(可在我的博客阅读更多文章)
最近有个需求是把以前字符串输出的log 改为json 格式,看了别人的例子,还是有些比较茫然,索性就把logging 整个翻了一边,做点小总结.
初看log
在程序中, log 的用处写代码的你用你知道,log 有等级,DEBUG, INFO,...之类,还会记录时间,log 发生的位置,在Python 中用的多的就是logging 这个标准库中的包了.当打log 的时候究竟发生了什么? 是如何把不同级别的log 输出到不同文件里,还能在控制台输出.......
最简单的用法
import logging
logging.basicConfig(format='%(levelname)s:%(message)s', level=logging.DEBUG)
logging.debug('This message should go to the log file')
logging.info('So should this')
logging.warning('And this, too')
1,第一行导入包
2,第二行利用basicConfig 对输出的格式,和输出级别做了限制
3, 后面分别输出了三条不同级别的 log
Logging Levels
Level | Numeric value |
---|---|
CRITICAL | 50 |
ERROR | 40 |
WARNING | 30 |
INFO | 20 |
DEBUG | 10 |
NOTSET | 0 |
共有几个等级, 每个等级对应一个Int 型整数 ,每个等级都会有一个方法与之对应,这样输出的内容就有了不同的等级.
logger 流程,
整个过程,还是不是很详细,贴个图吧, 现在看还太早,也说不清真个过程到底发生了什么,先放着,回头来看会比较好懂.
读代码
代码结构
logging 在源码中有三个文件,结构如下:
├── config.py
├── handlers.py
└── __init__.py
_int.py中实现了基础功能,主要的逻辑就在这个文件中
handlers.py 是一些Handlers (用处后面会明白)用起来很方便的.
config.py 是对配置做处理的方法.
objects
LogRecord Objects
每一次log 都会实例化一个Record 对象,这个对象有很多属性,最后对LogRecord 做一下format 就输出了,格式化的log ,里面就基本就是这个对象的属性了。
class LogRecord(object):
def __init__(self, name, level, pathname, lineno,
msg, args, exc_info, func=None):
ct = time.time()
self.name = name
self.msg = msg
if (args and len(args) == 1 and isinstance(args[0], collections.Mapping)
and args[0]):
args = args[0]
self.args = args
self.levelname = getLevelName(level)
self.levelno = level
self.pathname = pathname
try:
self.filename = os.path.basename(pathname)
self.module = os.path.splitext(self.filename)[0]
except (TypeError, ValueError, AttributeError):
self.filename = pathname
self.module = "Unknown module"
self.exc_info = exc_info
self.exc_text = None # used to cache the traceback text
self.lineno = lineno
self.funcName = func
self.created = ct
self.msecs = (ct - long(ct)) * 1000
self.relativeCreated = (self.created - _startTime) * 1000
if logThreads and thread:
self.thread = thread.get_ident()
self.threadName = threading.current_thread().name
else:
self.thread = None
self.threadName = None
if not logMultiprocessing:
self.processName = None
else:
self.processName = 'MainProcess'
mp = sys.modules.get('multiprocessing')
if mp is not None:
try:
self.processName = mp.current_process().name
except StandardError:
pass
if logProcesses and hasattr(os, 'getpid'):
self.process = os.getpid()
else:
self.process = None
def __str__(self):
return '<LogRecord: %s, %s, %s, %s, "%s">'%(self.name, self.levelno,
self.pathname, self.lineno, self.msg)
def getMessage(self):
pass
看代码就发现, 这个类没做什么事情,就是一个model 而已, 有一个得到msg 的方法
Formatter Objects
Formatter 就是对Record 专门格式化的对象,它有一个format 方法,我们实现这个方法就能 做到不同的输出,我的需求是做json 格式的log 其实关键就在写一个Formatter 就好了
class Formatter(object):
converter = time.localtime
def __init__(self, fmt=None, datefmt=None):
if fmt:
self._fmt = fmt
else:
self._fmt = "%(message)s"
self.datefmt = datefmt
def formatTime(self, record, datefmt=None):
pass
def formatException(self, ei):
pass
def usesTime(self):
return self._fmt.find("%(asctime)") >= 0
def format(self, record):
pass
删掉源代码中的实现细节,这个类里面主要的是format 方法,这是默认最基本的Formater ,还有专门对exception ,时间做格式化的方法。具体是哪个,看方法名就很清楚了,具体每个方法怎么实现的,一眼也就懂了。fmt 是制定格式化的,具体怎么指定在最基础的用法中就有例子,datefmt 是对时间格式的指定。
Filter Objects
这个类是Logger 和Handler 的基类,主要有一个Filter 方法,和一个filters 属性
Handler Objects
叫Handler 的类还真的不少,在SocketServer 中也有看到,具体的功能都在Handler 中.在这里,组合所有的Formatter ,和控制log 的输出的方向,继承自Filter.
def __init__(self, level=NOTSET):
Filterer.__init__(self)
self._name = None
self.level = _checkLevel(level)
self.formatter = None
_addHandlerRef(self)
self.createLock()
在init方法中看到,Handler 也有一个属性,通过把自身的属性和LogRecord 的level对比来决定是否处理这个LogRecord 的。每个Handler 都有一个Formatter 属性,其实就是上面介绍的Formatter 。Handler 就是来控制LogRecord 和Formatter 的,它还可以控制输出的方式,在后面会有,StreamHandler,FileHandler等。通过名称也就能明白具体能干什么,这就是编程取名的智慧。
Logger Objects
这个类通常会通过getLogger()或者getLogger(name)来得到,不会直接new 一个出来.它会有*info(msg, args, kwargs) ,warn(msg, *args, kwargs)等方法,
def __init__(self, name, level=NOTSET):
Filterer.__init__(self)
self.name = name
self.level = _checkLevel(level)
self.parent = Noneou
self.handlers = []
self.disabled = 0
从init方法中能看到handlers 属性,这是一个list ,每个LogRecord 通过Handlers 不同的handlers 就能以不同的格式输出到不同的地方了。每个Logger 可以通过addHandler(hdlr)方法来添加各种Handler,
知道这些你就基本可以随意定制化了
下面就是我实现的json 格式的Formater,支持控制台颜色变化,当然前提是你的控制终端支持(Ubuntu14.04测试通过)
import re
import logging
import socket
import json
import traceback
import datetime
import time
try:
from collections import OrderedDict
except ImportError:
pass
RESERVED_ATTRS = (
'args', 'asctime', 'created', 'exc_info', 'exc_text', 'filename',
'funcName', 'levelname', 'levelno', 'lineno', 'module',
'msecs', 'message', 'msg', 'name', 'pathname', 'process',
'processName', 'relativeCreated', 'stack_info', 'thread', 'threadName')
RESERVED_ATTR_HASH = dict(zip(RESERVED_ATTRS, RESERVED_ATTRS))
COLORS ={
'HEADER' : '\\033[95m',
'INFO' : '\\033[94m',
'DEBUG' : '\\033[92m',
'WARNING' : '\\033[93m',
'ERROR' : '\\033[91m',
'ENDC' : '\\033[0m',
}
def merge_record_extra(record, target, reserved=RESERVED_ATTR_HASH):
for key, value in record.__dict__.items():
if (key not in reserved
and not (hasattr(key, "startswith")
and key.startswith('_'))):
target[key] = value
return target
def get_host_info():
host_name = ''
local_ip = ''
try:
host_name = socket.gethostname()
local_ip = socket.gethostbyname(host_name)
except Exception, e:
pass
return host_name, local_ip
class JsonFormatterBase(logging.Formatter):
def __init__(self, *args, **kwargs):
logging.Formatter.__init__(self, *args, **kwargs)
self._required_fields = self.parse()
self._skip_fields = dict(zip(self._required_fields,self._required_fields))
self._skip_fields.update(RESERVED_ATTR_HASH)
def parse(self):
standard_formatters = re.compile(r'\\((.+?)\\)', re.IGNORECASE)
return standard_formatters.findall(self._fmt)
def add_fields(self, record ):
log_record = {}
for field in self._required_fields:
log_record[field] = record.__dict__.get(field)
host_name , local_ip = get_host_info()
log_record[u'@hostName'] = host_name
log_record[u'@localIp'] = local_ip
return log_record
#merge_record_extra(record, log_record, reserved=self._skip_fields)
def process_log_record(self, log_record):
"""
Override this method to implement custom logic
on the possibly ordered dictionary.
"""
try:
new_record = OrderedDict()
except Exception, e:
return log_record
key_list = [
'asctime',
'levelname',
'@hostName',
'@localIp',
'threadName',
'thread',
'name',
'pathname',
'lineno',
'message',
]
for k in key_list:
new_record[k] = log_record.get(k)
new_record.update(log_record)
return new_record
def jsonify_log_record(self, log_record):
"""Returns a json string of the log record."""
return json.dumps(log_record, ensure_ascii=False)
def format_col(self, message_str, level_name):
"""
是否需要颜色
"""
return message_str
def formatTime(self, record, datefmt=None):
ct = self.converter(record.created)
if datefmt:
s = time.strftime(datefmt, ct)
else:
t = time.strftime("%Y-%m-%d %H:%M:%S", ct)
s = "%s.%03d" % (t, record.msecs)
return s
def format(self, record):
if isinstance(record.msg, dict):
record.message = record.msg
elif isinstance(record.msg, list) or isinstance(record.msg, tuple):
record.message = record.msg
elif isinstance(record.msg, basestring):
record.message = record.getMessage().split('\\n')
elif isinstance(record.msg, Exception):
record.message = traceback.format_exc(record.msg).split('\\n')
else :
record.message = repr(record.msg)
if "asctime" in self._required_fields:
record.asctime = self.formatTime(record, self.datefmt)
#
# if record.exc_info and not message_dict.get('exc_info'):
# message_dict['message'] = traceback.format_exception(*record.exc_info)
log_record = self.add_fields(record)
log_record = self.process_log_record(log_record)
message_str = self.jsonify_log_record(log_record)
message_str = self.format_col(message_str, level_name=record.levelname)
return message_str
class ConsoleFormater(JsonFormatterBase):
def __init__(self, *args, **kwargs):
JsonFormatterBase.__init__(self, *args, **kwargs)
def format_col(self, message_str, level_name):
if level_name in COLORS.keys():
message_str = COLORS.get(level_name) + message_str + COLORS.get('ENDC')
return message_str
def jsonify_log_record(self, log_record):
return json.dumps(log_record, ensure_ascii=False, indent=4)
class JsonFileFormater(JsonFormatterBase):
def __init__(self, *args, **kewars):
JsonFormatterBase.__init__(self, *args, **kewars)
def jsonify_log_record(self, log_record):
return json.dumps(log_record, ensure_ascii=False)
配置
很多时候我们并不是这样自己去实现一些Handler ,Formater ,之类的代码,用logging 提供的config 就能做到了,如何写config下面举个例子解释下,
SC_LOGGING_CONF = {
"version": 1,
"disable_existing_loggers": False,
"formatters": {
"simple": {
"format": "%(asctime)s [%(levelname)s] [%(threadName)s:%(thread)d] [%(name)s:%(lineno)d] - %(message)s"
}
},
"handlers": {
"console": {
"class": "logging.StreamHandler",
"level": "DEBUG",
"formatter": "simple",
"stream": "ext://sys.stdout"
},
"info_file_handler": {
"class": "logging.handlers.RotatingFileHandler",
"level": "INFO",
"formatter": "simple",
"filename": PATH + "info-" + date.today().isoformat() + ".log",
"maxBytes": 10485760,
"backupCount": 20,
"encoding": "utf8"
},
"error_file_handler": {
"class": "logging.handlers.RotatingFileHandler",
"level": "ERROR",
"formatter": "simple",
"filename": PATH + "errors-" + date.today().isoformat() + ".log",
"maxBytes": 10485760,
"backupCount": 20,
"encoding": "utf8"
}
},
"": {
"level": "INFO",
"handlers": ["console", "info_file_handler", "error_file_handler"]
}
}
}
首先定义了一个formater 叫simaple , 然后定义了三个Handler ,分别是输出到控制台,输出到文件和info,error的。
logging.config.dictConfig(CONFIG.SC_LOGGING_CONF)
通过这句就能让这些配置产生效果了,这也是config.py做的事情,不需要写很多代码也能定制个性化的log.。