no log tears

I personally found the built-in logging library to be a bit messy to use, so I created this library. Now, logging is easy, and there are no more tears!

Features

• easy to configure: set up logging via environment variables or a configuration file (thanks to pydantic-settings).
• extended base configuration:
  • predefined configuration: use built-in formatters and handlers or create your own.
  • customizable default configuration with environment variables: e.g. LOGGING__LEVEL=INFO and your process will log at the INFO level. Or LOGGING__FORMATTER=verbose to use the predefined verbose formatter.
  • capture python warnings: log warnings.warn messages to the logger.
• mixins: easily add logging to your classes.
  • LoggerMixin: adds _logger property to your instance and adds self to log records (instance hex id).
  • LoggableMixin: converts instance to string based on logging level of the class.
  • LogMixin: adds a _log to your class and allows you to bind additional context to loggers with _log_extra method.
• log context: easily bind additional context to loggers.
• traceback limit: Set the number of stack frames to display in log messages.

Dependencies

• python 3.9+

Optional:

• pydantic
• pydantic-settings

Installation

pip install no-log-tears

Usage

Simple logging (execute python script with LOGGING__LEVEL=INFO environment variable):

from no_log_tears import get_logger

log = get_logger(__name__)
log.info("Hello, world!")
# Output: `2025-02-19 20:42:28,928 INFO    __main__                                           Hello, world!`

Set up logging with a configuration file:

# logging.yaml
root:
  level: INFO
  handlers: [ console ]
handlers:
  console:
    formatter: custom
    stream: ext://sys.stdout
formatters:
  custom:
    (): no_log_tears.formatters.SoftFormatter
    fmt: '%(asctime)s %(levelname)s %(message)s %(my_field)s %(__other__)s'

# automatically loads logging configuration from `logging.yaml` on import
from no_log_tears import get_logger

log = get_logger(__name__)
log.info("Hello, world!", my_field="my_value", my_other_field="my_other_value")
# Output: `2025-02-19 20:42:28,928 INFO Hello, world! my_value {'my_other_field': 'my_other_value'}`

Use the LogMixin class to easily log messages with additional context:

from no_log_tears import LogMixin


class MyClass(LogMixin):
    def __init__(self, my_field: str) -> None:
        self.__my_field = my_field

    def do_stuff(self, foo: str) -> str:
        # bind `foo` value to the logger
        log = self._log(foo=foo)

        # log a message with `foo` value
        log.debug("Doing stuff ...")

        spam = foo + ":eggs"
        log = log(spam=spam)

        # log a message with `foo`, `spam` and `magic` values
        log.info("Stuff done!", magic="Yes!")

        return spam