Logging in Kerch

Many classes throughout this package display various messages during instantiation and usage. This abstract from which they descend allows the dislpaying of those messages and control which are printed using the logging_level attribute.

Functionalities

This abstract class has only one purpose: adding a _logger attribute meant to log various messages across the package. Doing it this way allows to get the name of the class instantiated and print more precise messages.

Logging Messages

The private property self._logger is an instance of the class logging.Logger of the Python standard library logging — Logging facility for Python. This allows logging messages by calling, e.g., self._logger.debug(message: str), self._logger.info(message: str), self._logger.warning(message: str) or self._logger.error(message: str). The messages will be automatically formatted to reference the appropriate class and give file and line information in debug mode.

Logging Level

A particular log level can be set for each instance using the attribute logging_level. If assigned to None,the default logging level will be used. This value is also set during instantiation using the logging_level argument. If nothing is specified, None is passed leading to the genral default logging level.

The log level always corresponds to an integer. We refer to logging — Logging facility for Python for the different values.

Default Logging Level

This is the default logging level that is used if logging_level is never specified (which corresponds to setting it to None). By default, the logging level is 30, which corresponds to a warning level. This default package value can also be changed and read using the following functions.

kerch.set_logging_level(level: int)[source]

Changes the default logging level of the kerch package.

Parameters:: level (int) – Default kerch logging value

Usage:

import kerch
import logging

kerch.set_logging_level(logging.DEBUG)

Warning

Changing the default logging value does not affect the already instantiated objects. We advise to set those values in the beginning of the code.

kerch.get_logging_level() → int[source]

Returns the default logging level of the kerch package.

Usage:

import kerch
import logging

default_level = kerch.get_logging_level()
default_level = logging.getLevelName(default_level)
print(default_level)

WARNING

Abstract Class

class kerch.feature.Logger(*args, **kwargs)[source]

Bases: object

Parameters:: logging_level (int, optional) – Logging level for this specific instance. If the value is None, the current default kerch global log level will be used. Defaults to None (default kerch logging level). We refer to the Logging in Kerch documentation for further information.

property _logger: Logger

Logger of the instance.

Usage:

import kerch
import logging

class MyClass(kerch.feature.Logger):
    def __init__(self, *args, **kwargs):
        super(MyClass, self).__init__(*args, **kwargs)
        self._logger.info('Instantiation done information.')
        self._logger.warn('Instantiation done warning.')

print('First class with default logging level:')
my_class1 = MyClass()

print('\nSecond instance with logging.INFO logging level:')
my_class2 = MyClass(logging_level=logging.INFO)

First class with default logging level:
WARNING:MyClass:Instantiation done warning.

Second instance with logging.INFO logging level:
INFO:MyClass:Instantiation done information.
WARNING:MyClass:Instantiation done warning.

property logging_level: int: Logging level of this specific instance. If the value is None, the current default kerch global log Level will be used. Defaults to None (default global kerch level). We refer to the Logging in Kerch documentation for further information.

Example

Default

import kerch

k = kerch.kernel.RBF()
k.sample = range(3)
print(k.K)                                          # first warning (the sigma is defined)

WARNING:RBF:The kernel bandwidth sigma has not been provided and is assigned by a heuristic (sigma=5.00e-01).
tensor([[1.0000e+00, 1.3534e-01, 3.3546e-04],
        [1.3534e-01, 1.0000e+00, 1.3534e-01],
        [3.3546e-04, 1.3534e-01, 1.0000e+00]])

Info

import kerch
import logging

k = kerch.kernel.RBF(logging_level=logging.INFO)    # first info (no sample initialized yet)
k.sample = range(3)                                 # second info (the sample is initialized)
print(k.K)                                          # first warning (the sigma is defined)

INFO:RBF:The sample cannot be initialized because no sample data has been provided nor the sample dimensions have been initialized yet.
WARNING:RBF:The kernel bandwidth sigma has not been provided and is assigned by a heuristic (sigma=5.00e-01).
tensor([[1.0000e+00, 1.3534e-01, 3.3546e-04],
        [1.3534e-01, 1.0000e+00, 1.3534e-01],
        [3.3546e-04, 1.3534e-01, 1.0000e+00]])

Inheritance

This is a base class that directly inherits from object.