Installation

Via pip:

pip install django-modelhandler

Usage

Add modelhandler to your INSTALLED_APPS, then, configure log handler in your desired way. For example, using dictConfig:

{
    'version': 1,
    'handlers': {
        'modelhandler': {
            'class': 'modelhandler.handlers.LogModel',
            'level': 'ERROR'
        }
    },
    'loggers': {
        'my_logger': {
            'handlers': ['modelhandler'],
            'level': 'ERROR'
        }
    }
}

Run migrations that will create a Log model:

python manage.py migrate

And now you can start logging in django model.

Getting logs:

from modelhandler.models import Log
# Get the latest log
log = Log.objects.latest()
log.name # logger name
log.level # logging level integer
log.levelname # logging level as string (DEBUG, INFO, etc.)
log.message # the log message
log.traceback # traceback, if exists. default: None
log.filename # filename (with ext) where the log was sent
log.funcName # function name where the log was sent
log.created # log creation datetime
log.formatted # the log message as if it was written in file. (with [datetime] [level] etc.)

If you have a django admin enabled, then you can browse your logs on model Log of application modelhandler. It has a customized modeladmin to enabale filtering by logger name and levelname, and searching by message.

If you would like to customize a log model (to alter models Meta), then just subclass a modelhandler.models.Log model, do whatever you want and add your model to LogModel handler parameters:

'handlers': {
    'modelhandler': {
        'class': 'modelhandler.handlers.LogModel',
        'model': 'path.to.your.model'
        'level': 'ERROR'
    }
}

If you using celery in your project then you might want to add some model cleaning tasks in CELERYBEAT_SCHEDULE:

CELERYBEAT_SCHEDULE = {
    'cleanup_day': {
        'task': 'modelhandler.tasks.cleanup_day',
        'schedule': timedelta(days=1)
    }, # OR
    'cleanup_week': {
        'task': 'modelhandler.tasks.cleanup_week',
        'schedule': timedelta(days=7)
    }, # OR
    'cleanup_month': {
        'task': 'modelhandler.tasks.cleanup_month',
        'schedule': timedelta(days=30)
    }
}

modelhandler.tasks.cleanup_day will delete all logs that are older than one day from time of task execution. modelhandler.tasks.cleanup_week and modelhandler.tasks.cleanup_month are similar.

If you want to customize the time of deletion, there is a task modelhandler.tasks.cleanup_logs that accepts a before parameter that must be a datetime object or None (in this case a value of timezone.now() will be taken). There is no magic: just Log.objects.filter(created__lte=before).delete()