Quick start

To get started, first make sure logquacious is installed:

$ pip install logquacious

You’ll also need to set up logging for your application. For this example, we’ll use a really simple configuration:

import logging

logging.basicConfig(format='%(levelname)s: %(message)s',
                    level=logging.DEBUG)

Note that this simple configuration is used for demonstration purposes, only. See the Logging Cookbook in the official Python docs for examples of options used for real logging configuration.

The main interface to logquacious is the LogManager, which can be used for normal logging:

import logquacious

log = logquacious.LogManager(__name__)

log.debug('Nothing to see here.')

Due to our simplified logging format defined earlier, that would output:

DEBUG: Nothing to see here.

That isn’t a very interesting example. In addition to basic logging, LogManager has a context attribute for use as a context manager:

>>> with log.context.debug('greetings'):
...    print('Hello!')
DEBUG: Enter greetings
Hello!
DEBUG: Exit greetings

The same attribute can be used as a decorator, as well:

@log.context.info
def divide(numerator, denominator):
    if denominator == 0:
        log.warning('Attempted division by zero. Returning None')
        return None
    return numerator / denominator

>>> divide(1, 0)
INFO: Call `divide()`
WARNING: Attempted division by zero. Returning None
INFO: Return from `divide`

Even better, you can log input arguments as well:

@log.context.info(show_args=True, show_kwargs=True)
def greet(name, char='-'):
    msg = 'Hello, {name}'.format(name=name)
    print(msg)
    print(char * len(msg))

>>> greet('Tony', char='~')
INFO: Call `greet('Tony', char='~')`
Hello, Tony
~~~~~~~~~~~
INFO: Return from `greet`

There’s also a special context manager for suppressing errors and logging:

with log.and_suppress(ValueError, msg="It's ok, mistakes happen"):
    raise ValueError('Test error')

[ERROR] It's ok, mistakes happen
Traceback (most recent call last):
  File "/Users/tyu/code/logquacious/logquacious/log_manager.py", line 103, in and_suppress
    yield
  File "scripts/example.py", line 26, in <module>
    raise ValueError('Test error')
ValueError: Test error

Note the traceback above is logged, not streamed to stderr.

Configuration

The message templates used by LogManager.context can be configured to your liking by passing a context_templates argument to LogManager:

log = logquacious.LogManager(__name__, context_templates={
    'context.start': '=============== Enter {label} ===============',
    'context.finish': '=============== Exit {label} ===============',
})

with log.context.debug('greetings'):
    print('Hello!')

[DEBUG] =============== Enter greetings ===============
Hello!
[DEBUG] =============== Exit greetings ===============

The general format for context_templates keys is:

[CONTEXT_TYPE.]('start'|'finish')[.LOG_LEVEL_NAME]

where square-brackets marks optional fields.

CONTEXT_TYPE can be any of the following:

• function: Template used when called as a decorator.

• context: Template used when called as a context manager.

LOG_LEVEL_NAME can be any of the following logging levels:

• DEBUG

• INFO

• WARNING

• ERROR

• CRITICAL

For example, consider the cascade graph for function.start.DEBUG, which looks like:

     function.start.DEBUG
          /       \
start.DEBUG       function.start
          \       /
            start

The cascade is performed using a breadth-first search. If function.start.DEBUG is not defined, check start.DEBUG then check function.start BEFORE checking start.

The default configuration is:

DEFAULT_TEMPLATES = {
    'start': 'Enter {label}',
    'finish': 'Exit {label}',
    'function.start': 'Call `{label}({arguments})`',
    'function.finish': 'Return from `{label}`',
}

Note that custom configuration updates these defaults. For example, if you want to if you want to skip logging on exit for all context managers and decorators, you’ll have set both 'finish' and 'function.finish' to None or an empty string.

As you can see above, two template variables may be passed to the template string: label and arguments. When called as a context manager, the label is the first argument to the context manager and arguments is always empty. When called as a decorator, label is the function’s __name__ and arguments a string representing input arguments, if show_args or show_kwargs parameters are True.

Credits

This package was created with Cookiecutter and the audreyr/cookiecutter-pypackage project template.

0.5.0 (2019-05-05)

• Backport stacklevel keyword argument from Python 3.8 and configure stacklevel such that logging utilities report the context (e.g. filename and line number) where logquacious utilities are called.

0.4.0 (2018-10-05)

• Fix config override behavior to extend rather than replace default templates

0.3.0 (2018-10-05)

• Add decorator support for log.and_suppress and log.and_reraise context managers

• Suppress logging for null/empty log message templates

0.2.0 (2018-10-03)

Changed default templates. In 0.1.0, the templates were:

DEFAULT_TEMPLATES = {
    'start': 'Start {label}',
    'finish': 'Finish {label}',
}

These defaults have been changed to:

DEFAULT_TEMPLATES = {
    'start': 'Enter {label}',
    'finish': 'Exit {label}',
    'function.start': 'Call `{label}({arguments})`',
    'function.finish': 'Return from `{label}`',
}

0.1.0 (2018-10-03)

• First release on PyPI.