2 framework.

The module contains both function-based and class-based context managers that can be used to measure execution times of code blocks, with optional logging support. These tools are essential for performance monitoring and profiling in neural network training and data processing pipelines.

Functions

get_realNamefunction: Extract the real name from objects, handling augmented classes.
timeit_cntcontextmanager: Context manager for timing code execution with logging support.

Classes

TimeitContextclass: Reusable context manager for collecting timing statistics across multiple runs.

Examples

Basic timing with logging:

>>> with timeit_cnt("data_processing"):
...     # Some data processing code
...     process_data()
data_processing execution time: 1234.567 ms

Reusable timing context:

>>> timer = TimeitContext()
>>> for i in range(3):
...     with timer():
...         time.sleep(0.1)
>>> print(f"Mean execution time: {timer.mean:.3f}s")
Mean execution time: 0.100s

Notes

This module integrates with the SNN2 logging system through the f_logger decorator, providing consistent logging behavior across the framework. All timing measurements are performed using Python’s time.time() function for high-resolution timing.

See also

SNN2.src.decorators.decorators: Decorator functions including f_logger
SNN2.src.io.logger: Logging utilities and LogHandler class

class SNN2.src.contextManagers.contextManagers.TimeitContext

Bases: object

A reusable context manager for timing multiple code executions.

This class provides a context manager that can be reused multiple times to measure execution times and collect statistics about performance.

all_times

List storing all recorded execution times in seconds.

Type:: list of float

execution_time

The most recent execution time in seconds.

Type:: float or None

n_runs

Number of times the context manager has been used.

Type:: int

Examples

>>> timer = TimeitContext()
>>> with timer():
...     time.sleep(0.1)
>>> print(f"Execution time: {timer.execution_time:.3f}s")
>>> print(f"Mean time: {timer.mean:.3f}s")

property mean

Calculate the mean execution time across all runs.

Returns:: The average execution time in seconds across all recorded runs.
Return type:: float
Raises:: ZeroDivisionError – If no runs have been recorded (all_times is empty).

Notes

This property computes the arithmetic mean of all execution times stored in the all_times list.

SNN2.src.contextManagers.contextManagers.get_realName(obj: Any)

Get the real name of an object, handling augmented classes.

Parameters:: obj (Any) – The object whose real name should be retrieved.
Returns:: The real name of the object. If the object is an augmented class with a custom name, returns the passed name, otherwise returns the qualified name.
Return type:: str

Notes

This function is designed to handle special cases where objects have been augmented and their original name has been stored in a c_passed_name attribute.

SNN2.src.contextManagers.contextManagers.timeit_cnt(name, active: bool = True, *, logger=None, write_msg=<function f_logger.<locals>.__dummy_log>, **kwargs)

Context manager for timing code execution with logging support.

This context manager measures the execution time of a code block and optionally logs the result using the provided logger.

Parameters:

name (str) – Name identifier for the timed operation, used in log messages.
*args (tuple) – Variable length argument list (not used in current implementation).
active (bool, default=True) – Whether to actively log and print timing results.
**kwargs (dict) – Keyword arguments containing logger and write_msg from f_logger decorator.

Yields:

None – Control is yielded to the code block being timed.

Notes

This function is decorated with @f_logger which automatically provides logger and write_msg in kwargs. The timing results are both logged and printed to stdout when active=True.

Examples

>>> with timeit_cnt("my_operation"):
...     # Some time-consuming operation
...     time.sleep(1)
my_operation execution time: 1000.123 ms