click-logging-config 1.1.0

Quick and easy logging parameters for click commands.

Quick and easy CLI logging options for click commands using a Python decorator.

I found myself implementing logging preferences repeatedly for utilities. Logging configuration is pretty simple, but for each new implementation I would find myself spending time researching the same options to refresh my memory and then implementing something slightly different than the last time. 🙄

click-logging-config is my attempt to stop the circle of re-implementation with settings that are useful enough, with configurability to change it if you don’t like the out-of-box behaviour. It’s proving to be pretty useful and I’m already using it across several of my other projects. 😄

It is released under the MIT license so you are free to use it in lots of different ways. As simple as it looks, a tool like this still represents research time and implementation effort, so please use the link below to help support development.

Support click-logging-config

Byting Chipmunk 🐿

Take a byte off.

Contents

• 1 Installation

• 2 Getting Started

• 3 Custom defaults

• 4 Console logging

• 5 File logging

1 Installation

The click-logging-config package is available from PyPI. Installing into a virtual environment is recommended.

python3 -m venv .venv; .venv/bin/pip install click-logging-config

2 Getting Started

Using click-logging-config is intended to be very simple. A single decorator applied to your click command or group adds some click options specifically for managing logging context.

import click
import logging
from click_logging import logging_parameters

log = logging.getLogger(__name__)

def do_something()
    pass

@click.command()
@click.option("--my-option", type=str)
# NOTE: Empty braces are required for hard-coded click-logging-config defaults.
@logging_parameters()
def my_command(my_option: str) -> None:
    log.info("doing something")
    try:
        do_something(my_option)
    except Exception as e:
        log.critical(f"something bad happened, {str(e)}")
        raise

Application of the @logging_parameters decorator must be applied immediately above your click command function and below any other click decorators such as arguments and options.

Having applied the decorator, your command now has the following options available to it.

--log-console-enable / --log-console-disable
                       Enable or disable console logging.
                       [default: log-console-disable]
--log-console-json-enable / --log-console-json-disable
                       Enable or disable console JSON logging.
                       [default: log-console-json-disable]
--log-file-enable / --log-file-disable
                       Enable or disable file logging.
                       [default: log-file-enable]
--log-file-json-enable / --log-file-json-disable
                       Enable or disable file JSON logging.
                       [default: log-file-json-enable]
--log-file FILE        The log file to write to.  [default: this.log]
--log-level [critical|error|warning|info|debug|notset]
                       Select logging level to apply to all enabled
                       log sinks.  [default: warning]

Note that the single log level configuration parameter applies to both console and file logging.

The internal defaults are configured for an interactive utility (run by a human in a terminal rather than via automation, or in a container). In summary,

• disabled console logging (allows your application to use console output, if needed)

• enabled file logging (1MB rotation size, with 10 rotation backups)

• “warning” log level

3 Custom defaults

If you don’t like the click-logging-config internal defaults for the options you can define your own. The LoggingConfiguration class is derived from pydantic.BaseModel, so one easy way to define your defaults is using a dictionary. You only need to define values you want to change - any other value will continue using the internal defaults.

import pathlib

import click
import logging
from click_logging import logging_parameters, LoggingConfiguration

log = logging.getLogger(__name__)

MY_LOGGING_DEFAULTS = LoggingConfiguration.parse_obj(
    {
        "file_logging": {
            # NOTE: file path must be specified using pathlib.Path
            "log_file_path": pathlib.Path("some_other.log"),
        },
        "log_level": "info",
    }
)

def do_something()
    pass

@click.command()
@click.option("--my-option", type=str)
@logging_parameters(MY_LOGGING_DEFAULTS)
def my_command(my_option: str) -> None:
    log.info("doing something")
    try:
        do_something(my_option)
    except Exception as e:
        log.critical(f"something bad happened, {str(e)}")
        raise

The table below summarizes the available settings for defaults. Otherwise review the LoggingConfiguration class definition .

Available top-level settings for logging defaults.

Setting	Type	Hard default	Description
log_level	str	warning	Define log level
enable_console_logging	boolean	False	Enable console logging
console_logging	dict		Console logging specific settings. See table below.
enable_file_logging	bool	True	Enable file logging
file_logging	dict		File logging specific settings. See table below.

Available console logging defaults.

Setting	Type	Hard default	Description
json_enabled	bool	False	Output JSON logs using json_log_formatter

Available file logging defaults.

Setting	Type	Hard default	Description
json_enabled	bool	True	Output JSON logs using json_log_formatter
log_file_path	pathlib.Path	./this.log	Path and name of log file.
file_rotation_size_megabytes	int	1	Maximum size of
max_rotation_backup_files	int	10	Maximum number of rotation backup files

4 Console logging

Console logging can be enabled or disabled, and there is an additional option to output line-by-line text based timestamped log entries, or JSON logging via the json_log_formatter framework. The format of text based log entries cannot be configured at this time and console logging is always emitted to stderr at this time.

5 File logging

File rotation on the file log is implemented as a “sensible default” - it cannot be disabled at this time, although you might be able to specify a maximum rotation of 1 to achieve the same end (not tested). The maximum rotation size can be specified as a configuration default. File logging itself can be enabled or disabled via defaults or the CLI options described above.

Similar to console logging the format can be as either text-based or JSON logging.