Customising default global logger (hashtag1090)

[vikramsubramanian]

As an example: I have several instances of the same script each running with its own config file that contains instance_id. The scripts use multiple classes spread across several modules. I would like to override the default logger adding custom fields (i.e. instance_id), setting opts, etc. when the script starts up, and be able to use this custom logger instance everywhere in the code, just like from loguru import logger. Anyone tried something like this before and succeeded? )

[mayil-ai[bot]]

Implementation ideas

To implement the feature described in the Jira ticket, you need to customize the default logger instance provided by Loguru to include additional context (e.g., instance_id) and settings (opts). Here's how you can achieve this:

• Create a function to configure the logger with the desired settings and context.
• Use the logger.configure() method to apply the settings globally.
• Use the logger.bind() method to add the instance_id to the extra context of the logger.
• Ensure that the customized logger is used across all modules by importing it after configuration.

Here's a high-level overview of the steps in code:

from loguru import logger

def configure_logger(instance_id):
    # Configure the logger with custom settings (opts)
    logger.configure(
        handlers=[
            dict(sink=sys.stderr, format="<green>{time}</green> <level>{message}</level>", level="INFO")
            # Add other handler configurations if needed
        ],
        levels=[
            dict(name="CUSTOM", no=25, color="<yellow>", icon="🔥"),
            # Add other custom levels if needed
        ],
        # Add other configurations if needed
    )

    # Bind 'instance_id' to the logger's extra context
    logger = logger.bind(instance_id=instance_id)

    return logger

# Usage in the main script or entry point of each instance
instance_specific_logger = configure_logger(instance_id="your_instance_id_here")

# Now, you can use 'instance_specific_logger' across your modules

In each module where logging is required, you would import instance_specific_logger and use it for logging:

from path.to.module.with.configured.logger import instance_specific_logger as logger

logger.info("This log will have the 'instance_id' attached.")

Remember to replace "your_instance_id_here" with the actual instance_id you retrieve from your config file for each script instance.

This approach ensures that all log messages across your modules will include the instance_id and any other custom fields you bind to the logger. The configure_logger function should be called once at the startup of each script instance to set up the logger appropriately.

Code snippets to check

loguru → __init__.py

This snippet initializes the default logger instance which is relevant to the issue as the user wants to customize the default logger with additional fields like 'instance_id'. https://github.com/Mayil-AI/loguru/blob/871de74acb2382fb3982d3fb8e224e4a82de9073/loguru/__init__.py#L1-L33

loguru → _logger.py

This snippet contains the 'add' method of the logger which is used to configure handlers. It is relevant because the user needs to understand how to add custom configurations to the logger. https://github.com/Mayil-AI/loguru/blob/871de74acb2382fb3982d3fb8e224e4a82de9073/loguru/_logger.py#L253-L2094 This snippet defines the Core class which holds the configuration for levels and handlers. It is relevant as the user may need to modify or extend the logger's core to add custom fields. https://github.com/Mayil-AI/loguru/blob/871de74acb2382fb3982d3fb8e224e4a82de9073/loguru/_logger.py#L125-L204

loguru → _handler.py

This snippet defines the Handler class which is responsible for handling log messages. Customizing the logger to add fields like 'instance_id' may involve creating a custom handler. https://github.com/Mayil-AI/loguru/blob/871de74acb2382fb3982d3fb8e224e4a82de9073/loguru/_handler.py#L32-L108