我是一名 Java 开发人员，已经过渡到 Python 3 开发。我需要在我的 Python 应用程序中配置日志记录，我想以类似于我在 Java 中配置 log4j 的方式来完成它。理想情况下，我希望我的 Python 日志记录配置文件采用 YAML 格式。我该怎么做？
我无法找到这个问题的全面答案。我最终弄明白了，并将在这里分享我的解决方案。
我的 YAML 格式的日志记录配置文件是这样的（我没有记录到文件，所以我把那部分注释掉了）：
---
version: 1
disable_existing_loggers: False
# -------------------------------------------------------------------------------------------------
# Define format of output logs (named 'simple').
formatters:
#simple:
# format: "XXX %(asctime)s - %(name)s - %(levelname)s - %(message)s"
console_formatter:
format: "%(asctime)s | %(levelname)8s | %(name)50s | %(funcName)50s | %(message)s"
# -------------------------------------------------------------------------------------------------
# REFERENCE: https://docs.python.org/3/library/logging.handlers.html
handlers:
## Create rotating file handler using 'simple' format.
#file_handler:
# class: logging.handlers.RotatingFileHandler
# level: INFO
# formatter: simple
# filename: jims-fantastic-log-file.log
# maxBytes: 10485760 # 10MB
# backupCount: 5
# encoding: utf8
# Create console handler
console_handler:
class: logging.StreamHandler
formatter: console_formatter
# -------------------------------------------------------------------------------------------------
# The 'root' logger (parent of all other loggers) is treated as a special case in the logging configuration.
# Any settings for the root logger become the default settings for all other loggers.
# You can override those defaults by providing a configuration for any loggers that need one.
root:
level: WARN
# handlers: [file_handler, console_handler]
handlers: [console_handler]
# Provider logger-specific configuration below.
# Must like log4j in Java, logger names are treated as a hierarchy. The developer will typically use their module
# names as logger names, so the top-level package is the 'parent' and all the subpackages are the 'children'.
loggers:
mypackage:
level: INFO
propagate: False
handlers: [console_handler]
mypackage.childpkg1:
level: WARNING
propagate: False
handlers: [console_handler]
mypackage.childpkg2:
level: DEBUG
propagate: False
handlers: [console_handler]
这是我为在我的应用程序中初始化日志记录而定义的微型类。它只是加载 YAML 文件：
import logging
import logging.config
import yaml
class LoggingUtils:
def __init__(self):
pass
@staticmethod
def configure_logging(path_to_logging_config_yaml_file: str):
# Initialize the logger once as the application starts up.
with open(path_to_logging_config_yaml_file, 'rt') as f:
config = yaml.safe_load(f.read())
logging.config.dictConfig(config)
# Get an instance of the logger and use it to write a log!
# Note: Do this AFTER the config is loaded above or it won't use the config.
logger = logging.getLogger(__name__)
logger.info("Configured the logger!")
参考我学这个的教程：