• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1:mod:`logging` --- Logging facility for Python
2==============================================
3
4.. module:: logging
5   :synopsis: Flexible event logging system for applications.
6
7.. moduleauthor:: Vinay Sajip <vinay_sajip@red-dove.com>
8.. sectionauthor:: Vinay Sajip <vinay_sajip@red-dove.com>
9
10**Source code:** :source:`Lib/logging/__init__.py`
11
12.. index:: pair: Errors; logging
13
14.. sidebar:: Important
15
16   This page contains the API reference information. For tutorial
17   information and discussion of more advanced topics, see
18
19   * :ref:`Basic Tutorial <logging-basic-tutorial>`
20   * :ref:`Advanced Tutorial <logging-advanced-tutorial>`
21   * :ref:`Logging Cookbook <logging-cookbook>`
22
23--------------
24
25This module defines functions and classes which implement a flexible event
26logging system for applications and libraries.
27
28The key benefit of having the logging API provided by a standard library module
29is that all Python modules can participate in logging, so your application log
30can include your own messages integrated with messages from third-party
31modules.
32
33The module provides a lot of functionality and flexibility.  If you are
34unfamiliar with logging, the best way to get to grips with it is to see the
35tutorials (see the links on the right).
36
37The basic classes defined by the module, together with their functions, are
38listed below.
39
40* Loggers expose the interface that application code directly uses.
41* Handlers send the log records (created by loggers) to the appropriate
42  destination.
43* Filters provide a finer grained facility for determining which log records
44  to output.
45* Formatters specify the layout of log records in the final output.
46
47
48.. _logger:
49
50Logger Objects
51--------------
52
53Loggers have the following attributes and methods.  Note that Loggers should
54*NEVER* be instantiated directly, but always through the module-level function
55``logging.getLogger(name)``.  Multiple calls to :func:`getLogger` with the same
56name will always return a reference to the same Logger object.
57
58The ``name`` is potentially a period-separated hierarchical value, like
59``foo.bar.baz`` (though it could also be just plain ``foo``, for example).
60Loggers that are further down in the hierarchical list are children of loggers
61higher up in the list.  For example, given a logger with a name of ``foo``,
62loggers with names of ``foo.bar``, ``foo.bar.baz``, and ``foo.bam`` are all
63descendants of ``foo``.  The logger name hierarchy is analogous to the Python
64package hierarchy, and identical to it if you organise your loggers on a
65per-module basis using the recommended construction
66``logging.getLogger(__name__)``.  That's because in a module, ``__name__``
67is the module's name in the Python package namespace.
68
69
70.. class:: Logger
71
72   .. attribute:: Logger.propagate
73
74      If this attribute evaluates to true, events logged to this logger will be
75      passed to the handlers of higher level (ancestor) loggers, in addition to
76      any handlers attached to this logger. Messages are passed directly to the
77      ancestor loggers' handlers - neither the level nor filters of the ancestor
78      loggers in question are considered.
79
80      If this evaluates to false, logging messages are not passed to the handlers
81      of ancestor loggers.
82
83      The constructor sets this attribute to ``True``.
84
85      .. note:: If you attach a handler to a logger *and* one or more of its
86         ancestors, it may emit the same record multiple times. In general, you
87         should not need to attach a handler to more than one logger - if you just
88         attach it to the appropriate logger which is highest in the logger
89         hierarchy, then it will see all events logged by all descendant loggers,
90         provided that their propagate setting is left set to ``True``. A common
91         scenario is to attach handlers only to the root logger, and to let
92         propagation take care of the rest.
93
94   .. method:: Logger.setLevel(level)
95
96      Sets the threshold for this logger to *level*. Logging messages which are less
97      severe than *level* will be ignored; logging messages which have severity *level*
98      or higher will be emitted by whichever handler or handlers service this logger,
99      unless a handler's level has been set to a higher severity level than *level*.
100
101      When a logger is created, the level is set to :const:`NOTSET` (which causes
102      all messages to be processed when the logger is the root logger, or delegation
103      to the parent when the logger is a non-root logger). Note that the root logger
104      is created with level :const:`WARNING`.
105
106      The term 'delegation to the parent' means that if a logger has a level of
107      NOTSET, its chain of ancestor loggers is traversed until either an ancestor with
108      a level other than NOTSET is found, or the root is reached.
109
110      If an ancestor is found with a level other than NOTSET, then that ancestor's
111      level is treated as the effective level of the logger where the ancestor search
112      began, and is used to determine how a logging event is handled.
113
114      If the root is reached, and it has a level of NOTSET, then all messages will be
115      processed. Otherwise, the root's level will be used as the effective level.
116
117      See :ref:`levels` for a list of levels.
118
119      .. versionchanged:: 3.2
120         The *level* parameter now accepts a string representation of the
121         level such as 'INFO' as an alternative to the integer constants
122         such as :const:`INFO`. Note, however, that levels are internally stored
123         as integers, and methods such as e.g. :meth:`getEffectiveLevel` and
124         :meth:`isEnabledFor` will return/expect to be passed integers.
125
126
127   .. method:: Logger.isEnabledFor(level)
128
129      Indicates if a message of severity *level* would be processed by this logger.
130      This method checks first the module-level level set by
131      ``logging.disable(level)`` and then the logger's effective level as determined
132      by :meth:`getEffectiveLevel`.
133
134
135   .. method:: Logger.getEffectiveLevel()
136
137      Indicates the effective level for this logger. If a value other than
138      :const:`NOTSET` has been set using :meth:`setLevel`, it is returned. Otherwise,
139      the hierarchy is traversed towards the root until a value other than
140      :const:`NOTSET` is found, and that value is returned. The value returned is
141      an integer, typically one of :const:`logging.DEBUG`, :const:`logging.INFO`
142      etc.
143
144
145   .. method:: Logger.getChild(suffix)
146
147      Returns a logger which is a descendant to this logger, as determined by the suffix.
148      Thus, ``logging.getLogger('abc').getChild('def.ghi')`` would return the same
149      logger as would be returned by ``logging.getLogger('abc.def.ghi')``. This is a
150      convenience method, useful when the parent logger is named using e.g. ``__name__``
151      rather than a literal string.
152
153      .. versionadded:: 3.2
154
155
156   .. method:: Logger.debug(msg, *args, **kwargs)
157
158      Logs a message with level :const:`DEBUG` on this logger. The *msg* is the
159      message format string, and the *args* are the arguments which are merged into
160      *msg* using the string formatting operator. (Note that this means that you can
161      use keywords in the format string, together with a single dictionary argument.)
162      No % formatting operation is performed on *msg* when no *args* are supplied.
163
164      There are four keyword arguments in *kwargs* which are inspected:
165      *exc_info*, *stack_info*, *stacklevel* and *extra*.
166
167      If *exc_info* does not evaluate as false, it causes exception information to be
168      added to the logging message. If an exception tuple (in the format returned by
169      :func:`sys.exc_info`) or an exception instance is provided, it is used;
170      otherwise, :func:`sys.exc_info` is called to get the exception information.
171
172      The second optional keyword argument is *stack_info*, which defaults to
173      ``False``. If true, stack information is added to the logging
174      message, including the actual logging call. Note that this is not the same
175      stack information as that displayed through specifying *exc_info*: The
176      former is stack frames from the bottom of the stack up to the logging call
177      in the current thread, whereas the latter is information about stack frames
178      which have been unwound, following an exception, while searching for
179      exception handlers.
180
181      You can specify *stack_info* independently of *exc_info*, e.g. to just show
182      how you got to a certain point in your code, even when no exceptions were
183      raised. The stack frames are printed following a header line which says:
184
185      .. code-block:: none
186
187          Stack (most recent call last):
188
189      This mimics the ``Traceback (most recent call last):`` which is used when
190      displaying exception frames.
191
192      The third optional keyword argument is *stacklevel*, which defaults to ``1``.
193      If greater than 1, the corresponding number of stack frames are skipped
194      when computing the line number and function name set in the :class:`LogRecord`
195      created for the logging event. This can be used in logging helpers so that
196      the function name, filename and line number recorded are not the information
197      for the helper function/method, but rather its caller. The name of this
198      parameter mirrors the equivalent one in the :mod:`warnings` module.
199
200      The fourth keyword argument is *extra* which can be used to pass a
201      dictionary which is used to populate the __dict__ of the :class:`LogRecord`
202      created for the logging event with user-defined attributes. These custom
203      attributes can then be used as you like. For example, they could be
204      incorporated into logged messages. For example::
205
206         FORMAT = '%(asctime)-15s %(clientip)s %(user)-8s %(message)s'
207         logging.basicConfig(format=FORMAT)
208         d = {'clientip': '192.168.0.1', 'user': 'fbloggs'}
209         logger = logging.getLogger('tcpserver')
210         logger.warning('Protocol problem: %s', 'connection reset', extra=d)
211
212      would print something like
213
214      .. code-block:: none
215
216         2006-02-08 22:20:02,165 192.168.0.1 fbloggs  Protocol problem: connection reset
217
218      The keys in the dictionary passed in *extra* should not clash with the keys used
219      by the logging system. (See the :class:`Formatter` documentation for more
220      information on which keys are used by the logging system.)
221
222      If you choose to use these attributes in logged messages, you need to exercise
223      some care. In the above example, for instance, the :class:`Formatter` has been
224      set up with a format string which expects 'clientip' and 'user' in the attribute
225      dictionary of the :class:`LogRecord`. If these are missing, the message will
226      not be logged because a string formatting exception will occur. So in this case,
227      you always need to pass the *extra* dictionary with these keys.
228
229      While this might be annoying, this feature is intended for use in specialized
230      circumstances, such as multi-threaded servers where the same code executes in
231      many contexts, and interesting conditions which arise are dependent on this
232      context (such as remote client IP address and authenticated user name, in the
233      above example). In such circumstances, it is likely that specialized
234      :class:`Formatter`\ s would be used with particular :class:`Handler`\ s.
235
236      .. versionchanged:: 3.2
237         The *stack_info* parameter was added.
238
239      .. versionchanged:: 3.5
240         The *exc_info* parameter can now accept exception instances.
241
242      .. versionchanged:: 3.8
243         The *stacklevel* parameter was added.
244
245
246   .. method:: Logger.info(msg, *args, **kwargs)
247
248      Logs a message with level :const:`INFO` on this logger. The arguments are
249      interpreted as for :meth:`debug`.
250
251
252   .. method:: Logger.warning(msg, *args, **kwargs)
253
254      Logs a message with level :const:`WARNING` on this logger. The arguments are
255      interpreted as for :meth:`debug`.
256
257      .. note:: There is an obsolete method ``warn`` which is functionally
258         identical to ``warning``. As ``warn`` is deprecated, please do not use
259         it - use ``warning`` instead.
260
261   .. method:: Logger.error(msg, *args, **kwargs)
262
263      Logs a message with level :const:`ERROR` on this logger. The arguments are
264      interpreted as for :meth:`debug`.
265
266
267   .. method:: Logger.critical(msg, *args, **kwargs)
268
269      Logs a message with level :const:`CRITICAL` on this logger. The arguments are
270      interpreted as for :meth:`debug`.
271
272
273   .. method:: Logger.log(level, msg, *args, **kwargs)
274
275      Logs a message with integer level *level* on this logger. The other arguments are
276      interpreted as for :meth:`debug`.
277
278
279   .. method:: Logger.exception(msg, *args, **kwargs)
280
281      Logs a message with level :const:`ERROR` on this logger. The arguments are
282      interpreted as for :meth:`debug`. Exception info is added to the logging
283      message. This method should only be called from an exception handler.
284
285
286   .. method:: Logger.addFilter(filter)
287
288      Adds the specified filter *filter* to this logger.
289
290
291   .. method:: Logger.removeFilter(filter)
292
293      Removes the specified filter *filter* from this logger.
294
295
296   .. method:: Logger.filter(record)
297
298      Apply this logger's filters to the record and return ``True`` if the
299      record is to be processed. The filters are consulted in turn, until one of
300      them returns a false value. If none of them return a false value, the record
301      will be processed (passed to handlers). If one returns a false value, no
302      further processing of the record occurs.
303
304
305   .. method:: Logger.addHandler(hdlr)
306
307      Adds the specified handler *hdlr* to this logger.
308
309
310   .. method:: Logger.removeHandler(hdlr)
311
312      Removes the specified handler *hdlr* from this logger.
313
314
315   .. method:: Logger.findCaller(stack_info=False, stacklevel=1)
316
317      Finds the caller's source filename and line number. Returns the filename, line
318      number, function name and stack information as a 4-element tuple. The stack
319      information is returned as ``None`` unless *stack_info* is ``True``.
320
321      The *stacklevel* parameter is passed from code calling the :meth:`debug`
322      and other APIs. If greater than 1, the excess is used to skip stack frames
323      before determining the values to be returned. This will generally be useful
324      when calling logging APIs from helper/wrapper code, so that the information
325      in the event log refers not to the helper/wrapper code, but to the code that
326      calls it.
327
328
329   .. method:: Logger.handle(record)
330
331      Handles a record by passing it to all handlers associated with this logger and
332      its ancestors (until a false value of *propagate* is found). This method is used
333      for unpickled records received from a socket, as well as those created locally.
334      Logger-level filtering is applied using :meth:`~Logger.filter`.
335
336
337   .. method:: Logger.makeRecord(name, level, fn, lno, msg, args, exc_info, func=None, extra=None, sinfo=None)
338
339      This is a factory method which can be overridden in subclasses to create
340      specialized :class:`LogRecord` instances.
341
342   .. method:: Logger.hasHandlers()
343
344      Checks to see if this logger has any handlers configured. This is done by
345      looking for handlers in this logger and its parents in the logger hierarchy.
346      Returns ``True`` if a handler was found, else ``False``. The method stops searching
347      up the hierarchy whenever a logger with the 'propagate' attribute set to
348      false is found - that will be the last logger which is checked for the
349      existence of handlers.
350
351      .. versionadded:: 3.2
352
353   .. versionchanged:: 3.7
354      Loggers can now be pickled and unpickled.
355
356.. _levels:
357
358Logging Levels
359--------------
360
361The numeric values of logging levels are given in the following table. These are
362primarily of interest if you want to define your own levels, and need them to
363have specific values relative to the predefined levels. If you define a level
364with the same numeric value, it overwrites the predefined value; the predefined
365name is lost.
366
367+--------------+---------------+
368| Level        | Numeric value |
369+==============+===============+
370| ``CRITICAL`` | 50            |
371+--------------+---------------+
372| ``ERROR``    | 40            |
373+--------------+---------------+
374| ``WARNING``  | 30            |
375+--------------+---------------+
376| ``INFO``     | 20            |
377+--------------+---------------+
378| ``DEBUG``    | 10            |
379+--------------+---------------+
380| ``NOTSET``   | 0             |
381+--------------+---------------+
382
383
384.. _handler:
385
386Handler Objects
387---------------
388
389Handlers have the following attributes and methods. Note that :class:`Handler`
390is never instantiated directly; this class acts as a base for more useful
391subclasses. However, the :meth:`__init__` method in subclasses needs to call
392:meth:`Handler.__init__`.
393
394.. class:: Handler
395
396   .. method:: Handler.__init__(level=NOTSET)
397
398      Initializes the :class:`Handler` instance by setting its level, setting the list
399      of filters to the empty list and creating a lock (using :meth:`createLock`) for
400      serializing access to an I/O mechanism.
401
402
403   .. method:: Handler.createLock()
404
405      Initializes a thread lock which can be used to serialize access to underlying
406      I/O functionality which may not be threadsafe.
407
408
409   .. method:: Handler.acquire()
410
411      Acquires the thread lock created with :meth:`createLock`.
412
413
414   .. method:: Handler.release()
415
416      Releases the thread lock acquired with :meth:`acquire`.
417
418
419   .. method:: Handler.setLevel(level)
420
421      Sets the threshold for this handler to *level*. Logging messages which are
422      less severe than *level* will be ignored. When a handler is created, the
423      level is set to :const:`NOTSET` (which causes all messages to be
424      processed).
425
426      See :ref:`levels` for a list of levels.
427
428      .. versionchanged:: 3.2
429         The *level* parameter now accepts a string representation of the
430         level such as 'INFO' as an alternative to the integer constants
431         such as :const:`INFO`.
432
433
434   .. method:: Handler.setFormatter(fmt)
435
436      Sets the :class:`Formatter` for this handler to *fmt*.
437
438
439   .. method:: Handler.addFilter(filter)
440
441      Adds the specified filter *filter* to this handler.
442
443
444   .. method:: Handler.removeFilter(filter)
445
446      Removes the specified filter *filter* from this handler.
447
448
449   .. method:: Handler.filter(record)
450
451      Apply this handler's filters to the record and return ``True`` if the
452      record is to be processed. The filters are consulted in turn, until one of
453      them returns a false value. If none of them return a false value, the record
454      will be emitted. If one returns a false value, the handler will not emit the
455      record.
456
457
458   .. method:: Handler.flush()
459
460      Ensure all logging output has been flushed. This version does nothing and is
461      intended to be implemented by subclasses.
462
463
464   .. method:: Handler.close()
465
466      Tidy up any resources used by the handler. This version does no output but
467      removes the handler from an internal list of handlers which is closed when
468      :func:`shutdown` is called. Subclasses should ensure that this gets called
469      from overridden :meth:`close` methods.
470
471
472   .. method:: Handler.handle(record)
473
474      Conditionally emits the specified logging record, depending on filters which may
475      have been added to the handler. Wraps the actual emission of the record with
476      acquisition/release of the I/O thread lock.
477
478
479   .. method:: Handler.handleError(record)
480
481      This method should be called from handlers when an exception is encountered
482      during an :meth:`emit` call. If the module-level attribute
483      ``raiseExceptions`` is ``False``, exceptions get silently ignored. This is
484      what is mostly wanted for a logging system - most users will not care about
485      errors in the logging system, they are more interested in application
486      errors. You could, however, replace this with a custom handler if you wish.
487      The specified record is the one which was being processed when the exception
488      occurred. (The default value of ``raiseExceptions`` is ``True``, as that is
489      more useful during development).
490
491
492   .. method:: Handler.format(record)
493
494      Do formatting for a record - if a formatter is set, use it. Otherwise, use the
495      default formatter for the module.
496
497
498   .. method:: Handler.emit(record)
499
500      Do whatever it takes to actually log the specified logging record. This version
501      is intended to be implemented by subclasses and so raises a
502      :exc:`NotImplementedError`.
503
504For a list of handlers included as standard, see :mod:`logging.handlers`.
505
506.. _formatter-objects:
507
508Formatter Objects
509-----------------
510
511.. currentmodule:: logging
512
513:class:`Formatter` objects have the following attributes and methods. They are
514responsible for converting a :class:`LogRecord` to (usually) a string which can
515be interpreted by either a human or an external system. The base
516:class:`Formatter` allows a formatting string to be specified. If none is
517supplied, the default value of ``'%(message)s'`` is used, which just includes
518the message in the logging call. To have additional items of information in the
519formatted output (such as a timestamp), keep reading.
520
521A Formatter can be initialized with a format string which makes use of knowledge
522of the :class:`LogRecord` attributes - such as the default value mentioned above
523making use of the fact that the user's message and arguments are pre-formatted
524into a :class:`LogRecord`'s *message* attribute.  This format string contains
525standard Python %-style mapping keys. See section :ref:`old-string-formatting`
526for more information on string formatting.
527
528The useful mapping keys in a :class:`LogRecord` are given in the section on
529:ref:`logrecord-attributes`.
530
531
532.. class:: Formatter(fmt=None, datefmt=None, style='%', validate=True)
533
534   Returns a new instance of the :class:`Formatter` class.  The instance is
535   initialized with a format string for the message as a whole, as well as a
536   format string for the date/time portion of a message.  If no *fmt* is
537   specified, ``'%(message)s'`` is used.  If no *datefmt* is specified, a format
538   is used which is described in the :meth:`formatTime` documentation.
539
540   The *style* parameter can be one of '%', '{' or '$' and determines how
541   the format string will be merged with its data: using one of %-formatting,
542   :meth:`str.format` or :class:`string.Template`. This only applies to the
543   format string *fmt* (e.g. ``'%(message)s'`` or ``{message}``), not to the
544   actual log messages passed to ``Logger.debug`` etc; see
545   :ref:`formatting-styles` for more information on using {- and $-formatting
546   for log messages.
547
548   .. versionchanged:: 3.2
549      The *style* parameter was added.
550
551   .. versionchanged:: 3.8
552      The *validate* parameter was added. Incorrect or mismatched style and fmt
553      will raise a ``ValueError``.
554      For example: ``logging.Formatter('%(asctime)s - %(message)s', style='{')``.
555
556   .. method:: format(record)
557
558      The record's attribute dictionary is used as the operand to a string
559      formatting operation. Returns the resulting string. Before formatting the
560      dictionary, a couple of preparatory steps are carried out. The *message*
561      attribute of the record is computed using *msg* % *args*. If the
562      formatting string contains ``'(asctime)'``, :meth:`formatTime` is called
563      to format the event time. If there is exception information, it is
564      formatted using :meth:`formatException` and appended to the message. Note
565      that the formatted exception information is cached in attribute
566      *exc_text*. This is useful because the exception information can be
567      pickled and sent across the wire, but you should be careful if you have
568      more than one :class:`Formatter` subclass which customizes the formatting
569      of exception information. In this case, you will have to clear the cached
570      value after a formatter has done its formatting, so that the next
571      formatter to handle the event doesn't use the cached value but
572      recalculates it afresh.
573
574      If stack information is available, it's appended after the exception
575      information, using :meth:`formatStack` to transform it if necessary.
576
577
578   .. method:: formatTime(record, datefmt=None)
579
580      This method should be called from :meth:`format` by a formatter which
581      wants to make use of a formatted time. This method can be overridden in
582      formatters to provide for any specific requirement, but the basic behavior
583      is as follows: if *datefmt* (a string) is specified, it is used with
584      :func:`time.strftime` to format the creation time of the
585      record. Otherwise, the format '%Y-%m-%d %H:%M:%S,uuu' is used, where the
586      uuu part is a millisecond value and the other letters are as per the
587      :func:`time.strftime` documentation.  An example time in this format is
588      ``2003-01-23 00:29:50,411``.  The resulting string is returned.
589
590      This function uses a user-configurable function to convert the creation
591      time to a tuple. By default, :func:`time.localtime` is used; to change
592      this for a particular formatter instance, set the ``converter`` attribute
593      to a function with the same signature as :func:`time.localtime` or
594      :func:`time.gmtime`. To change it for all formatters, for example if you
595      want all logging times to be shown in GMT, set the ``converter``
596      attribute in the ``Formatter`` class.
597
598      .. versionchanged:: 3.3
599         Previously, the default format was hard-coded as in this example:
600         ``2010-09-06 22:38:15,292`` where the part before the comma is
601         handled by a strptime format string (``'%Y-%m-%d %H:%M:%S'``), and the
602         part after the comma is a millisecond value. Because strptime does not
603         have a format placeholder for milliseconds, the millisecond value is
604         appended using another format string, ``'%s,%03d'`` --- and both of these
605         format strings have been hardcoded into this method. With the change,
606         these strings are defined as class-level attributes which can be
607         overridden at the instance level when desired. The names of the
608         attributes are ``default_time_format`` (for the strptime format string)
609         and ``default_msec_format`` (for appending the millisecond value).
610
611      .. versionchanged:: 3.9
612         The ``default_msec_format`` can be ``None``.
613
614   .. method:: formatException(exc_info)
615
616      Formats the specified exception information (a standard exception tuple as
617      returned by :func:`sys.exc_info`) as a string. This default implementation
618      just uses :func:`traceback.print_exception`. The resulting string is
619      returned.
620
621   .. method:: formatStack(stack_info)
622
623      Formats the specified stack information (a string as returned by
624      :func:`traceback.print_stack`, but with the last newline removed) as a
625      string. This default implementation just returns the input value.
626
627.. _filter:
628
629Filter Objects
630--------------
631
632``Filters`` can be used by ``Handlers`` and ``Loggers`` for more sophisticated
633filtering than is provided by levels. The base filter class only allows events
634which are below a certain point in the logger hierarchy. For example, a filter
635initialized with 'A.B' will allow events logged by loggers 'A.B', 'A.B.C',
636'A.B.C.D', 'A.B.D' etc. but not 'A.BB', 'B.A.B' etc. If initialized with the
637empty string, all events are passed.
638
639
640.. class:: Filter(name='')
641
642   Returns an instance of the :class:`Filter` class. If *name* is specified, it
643   names a logger which, together with its children, will have its events allowed
644   through the filter. If *name* is the empty string, allows every event.
645
646
647   .. method:: filter(record)
648
649      Is the specified record to be logged? Returns zero for no, nonzero for
650      yes. If deemed appropriate, the record may be modified in-place by this
651      method.
652
653Note that filters attached to handlers are consulted before an event is
654emitted by the handler, whereas filters attached to loggers are consulted
655whenever an event is logged (using :meth:`debug`, :meth:`info`,
656etc.), before sending an event to handlers. This means that events which have
657been generated by descendant loggers will not be filtered by a logger's filter
658setting, unless the filter has also been applied to those descendant loggers.
659
660You don't actually need to subclass ``Filter``: you can pass any instance
661which has a ``filter`` method with the same semantics.
662
663.. versionchanged:: 3.2
664   You don't need to create specialized ``Filter`` classes, or use other
665   classes with a ``filter`` method: you can use a function (or other
666   callable) as a filter. The filtering logic will check to see if the filter
667   object has a ``filter`` attribute: if it does, it's assumed to be a
668   ``Filter`` and its :meth:`~Filter.filter` method is called. Otherwise, it's
669   assumed to be a callable and called with the record as the single
670   parameter. The returned value should conform to that returned by
671   :meth:`~Filter.filter`.
672
673Although filters are used primarily to filter records based on more
674sophisticated criteria than levels, they get to see every record which is
675processed by the handler or logger they're attached to: this can be useful if
676you want to do things like counting how many records were processed by a
677particular logger or handler, or adding, changing or removing attributes in
678the :class:`LogRecord` being processed. Obviously changing the LogRecord needs
679to be done with some care, but it does allow the injection of contextual
680information into logs (see :ref:`filters-contextual`).
681
682.. _log-record:
683
684LogRecord Objects
685-----------------
686
687:class:`LogRecord` instances are created automatically by the :class:`Logger`
688every time something is logged, and can be created manually via
689:func:`makeLogRecord` (for example, from a pickled event received over the
690wire).
691
692
693.. class:: LogRecord(name, level, pathname, lineno, msg, args, exc_info, func=None, sinfo=None)
694
695   Contains all the information pertinent to the event being logged.
696
697   The primary information is passed in :attr:`msg` and :attr:`args`, which
698   are combined using ``msg % args`` to create the :attr:`message` field of the
699   record.
700
701   :param name:  The name of the logger used to log the event represented by
702                 this LogRecord. Note that this name will always have this
703                 value, even though it may be emitted by a handler attached to
704                 a different (ancestor) logger.
705   :param level: The numeric level of the logging event (one of DEBUG, INFO etc.)
706                 Note that this is converted to *two* attributes of the LogRecord:
707                 ``levelno`` for the numeric value and ``levelname`` for the
708                 corresponding level name.
709   :param pathname: The full pathname of the source file where the logging call
710                    was made.
711   :param lineno: The line number in the source file where the logging call was
712                  made.
713   :param msg: The event description message, possibly a format string with
714               placeholders for variable data.
715   :param args: Variable data to merge into the *msg* argument to obtain the
716                event description.
717   :param exc_info: An exception tuple with the current exception information,
718                    or ``None`` if no exception information is available.
719   :param func: The name of the function or method from which the logging call
720                was invoked.
721   :param sinfo: A text string representing stack information from the base of
722                 the stack in the current thread, up to the logging call.
723
724   .. method:: getMessage()
725
726      Returns the message for this :class:`LogRecord` instance after merging any
727      user-supplied arguments with the message. If the user-supplied message
728      argument to the logging call is not a string, :func:`str` is called on it to
729      convert it to a string. This allows use of user-defined classes as
730      messages, whose ``__str__`` method can return the actual format string to
731      be used.
732
733   .. versionchanged:: 3.2
734      The creation of a :class:`LogRecord` has been made more configurable by
735      providing a factory which is used to create the record. The factory can be
736      set using :func:`getLogRecordFactory` and :func:`setLogRecordFactory`
737      (see this for the factory's signature).
738
739   This functionality can be used to inject your own values into a
740   :class:`LogRecord` at creation time. You can use the following pattern::
741
742      old_factory = logging.getLogRecordFactory()
743
744      def record_factory(*args, **kwargs):
745          record = old_factory(*args, **kwargs)
746          record.custom_attribute = 0xdecafbad
747          return record
748
749      logging.setLogRecordFactory(record_factory)
750
751   With this pattern, multiple factories could be chained, and as long
752   as they don't overwrite each other's attributes or unintentionally
753   overwrite the standard attributes listed above, there should be no
754   surprises.
755
756
757.. _logrecord-attributes:
758
759LogRecord attributes
760--------------------
761
762The LogRecord has a number of attributes, most of which are derived from the
763parameters to the constructor. (Note that the names do not always correspond
764exactly between the LogRecord constructor parameters and the LogRecord
765attributes.) These attributes can be used to merge data from the record into
766the format string. The following table lists (in alphabetical order) the
767attribute names, their meanings and the corresponding placeholder in a %-style
768format string.
769
770If you are using {}-formatting (:func:`str.format`), you can use
771``{attrname}`` as the placeholder in the format string. If you are using
772$-formatting (:class:`string.Template`), use the form ``${attrname}``. In
773both cases, of course, replace ``attrname`` with the actual attribute name
774you want to use.
775
776In the case of {}-formatting, you can specify formatting flags by placing them
777after the attribute name, separated from it with a colon. For example: a
778placeholder of ``{msecs:03d}`` would format a millisecond value of ``4`` as
779``004``. Refer to the :meth:`str.format` documentation for full details on
780the options available to you.
781
782+----------------+-------------------------+-----------------------------------------------+
783| Attribute name | Format                  | Description                                   |
784+================+=========================+===============================================+
785| args           | You shouldn't need to   | The tuple of arguments merged into ``msg`` to |
786|                | format this yourself.   | produce ``message``, or a dict whose values   |
787|                |                         | are used for the merge (when there is only one|
788|                |                         | argument, and it is a dictionary).            |
789+----------------+-------------------------+-----------------------------------------------+
790| asctime        | ``%(asctime)s``         | Human-readable time when the                  |
791|                |                         | :class:`LogRecord` was created.  By default   |
792|                |                         | this is of the form '2003-07-08 16:49:45,896' |
793|                |                         | (the numbers after the comma are millisecond  |
794|                |                         | portion of the time).                         |
795+----------------+-------------------------+-----------------------------------------------+
796| created        | ``%(created)f``         | Time when the :class:`LogRecord` was created  |
797|                |                         | (as returned by :func:`time.time`).           |
798+----------------+-------------------------+-----------------------------------------------+
799| exc_info       | You shouldn't need to   | Exception tuple (à la ``sys.exc_info``) or,   |
800|                | format this yourself.   | if no exception has occurred, ``None``.       |
801+----------------+-------------------------+-----------------------------------------------+
802| filename       | ``%(filename)s``        | Filename portion of ``pathname``.             |
803+----------------+-------------------------+-----------------------------------------------+
804| funcName       | ``%(funcName)s``        | Name of function containing the logging call. |
805+----------------+-------------------------+-----------------------------------------------+
806| levelname      | ``%(levelname)s``       | Text logging level for the message            |
807|                |                         | (``'DEBUG'``, ``'INFO'``, ``'WARNING'``,      |
808|                |                         | ``'ERROR'``, ``'CRITICAL'``).                 |
809+----------------+-------------------------+-----------------------------------------------+
810| levelno        | ``%(levelno)s``         | Numeric logging level for the message         |
811|                |                         | (:const:`DEBUG`, :const:`INFO`,               |
812|                |                         | :const:`WARNING`, :const:`ERROR`,             |
813|                |                         | :const:`CRITICAL`).                           |
814+----------------+-------------------------+-----------------------------------------------+
815| lineno         | ``%(lineno)d``          | Source line number where the logging call was |
816|                |                         | issued (if available).                        |
817+----------------+-------------------------+-----------------------------------------------+
818| message        | ``%(message)s``         | The logged message, computed as ``msg %       |
819|                |                         | args``. This is set when                      |
820|                |                         | :meth:`Formatter.format` is invoked.          |
821+----------------+-------------------------+-----------------------------------------------+
822| module         | ``%(module)s``          | Module (name portion of ``filename``).        |
823+----------------+-------------------------+-----------------------------------------------+
824| msecs          | ``%(msecs)d``           | Millisecond portion of the time when the      |
825|                |                         | :class:`LogRecord` was created.               |
826+----------------+-------------------------+-----------------------------------------------+
827| msg            | You shouldn't need to   | The format string passed in the original      |
828|                | format this yourself.   | logging call. Merged with ``args`` to         |
829|                |                         | produce ``message``, or an arbitrary object   |
830|                |                         | (see :ref:`arbitrary-object-messages`).       |
831+----------------+-------------------------+-----------------------------------------------+
832| name           | ``%(name)s``            | Name of the logger used to log the call.      |
833+----------------+-------------------------+-----------------------------------------------+
834| pathname       | ``%(pathname)s``        | Full pathname of the source file where the    |
835|                |                         | logging call was issued (if available).       |
836+----------------+-------------------------+-----------------------------------------------+
837| process        | ``%(process)d``         | Process ID (if available).                    |
838+----------------+-------------------------+-----------------------------------------------+
839| processName    | ``%(processName)s``     | Process name (if available).                  |
840+----------------+-------------------------+-----------------------------------------------+
841| relativeCreated| ``%(relativeCreated)d`` | Time in milliseconds when the LogRecord was   |
842|                |                         | created, relative to the time the logging     |
843|                |                         | module was loaded.                            |
844+----------------+-------------------------+-----------------------------------------------+
845| stack_info     | You shouldn't need to   | Stack frame information (where available)     |
846|                | format this yourself.   | from the bottom of the stack in the current   |
847|                |                         | thread, up to and including the stack frame   |
848|                |                         | of the logging call which resulted in the     |
849|                |                         | creation of this record.                      |
850+----------------+-------------------------+-----------------------------------------------+
851| thread         | ``%(thread)d``          | Thread ID (if available).                     |
852+----------------+-------------------------+-----------------------------------------------+
853| threadName     | ``%(threadName)s``      | Thread name (if available).                   |
854+----------------+-------------------------+-----------------------------------------------+
855
856.. versionchanged:: 3.1
857   *processName* was added.
858
859
860.. _logger-adapter:
861
862LoggerAdapter Objects
863---------------------
864
865:class:`LoggerAdapter` instances are used to conveniently pass contextual
866information into logging calls. For a usage example, see the section on
867:ref:`adding contextual information to your logging output <context-info>`.
868
869.. class:: LoggerAdapter(logger, extra)
870
871   Returns an instance of :class:`LoggerAdapter` initialized with an
872   underlying :class:`Logger` instance and a dict-like object.
873
874   .. method:: process(msg, kwargs)
875
876      Modifies the message and/or keyword arguments passed to a logging call in
877      order to insert contextual information. This implementation takes the object
878      passed as *extra* to the constructor and adds it to *kwargs* using key
879      'extra'. The return value is a (*msg*, *kwargs*) tuple which has the
880      (possibly modified) versions of the arguments passed in.
881
882In addition to the above, :class:`LoggerAdapter` supports the following
883methods of :class:`Logger`: :meth:`~Logger.debug`, :meth:`~Logger.info`,
884:meth:`~Logger.warning`, :meth:`~Logger.error`, :meth:`~Logger.exception`,
885:meth:`~Logger.critical`, :meth:`~Logger.log`, :meth:`~Logger.isEnabledFor`,
886:meth:`~Logger.getEffectiveLevel`, :meth:`~Logger.setLevel` and
887:meth:`~Logger.hasHandlers`. These methods have the same signatures as their
888counterparts in :class:`Logger`, so you can use the two types of instances
889interchangeably.
890
891.. versionchanged:: 3.2
892   The :meth:`~Logger.isEnabledFor`, :meth:`~Logger.getEffectiveLevel`,
893   :meth:`~Logger.setLevel` and :meth:`~Logger.hasHandlers` methods were added
894   to :class:`LoggerAdapter`.  These methods delegate to the underlying logger.
895
896
897Thread Safety
898-------------
899
900The logging module is intended to be thread-safe without any special work
901needing to be done by its clients. It achieves this though using threading
902locks; there is one lock to serialize access to the module's shared data, and
903each handler also creates a lock to serialize access to its underlying I/O.
904
905If you are implementing asynchronous signal handlers using the :mod:`signal`
906module, you may not be able to use logging from within such handlers. This is
907because lock implementations in the :mod:`threading` module are not always
908re-entrant, and so cannot be invoked from such signal handlers.
909
910
911Module-Level Functions
912----------------------
913
914In addition to the classes described above, there are a number of module-level
915functions.
916
917
918.. function:: getLogger(name=None)
919
920   Return a logger with the specified name or, if name is ``None``, return a
921   logger which is the root logger of the hierarchy. If specified, the name is
922   typically a dot-separated hierarchical name like *'a'*, *'a.b'* or *'a.b.c.d'*.
923   Choice of these names is entirely up to the developer who is using logging.
924
925   All calls to this function with a given name return the same logger instance.
926   This means that logger instances never need to be passed between different parts
927   of an application.
928
929
930.. function:: getLoggerClass()
931
932   Return either the standard :class:`Logger` class, or the last class passed to
933   :func:`setLoggerClass`. This function may be called from within a new class
934   definition, to ensure that installing a customized :class:`Logger` class will
935   not undo customizations already applied by other code. For example::
936
937      class MyLogger(logging.getLoggerClass()):
938          # ... override behaviour here
939
940
941.. function:: getLogRecordFactory()
942
943   Return a callable which is used to create a :class:`LogRecord`.
944
945   .. versionadded:: 3.2
946      This function has been provided, along with :func:`setLogRecordFactory`,
947      to allow developers more control over how the :class:`LogRecord`
948      representing a logging event is constructed.
949
950   See :func:`setLogRecordFactory` for more information about the how the
951   factory is called.
952
953.. function:: debug(msg, *args, **kwargs)
954
955   Logs a message with level :const:`DEBUG` on the root logger. The *msg* is the
956   message format string, and the *args* are the arguments which are merged into
957   *msg* using the string formatting operator. (Note that this means that you can
958   use keywords in the format string, together with a single dictionary argument.)
959
960   There are three keyword arguments in *kwargs* which are inspected: *exc_info*
961   which, if it does not evaluate as false, causes exception information to be
962   added to the logging message. If an exception tuple (in the format returned by
963   :func:`sys.exc_info`) or an exception instance is provided, it is used;
964   otherwise, :func:`sys.exc_info` is called to get the exception information.
965
966   The second optional keyword argument is *stack_info*, which defaults to
967   ``False``. If true, stack information is added to the logging
968   message, including the actual logging call. Note that this is not the same
969   stack information as that displayed through specifying *exc_info*: The
970   former is stack frames from the bottom of the stack up to the logging call
971   in the current thread, whereas the latter is information about stack frames
972   which have been unwound, following an exception, while searching for
973   exception handlers.
974
975   You can specify *stack_info* independently of *exc_info*, e.g. to just show
976   how you got to a certain point in your code, even when no exceptions were
977   raised. The stack frames are printed following a header line which says:
978
979   .. code-block:: none
980
981       Stack (most recent call last):
982
983   This mimics the ``Traceback (most recent call last):`` which is used when
984   displaying exception frames.
985
986   The third optional keyword argument is *extra* which can be used to pass a
987   dictionary which is used to populate the __dict__ of the LogRecord created for
988   the logging event with user-defined attributes. These custom attributes can then
989   be used as you like. For example, they could be incorporated into logged
990   messages. For example::
991
992      FORMAT = '%(asctime)-15s %(clientip)s %(user)-8s %(message)s'
993      logging.basicConfig(format=FORMAT)
994      d = {'clientip': '192.168.0.1', 'user': 'fbloggs'}
995      logging.warning('Protocol problem: %s', 'connection reset', extra=d)
996
997   would print something like:
998
999   .. code-block:: none
1000
1001      2006-02-08 22:20:02,165 192.168.0.1 fbloggs  Protocol problem: connection reset
1002
1003   The keys in the dictionary passed in *extra* should not clash with the keys used
1004   by the logging system. (See the :class:`Formatter` documentation for more
1005   information on which keys are used by the logging system.)
1006
1007   If you choose to use these attributes in logged messages, you need to exercise
1008   some care. In the above example, for instance, the :class:`Formatter` has been
1009   set up with a format string which expects 'clientip' and 'user' in the attribute
1010   dictionary of the LogRecord. If these are missing, the message will not be
1011   logged because a string formatting exception will occur. So in this case, you
1012   always need to pass the *extra* dictionary with these keys.
1013
1014   While this might be annoying, this feature is intended for use in specialized
1015   circumstances, such as multi-threaded servers where the same code executes in
1016   many contexts, and interesting conditions which arise are dependent on this
1017   context (such as remote client IP address and authenticated user name, in the
1018   above example). In such circumstances, it is likely that specialized
1019   :class:`Formatter`\ s would be used with particular :class:`Handler`\ s.
1020
1021   .. versionchanged:: 3.2
1022      The *stack_info* parameter was added.
1023
1024.. function:: info(msg, *args, **kwargs)
1025
1026   Logs a message with level :const:`INFO` on the root logger. The arguments are
1027   interpreted as for :func:`debug`.
1028
1029
1030.. function:: warning(msg, *args, **kwargs)
1031
1032   Logs a message with level :const:`WARNING` on the root logger. The arguments
1033   are interpreted as for :func:`debug`.
1034
1035   .. note:: There is an obsolete function ``warn`` which is functionally
1036      identical to ``warning``. As ``warn`` is deprecated, please do not use
1037      it - use ``warning`` instead.
1038
1039
1040.. function:: error(msg, *args, **kwargs)
1041
1042   Logs a message with level :const:`ERROR` on the root logger. The arguments are
1043   interpreted as for :func:`debug`.
1044
1045
1046.. function:: critical(msg, *args, **kwargs)
1047
1048   Logs a message with level :const:`CRITICAL` on the root logger. The arguments
1049   are interpreted as for :func:`debug`.
1050
1051
1052.. function:: exception(msg, *args, **kwargs)
1053
1054   Logs a message with level :const:`ERROR` on the root logger. The arguments are
1055   interpreted as for :func:`debug`. Exception info is added to the logging
1056   message. This function should only be called from an exception handler.
1057
1058.. function:: log(level, msg, *args, **kwargs)
1059
1060   Logs a message with level *level* on the root logger. The other arguments are
1061   interpreted as for :func:`debug`.
1062
1063   .. note:: The above module-level convenience functions, which delegate to the
1064      root logger, call :func:`basicConfig` to ensure that at least one handler
1065      is available. Because of this, they should *not* be used in threads,
1066      in versions of Python earlier than 2.7.1 and 3.2, unless at least one
1067      handler has been added to the root logger *before* the threads are
1068      started. In earlier versions of Python, due to a thread safety shortcoming
1069      in :func:`basicConfig`, this can (under rare circumstances) lead to
1070      handlers being added multiple times to the root logger, which can in turn
1071      lead to multiple messages for the same event.
1072
1073.. function:: disable(level=CRITICAL)
1074
1075   Provides an overriding level *level* for all loggers which takes precedence over
1076   the logger's own level. When the need arises to temporarily throttle logging
1077   output down across the whole application, this function can be useful. Its
1078   effect is to disable all logging calls of severity *level* and below, so that
1079   if you call it with a value of INFO, then all INFO and DEBUG events would be
1080   discarded, whereas those of severity WARNING and above would be processed
1081   according to the logger's effective level. If
1082   ``logging.disable(logging.NOTSET)`` is called, it effectively removes this
1083   overriding level, so that logging output again depends on the effective
1084   levels of individual loggers.
1085
1086   Note that if you have defined any custom logging level higher than
1087   ``CRITICAL`` (this is not recommended), you won't be able to rely on the
1088   default value for the *level* parameter, but will have to explicitly supply a
1089   suitable value.
1090
1091   .. versionchanged:: 3.7
1092      The *level* parameter was defaulted to level ``CRITICAL``. See
1093      :issue:`28524` for more information about this change.
1094
1095.. function:: addLevelName(level, levelName)
1096
1097   Associates level *level* with text *levelName* in an internal dictionary, which is
1098   used to map numeric levels to a textual representation, for example when a
1099   :class:`Formatter` formats a message. This function can also be used to define
1100   your own levels. The only constraints are that all levels used must be
1101   registered using this function, levels should be positive integers and they
1102   should increase in increasing order of severity.
1103
1104   .. note:: If you are thinking of defining your own levels, please see the
1105      section on :ref:`custom-levels`.
1106
1107.. function:: getLevelName(level)
1108
1109   Returns the textual representation of logging level *level*. If the level is one
1110   of the predefined levels :const:`CRITICAL`, :const:`ERROR`, :const:`WARNING`,
1111   :const:`INFO` or :const:`DEBUG` then you get the corresponding string. If you
1112   have associated levels with names using :func:`addLevelName` then the name you
1113   have associated with *level* is returned. If a numeric value corresponding to one
1114   of the defined levels is passed in, the corresponding string representation is
1115   returned. Otherwise, the string 'Level %s' % level is returned.
1116
1117   .. note:: Levels are internally integers (as they need to be compared in the
1118      logging logic). This function is used to convert between an integer level
1119      and the level name displayed in the formatted log output by means of the
1120      ``%(levelname)s`` format specifier (see :ref:`logrecord-attributes`).
1121
1122   .. versionchanged:: 3.4
1123      In Python versions earlier than 3.4, this function could also be passed a
1124      text level, and would return the corresponding numeric value of the level.
1125      This undocumented behaviour was considered a mistake, and was removed in
1126      Python 3.4, but reinstated in 3.4.2 due to retain backward compatibility.
1127
1128.. function:: makeLogRecord(attrdict)
1129
1130   Creates and returns a new :class:`LogRecord` instance whose attributes are
1131   defined by *attrdict*. This function is useful for taking a pickled
1132   :class:`LogRecord` attribute dictionary, sent over a socket, and reconstituting
1133   it as a :class:`LogRecord` instance at the receiving end.
1134
1135
1136.. function:: basicConfig(**kwargs)
1137
1138   Does basic configuration for the logging system by creating a
1139   :class:`StreamHandler` with a default :class:`Formatter` and adding it to the
1140   root logger. The functions :func:`debug`, :func:`info`, :func:`warning`,
1141   :func:`error` and :func:`critical` will call :func:`basicConfig` automatically
1142   if no handlers are defined for the root logger.
1143
1144   This function does nothing if the root logger already has handlers
1145   configured, unless the keyword argument *force* is set to ``True``.
1146
1147   .. note:: This function should be called from the main thread
1148      before other threads are started. In versions of Python prior to
1149      2.7.1 and 3.2, if this function is called from multiple threads,
1150      it is possible (in rare circumstances) that a handler will be added
1151      to the root logger more than once, leading to unexpected results
1152      such as messages being duplicated in the log.
1153
1154   The following keyword arguments are supported.
1155
1156   .. tabularcolumns:: |l|L|
1157
1158   +--------------+---------------------------------------------+
1159   | Format       | Description                                 |
1160   +==============+=============================================+
1161   | *filename*   | Specifies that a FileHandler be created,    |
1162   |              | using the specified filename, rather than a |
1163   |              | StreamHandler.                              |
1164   +--------------+---------------------------------------------+
1165   | *filemode*   | If *filename* is specified, open the file   |
1166   |              | in this :ref:`mode <filemodes>`. Defaults   |
1167   |              | to ``'a'``.                                 |
1168   +--------------+---------------------------------------------+
1169   | *format*     | Use the specified format string for the     |
1170   |              | handler.                                    |
1171   +--------------+---------------------------------------------+
1172   | *datefmt*    | Use the specified date/time format, as      |
1173   |              | accepted by :func:`time.strftime`.          |
1174   +--------------+---------------------------------------------+
1175   | *style*      | If *format* is specified, use this style    |
1176   |              | for the format string. One of ``'%'``,      |
1177   |              | ``'{'`` or ``'$'`` for :ref:`printf-style   |
1178   |              | <old-string-formatting>`,                   |
1179   |              | :meth:`str.format` or                       |
1180   |              | :class:`string.Template` respectively.      |
1181   |              | Defaults to ``'%'``.                        |
1182   +--------------+---------------------------------------------+
1183   | *level*      | Set the root logger level to the specified  |
1184   |              | :ref:`level <levels>`.                      |
1185   +--------------+---------------------------------------------+
1186   | *stream*     | Use the specified stream to initialize the  |
1187   |              | StreamHandler. Note that this argument is   |
1188   |              | incompatible with *filename* - if both      |
1189   |              | are present, a ``ValueError`` is raised.    |
1190   +--------------+---------------------------------------------+
1191   | *handlers*   | If specified, this should be an iterable of |
1192   |              | already created handlers to add to the root |
1193   |              | logger. Any handlers which don't already    |
1194   |              | have a formatter set will be assigned the   |
1195   |              | default formatter created in this function. |
1196   |              | Note that this argument is incompatible     |
1197   |              | with *filename* or *stream* - if both       |
1198   |              | are present, a ``ValueError`` is raised.    |
1199   +--------------+---------------------------------------------+
1200   | *force*      | If this keyword argument is specified as    |
1201   |              | true, any existing handlers attached to the |
1202   |              | root logger are removed and closed, before  |
1203   |              | carrying out the configuration as specified |
1204   |              | by the other arguments.                     |
1205   +--------------+---------------------------------------------+
1206   | *encoding*   | If this keyword argument is specified along |
1207   |              | with *filename*, its value is used when the |
1208   |              | FileHandler is created, and thus used when  |
1209   |              | opening the output file.                    |
1210   +--------------+---------------------------------------------+
1211   | *errors*     | If this keyword argument is specified along |
1212   |              | with *filename*, its value is used when the |
1213   |              | FileHandler is created, and thus used when  |
1214   |              | opening the output file. If not specified,  |
1215   |              | the value 'backslashreplace' is used. Note  |
1216   |              | that if ``None`` is specified, it will be   |
1217   |              | passed as such to func:`open`, which means  |
1218   |              | that it will be treated the same as passing |
1219   |              | 'errors'.                                   |
1220   +--------------+---------------------------------------------+
1221
1222   .. versionchanged:: 3.2
1223      The *style* argument was added.
1224
1225   .. versionchanged:: 3.3
1226      The *handlers* argument was added. Additional checks were added to
1227      catch situations where incompatible arguments are specified (e.g.
1228      *handlers* together with *stream* or *filename*, or *stream*
1229      together with *filename*).
1230
1231   .. versionchanged:: 3.8
1232      The *force* argument was added.
1233
1234   .. versionchanged:: 3.9
1235      The *encoding* and *errors* arguments were added.
1236
1237.. function:: shutdown()
1238
1239   Informs the logging system to perform an orderly shutdown by flushing and
1240   closing all handlers. This should be called at application exit and no
1241   further use of the logging system should be made after this call.
1242
1243   When the logging module is imported, it registers this function as an exit
1244   handler (see :mod:`atexit`), so normally there's no need to do that
1245   manually.
1246
1247
1248.. function:: setLoggerClass(klass)
1249
1250   Tells the logging system to use the class *klass* when instantiating a logger.
1251   The class should define :meth:`__init__` such that only a name argument is
1252   required, and the :meth:`__init__` should call :meth:`Logger.__init__`. This
1253   function is typically called before any loggers are instantiated by applications
1254   which need to use custom logger behavior. After this call, as at any other
1255   time, do not instantiate loggers directly using the subclass: continue to use
1256   the :func:`logging.getLogger` API to get your loggers.
1257
1258
1259.. function:: setLogRecordFactory(factory)
1260
1261   Set a callable which is used to create a :class:`LogRecord`.
1262
1263   :param factory: The factory callable to be used to instantiate a log record.
1264
1265   .. versionadded:: 3.2
1266      This function has been provided, along with :func:`getLogRecordFactory`, to
1267      allow developers more control over how the :class:`LogRecord` representing
1268      a logging event is constructed.
1269
1270   The factory has the following signature:
1271
1272   ``factory(name, level, fn, lno, msg, args, exc_info, func=None, sinfo=None, **kwargs)``
1273
1274      :name: The logger name.
1275      :level: The logging level (numeric).
1276      :fn: The full pathname of the file where the logging call was made.
1277      :lno: The line number in the file where the logging call was made.
1278      :msg: The logging message.
1279      :args: The arguments for the logging message.
1280      :exc_info: An exception tuple, or ``None``.
1281      :func: The name of the function or method which invoked the logging
1282             call.
1283      :sinfo: A stack traceback such as is provided by
1284              :func:`traceback.print_stack`, showing the call hierarchy.
1285      :kwargs: Additional keyword arguments.
1286
1287
1288Module-Level Attributes
1289-----------------------
1290
1291.. attribute:: lastResort
1292
1293   A "handler of last resort" is available through this attribute. This
1294   is a :class:`StreamHandler` writing to ``sys.stderr`` with a level of
1295   ``WARNING``, and is used to handle logging events in the absence of any
1296   logging configuration. The end result is to just print the message to
1297   ``sys.stderr``. This replaces the earlier error message saying that
1298   "no handlers could be found for logger XYZ". If you need the earlier
1299   behaviour for some reason, ``lastResort`` can be set to ``None``.
1300
1301   .. versionadded:: 3.2
1302
1303Integration with the warnings module
1304------------------------------------
1305
1306The :func:`captureWarnings` function can be used to integrate :mod:`logging`
1307with the :mod:`warnings` module.
1308
1309.. function:: captureWarnings(capture)
1310
1311   This function is used to turn the capture of warnings by logging on and
1312   off.
1313
1314   If *capture* is ``True``, warnings issued by the :mod:`warnings` module will
1315   be redirected to the logging system. Specifically, a warning will be
1316   formatted using :func:`warnings.formatwarning` and the resulting string
1317   logged to a logger named ``'py.warnings'`` with a severity of :const:`WARNING`.
1318
1319   If *capture* is ``False``, the redirection of warnings to the logging system
1320   will stop, and warnings will be redirected to their original destinations
1321   (i.e. those in effect before ``captureWarnings(True)`` was called).
1322
1323
1324.. seealso::
1325
1326   Module :mod:`logging.config`
1327      Configuration API for the logging module.
1328
1329   Module :mod:`logging.handlers`
1330      Useful handlers included with the logging module.
1331
1332   :pep:`282` - A Logging System
1333      The proposal which described this feature for inclusion in the Python standard
1334      library.
1335
1336   `Original Python logging package <https://www.red-dove.com/python_logging.html>`_
1337      This is the original source for the :mod:`logging` package.  The version of the
1338      package available from this site is suitable for use with Python 1.5.2, 2.1.x
1339      and 2.2.x, which do not include the :mod:`logging` package in the standard
1340      library.
1341