logtool 0.4.post2

Methods and tools that assist logging

Methods and tools that assist logging. Can be installed from PyPI:

$ pip install logtool

log_call

A decorator for function and method definitions that logs at DEBUG level a variety of data about every call made to that entrypoint.

Intended to supercede @log_func and log_func_noargs (see below). See log_func for example output.

Optional arguments:

log_enter

Log entrance to the decorated method. Defaults to True.

log_args

Log the arguments passed to the decorated method. Defaults to True.

log_exit

Log exit/returns from the decorated method along with the execution time. Defaults to True.

log_rc

Log the value returned by the decorated method. Defaults to True.

log_trace

Log each line of the decorated method as it is executed. Defaults to False.

log_level

Log level to use for the logging of the call. Defaults to logging.DBEUG.

@logtool.log_call
def a_method (...):
 ...etc...

@logtool.log_call (log_args = False, log_rc = False)
def big_complex_data (...):
 ...etc...

log_func

A decorator for function and method definitions that logs at DEBUG level every call to that function or method along with its arguments.

eg

@logtool.log_wrap
def my_method (self, *args):
  ...stuff here...

Resulting log entry from a real production usage (with a few of the argumentvalues redacted):

Entered: function:test_tool.toolwrapper:email_report ((<test_tool.meshtool.Wrapper object at 0x7f19d4879c10>, path(u'../file.ext'), '[e-mail]', '[e-mail]', 'Interesting subject header') {})

The {} at the end shows that there were no named arguments passed to that call, else they would be shown there.

log_func_noargs

A decorator for function and method definitions that logs at DEBUG level every call to that function or method but without any arguments. This can be useful when traversing and dumping the arguments would be execssively expensive, or would potentially create infinite loops.

eg

@logtool.log_wrap_noargs
def my_method (self, *args):
  ...stuff here...

log_fault

Logs an exception in a standardised form, including the source file and line number of the exception, and if logging at DEBUG level, also logs a stack trace along with all the variables in each stack frame. eg

In WARN or higher mode:

CRITICAL <log_fault_impl:log_fault(24)> FAULT: /usr/local/lib/python2.7/dist-packages/workerd-0.1.26_gbb342e2-py2.7.egg/workerd/do.py(243): IOError(28, 'No space left on device')

When logging at DEBUG:

CRITICAL <log_fault_impl:log_fault(24)> FAULT: /usr/local/lib/python2.7/dist-packages/workerd-0.1.26_gbb342e2-py2.7.egg/workerd/do.py(243): IOError(28, 'No space left on device')
DEBUG <log_fault_impl:log_fault(26)> Locals by frame, innermost last:
DEBUG <log_fault_impl:log_fault(30)> Frame run in /usr/local/lib/python2.7/dist-packages/workerd-0.1.26_gbb342e2-py2.7.egg/workerd/do.py at line 248
DEBUG <log_fault_impl:log_fault(40)>                 self = <workerd.do.Do object at 0x7f5709e3d490>
DEBUG <log_fault_impl:log_fault(40)>                    e = [Errno 28] No space left on device
DEBUG <log_fault_impl:log_fault(40)>                   rc = 0
DEBUG <log_fault_impl:log_fault(30)> Frame wrapper_args in build/bdist.linux-x86_64/egg/mppy/log_wrap.py at line 27
DEBUG <log_fault_impl:log_fault(40)>                 args = (<workerd.do.Do object at 0x7f5709e3d490>,)
DEBUG <log_fault_impl:log_fault(40)>                   fn = <function do_job at 0x7f570a2936e0>
DEBUG <log_fault_impl:log_fault(40)>               kwargs = {}
DEBUG <log_fault_impl:log_fault(30)> Frame do_job in /usr/local/lib/python2.7/dist-packages/workerd-0.1.26_gbb342e2-py2.7.egg/workerd/do.py at line 227
DEBUG <log_fault_impl:log_fault(40)>                  toc = 1410867312.58
DEBUG <log_fault_impl:log_fault(40)>                 self = <workerd.do.Do object at 0x7f5709e3d490>
DEBUG <log_fault_impl:log_fault(40)>                  tic = 1410842559.54
DEBUG <log_fault_impl:log_fault(40)>                   rc = -99
DEBUG <log_fault_impl:log_fault(30)> Frame __setitem__ in build/bdist.linux-x86_64/egg/mppy/jsondict.py at line 69
DEBUG <log_fault_impl:log_fault(40)>                 self = {u'status': u'pending', u'notified_for': u'pending
DEBUG <log_fault_impl:log_fault(40)>                  key = execution_time
DEBUG <log_fault_impl:log_fault(40)>                  val = 24753.043578
DEBUG <log_fault_impl:log_fault(40)>               kwargs = {}
DEBUG <log_fault_impl:log_fault(30)> Frame wrapper in build/bdist.linux-x86_64/egg/mppy/jsondict.py at line 80
DEBUG <log_fault_impl:log_fault(40)>                 self = {u'status': u'pending', u'notified_for': u'pending
DEBUG <log_fault_impl:log_fault(40)>               kwargs = {}
DEBUG <log_fault_impl:log_fault(40)>                 attr = <bound method JsonDict.save of {u'status': u'pendi
DEBUG <log_fault_impl:log_fault(40)>                 args = ()
DEBUG <log_fault_impl:log_fault(40)>           was_loaded = True
DEBUG <log_fault_impl:log_fault(30)> Frame save in build/bdist.linux-x86_64/egg/mppy/jsondict.py at line 46
DEBUG <log_fault_impl:log_fault(40)>                force = False
DEBUG <log_fault_impl:log_fault(40)>                 self = {u'status': u'pending', u'notified_for': u'pending
DEBUG <log_fault_impl:log_fault(40)>                   fd = 5
DEBUG <log_fault_impl:log_fault(40)>                   fn = /var/spool/matterport/workerd/generate_mesh/d34fea

time_str

Simply returns a time_t (seconds since the epoch, possibly fractional) in a simple consistent string form suitable for logfiles, reports and the like.

See below under now for an example.

now

Reurns a tuple of the current time as a time_t, and its matching time_str. Getting the two together allows the string to be used for logs and the like, and the time_t to be used as a numeric. eg:

$ ipython
Python 2.7.6 (default, Mar 22 2014, 22:59:56)
Type "copyright", "credits" or "license" for more information.

IPython 1.2.1 -- An enhanced Interactive Python.
?         -> Introduction and overview of IPython's features.
%quickref -> Quick reference.
help      -> Python's own help system.
object?   -> Details about 'object', use 'object??' for extra details.

In [1]: import logtool

In [2]: logtool.now ()
Out[2]: (1411075417, '21:23:37 Thu 18 Sep 2014 Z+0000')

In [3]: logtool.time_str (logtool.now ()[0])
Out[3]: '14:23:42 Thu 18 Sep 2014 Z+0000'