Module Gnumed.pycommon.gmLog2

GNUmed logging framework setup.

All error logging, user notification and otherwise unhandled exception handling should go through classes or functions of this module.

Theory of operation:

This module tailors the standard logging framework to the needs of GNUmed.

By importing gmLog2 into your code you'll get the root logger send to a unicode file with messages in a format useful for debugging. The filename is either taken from the command line (–log-file=…) or derived from the name of the main application.

The log file will be found in one of the following standard locations:

1) given on the command line as "–log-file=LOGFILE" 2) ~/./.log 3) /dir/of/binary/.log (mainly for DOS/Windows)

where is derived from the name of the main application.

If you want to specify just a directory for the log file you must end the –log-file definition with a slash.

By importing "logging" and getting a logger your modules never need to worry about the real message destination or whether at any given time there's a valid logger available.

Your MAIN module simply imports gmLog2 and all other modules will merrily and automagically start logging away.

Ad hoc call stack logging recipe:

    call_stack = inspect.stack()
    call_stack.reverse()
    for idx in range(1, len(call_stack)):
            caller = call_stack[idx]
            _log.debug('%s[%s] @ [%s] in [%s]', ' '* idx, caller[3], caller[2], caller[1])
    del call_stack
Expand source code
"""GNUmed logging framework setup.

All error logging, user notification and otherwise unhandled 
exception handling should go through classes or functions of 
this module.

Theory of operation:

This module tailors the standard logging framework to
the needs of GNUmed.

By importing gmLog2 into your code you'll get the root
logger send to a unicode file with messages in a format useful
for debugging. The filename is either taken from the
command line (--log-file=...) or derived from the name
of the main application.

The log file will be found in one of the following standard
locations:

1) given on the command line as "--log-file=LOGFILE"
2) ~/.<base_name>/<base_name>.log
3) /dir/of/binary/<base_name>.log               (mainly for DOS/Windows)

where <base_name> is derived from the name
of the main application.

If you want to specify just a directory for the log file you
must end the --log-file definition with a slash.

By importing "logging" and getting a logger your modules
never need to worry about the real message destination or whether
at any given time there's a valid logger available.

Your MAIN module simply imports gmLog2 and all other modules
will merrily and automagically start logging away.


Ad hoc call stack logging recipe:

        call_stack = inspect.stack()
        call_stack.reverse()
        for idx in range(1, len(call_stack)):
                caller = call_stack[idx]
                _log.debug('%s[%s] @ [%s] in [%s]', ' '* idx, caller[3], caller[2], caller[1])
        del call_stack
"""
# TODO:
# - exception()
# - ascii_ctrl2mnemonic()
#========================================================================
__author__  = "K. Hilbert <Karsten.Hilbert@gmx.net>"
__license__ = "GPL v2 or later (details at https://www.gnu.org)"


# stdlib
import logging
import sys
import os
import datetime as pydt
import random


_logfile_name = None
_logfile = None
_logfile_name_printed = False


# table used for cooking non-printables
AsciiName = ['<#0-0x00-nul>',
                         '<#1-0x01-soh>',
                         '<#2-0x02-stx>',
                         '<#3-0x03-etx>',
                         '<#4-0x04-eot>',
                         '<#5-0x05-enq>',
                         '<#6-0x06-ack>',
                         '<#7-0x07-bel>',
                         '<#8-0x08-bs>',
                         '<#9-0x09-ht>',
                         '<#10-0x0A-lf>',
                         '<#11-0x0B-vt>',
                         '<#12-0x0C-ff>',
                         '<#13-0x0D-cr>',
                         '<#14-0x0E-so>',
                         '<#15-0x0F-si>',
                         '<#16-0x10-dle>',
                         '<#17-0x11-dc1/xon>',
                         '<#18-0x12-dc2>',
                         '<#19-0x13-dc3/xoff>',
                         '<#20-0x14-dc4>',
                         '<#21-0x15-nak>',
                         '<#22-0x16-syn>',
                         '<#23-0x17-etb>',
                         '<#24-0x18-can>',
                         '<#25-0x19-em>',
                         '<#26-0x1A-sub>',
                         '<#27-0x1B-esc>',
                         '<#28-0x1C-fs>',
                         '<#29-0x1D-gs>',
                         '<#30-0x1E-rs>',
                         '<#31-0x1F-us>'
                        ]

# msg = reduce(lambda x, y: x+y, (map(self.__char2AsciiName, list(tmp))), '')
#
#       def __char2AsciiName(self, aChar):
#               try:
#                       return AsciiName[ord(aChar)]
#               except IndexError:
#                       return aChar
#
#       def __tracestack(self):
#               """extract data from the current execution stack
#
#               this is rather fragile, I guess
#               """
#               stack = traceback.extract_stack()
#               self.__modulename = stack[-4][0]
#               self.__linenumber = stack[-4][1]
#               self.__functionname = stack[-4][2]
#               if (self.__functionname == "?"):
#                       self.__functionname = "Main"

#===============================================================
# external API
#===============================================================
def flush():
        """Log a <synced> line and flush handlers."""
        logger = logging.getLogger('gm.logging')
        logger.critical('-------- synced log file -------------------------------')
        root_logger = logging.getLogger()
        for handler in root_logger.handlers:
                handler.flush()

#===============================================================
def log_instance_state(instance):
        """Log the state of a class instance."""
        logger = logging.getLogger('gm.logging')
        logger.debug('state of %s', instance)
        for attr in [ a for a in dir(instance) if not a.startswith('__') ]:
                try:
                        val = getattr(instance, attr)
                except AttributeError:
                        val = '<cannot access>'
                logger.debug('  %s: %s', attr, val)

#===============================================================
def log_stack_trace(message:str=None, t=None, v=None, tb=None):
        """Log exception details and stack trace.

        (t,v,tb) are what sys.exc_info() returns.

        If any of (t,v,tb) is None it is attempted to be
        retrieved from sys.exc_info().

        Args:
                message: arbitrary message to add in
                t: an exception type
                v: an exception value
                tb: a traceback object
        """

        logger = logging.getLogger('gm.logging')

        if t is None:
                t = sys.exc_info()[0]
        if v is None:
                v = sys.exc_info()[1]
        if tb is None:
                tb = sys.exc_info()[2]
        if tb is None:
                logger.debug('sys.exc_info() did not return a traceback object, trying sys.last_traceback')
                try:
                        tb = sys.last_traceback
                except AttributeError:
                        logger.debug('no stack to trace (no exception information available)')
                        return

        # log exception details
        logger.debug('exception: %s', v)
        logger.debug('type: %s', t)
        logger.debug('list of attributes:')
        for attr in [ a for a in dir(v) if not a.startswith('__') ]:
                try:
                        val = getattr(v, attr)
                except AttributeError:
                        val = '<cannot access>'
                logger.debug('  %s: %s', attr, val)

        # make sure we don't leave behind a binding
        # to the traceback as warned against in
        # sys.exc_info() documentation
        try:
                # recurse back to root caller
                while 1:
                        if not tb.tb_next:
                                break
                        tb = tb.tb_next
                # put the frames on a stack
                stack_of_frames = []
                frame = tb.tb_frame
                while frame:
                        stack_of_frames.append(frame)
                        frame = frame.f_back
        finally:
                del tb
        stack_of_frames.reverse()

        if message is not None:
                logger.debug(message)
        logger.debug('stack trace follows:')
        logger.debug('(locals by frame, outmost frame first)')
        for frame in stack_of_frames:
                logger.debug (
                        '--- frame [%s]: #%s, %s -------------------',
                        frame.f_code.co_name,
                        frame.f_lineno,
                        frame.f_code.co_filename
                )
                for varname, value in frame.f_locals.items():
                        if varname == '__doc__':
                                continue
                        logger.debug('%20s = %s', varname, value)

#---------------------------------------------------------------
def log_multiline(level, message:str=None, line_prefix:str=None, text:str=None):
        """Log multi-line text in a standard format.

        Args:
                level: a log level
                message: an arbitrary message to add in
                line_prefix: a string to prefix lines with
                text: the multi-line text to log
        """
        if text is None:
                return

        if message is None:
                message = 'multiline text:'
        if line_prefix is None:
                line_template = '  > %s'
        else:
                line_template = '%s: %%s' % line_prefix
        lines2log = [message]
        lines2log.extend([ line_template % line for line in text.split('\n') ])
        logger = logging.getLogger('gm.logging')
        logger.log(level, '\n'.join(lines2log))

#---------------------------------------------------------------
__current_log_step = 1

def log_step(level:int=logging.DEBUG, message:str=None, restart:bool=False):
        if restart:
                global __current_log_step
                __current_log_step = 1
        logger = logging.getLogger('gm.logging')
        if message:
                logger.log(level, '%s - %s', __current_log_step, message)
        else:
                logger.log(level, '%s', __current_log_step)
        __current_log_step += 1

#---------------------------------------------------------------
def print_logfile_name():
        global _logfile_name_printed
        if _logfile_name_printed:
                return

        print('Log file:', _logfile_name)
        _logfile_name_printed = True

#===============================================================
# internal API
#===============================================================
__words2hide = []

def add_word2hide(word:str):
        """Add a string to the list of strings to be scrubbed from logging output.

        Useful for hiding credentials etc.
        """
        if word is None:
                return

        if word.strip() == '':
                return

        if word not in __words2hide:
                __words2hide.append(str(word))

#---------------------------------------------------------------
__original_logger_write_func = None

def __safe_logger_write_func(s):
        for word in __words2hide:
                # throw away up to 4 bits (plus the randint() cost)
                random.getrandbits(random.randint(1, 4))
                # from there generate a replacement string valid for
                # *this* round of replacements of *this* word,
                # this approach won't mitigate guessing trivial passwords
                # from replacements of known data (a known-plaintext attack)
                # but will make automated searching for replaced strings
                # in the log more difficult
                bummer = hex(random.randint(0, sys.maxsize)).lstrip('0x')
                s = s.replace(word, bummer)
        __original_logger_write_func(s)

#---------------------------------------------------------------
def __setup_logging():

        global _logfile
        if _logfile is not None:
                return True

        if not __get_logfile_name():
                return False

        _logfile = open(_logfile_name, mode = 'wt', encoding = 'utf8', errors = 'replace')
        global __original_logger_write_func
        __original_logger_write_func = _logfile.write
        _logfile.write = __safe_logger_write_func

        # setup
        fmt = '%(asctime)s  %(levelname)-8s  %(name)-12s  [%(thread)d %(threadName)-10s]  (%(pathname)s::%(funcName)s() #%(lineno)d): %(message)s'
        logging.basicConfig (
                format = fmt,
                datefmt = '%Y-%m-%d %H:%M:%S',
                level = logging.DEBUG,
                stream = _logfile
        )
        logging.captureWarnings(True)
        logger = logging.getLogger()

        # start logging
        #logger = logging.getLogger('gm.logging')
        logger.critical('-------- start of logging ------------------------------')
        logger.info('log file is <%s>', _logfile_name)
        logger.info('log level is [%s]', logging.getLevelName(logger.getEffectiveLevel()))
        logger.info('log file encoding is <utf8>')
        logger.debug('log file .write() patched from original %s to patched %s', __original_logger_write_func, __safe_logger_write_func)

#---------------------------------------------------------------
def __get_logfile_name():

        global _logfile_name
        if _logfile_name is not None:
                return _logfile_name

        def_log_basename = os.path.splitext(os.path.basename(sys.argv[0]))[0]
        default_logfile_name = '%s-%s-%s.log' % (
                def_log_basename,
                pydt.datetime.now().strftime('%Y_%m_%d-%H_%M_%S'),
                os.getpid()
        )

        # given on command line ?
        for option in sys.argv[1:]:
                if option.startswith('--log-file='):
                        (opt_name, value) = option.split('=')
                        (dir_name, file_name) = os.path.split(value)
                        if dir_name == '':
                                dir_name = '.'
                        if file_name == '':
                                file_name = default_logfile_name
                        _logfile_name = os.path.abspath(os.path.expanduser(os.path.join(dir_name, file_name)))
                        return True

        # else store it in ~/.local/gnumed/logs/def_log_basename/default_logfile_name
        dir_name = os.path.expanduser(os.path.join('~', '.local', 'gnumed', 'logs', def_log_basename))
        try:
                os.makedirs(dir_name)
        except OSError as e:
                if (e.errno == 17) and not os.path.isdir(dir_name):
                        raise

        _logfile_name = os.path.join(dir_name, default_logfile_name)

        return True

#===============================================================
# main
#---------------------------------------------------------------
__setup_logging()

if __name__ == '__main__':

        if len(sys.argv) < 2:
                sys.exit()

        if sys.argv[1] != 'test':
                sys.exit()

        print_logfile_name()
        #-----------------------------------------------------------
        def test():
                print_logfile_name()
                log_step(message = 'testing')
                logger = logging.getLogger('gmLog2.test')
                log_step()
                logger.error('test %s', [1,2,3])
                log_step(message = 'still testing')
                logger.error("I expected to see %s::test()" % __file__)
                log_step(message = 'restarting', restart = True)
                add_word2hide('super secret passphrase')
                log_step()
                logger.debug('credentials: super secret passphrase')
                log_step()
                try:
                        int(None)
                except Exception:
                        logger.exception('unhandled exception')
                        log_stack_trace()
                flush()
                log_step(message = 'done')
        #-----------------------------------------------------------
        test()

Functions

def add_word2hide(word: str)

Add a string to the list of strings to be scrubbed from logging output.

Useful for hiding credentials etc.

Expand source code
def add_word2hide(word:str):
        """Add a string to the list of strings to be scrubbed from logging output.

        Useful for hiding credentials etc.
        """
        if word is None:
                return

        if word.strip() == '':
                return

        if word not in __words2hide:
                __words2hide.append(str(word))
def flush()

Log a line and flush handlers.

Expand source code
def flush():
        """Log a <synced> line and flush handlers."""
        logger = logging.getLogger('gm.logging')
        logger.critical('-------- synced log file -------------------------------')
        root_logger = logging.getLogger()
        for handler in root_logger.handlers:
                handler.flush()
def log_instance_state(instance)

Log the state of a class instance.

Expand source code
def log_instance_state(instance):
        """Log the state of a class instance."""
        logger = logging.getLogger('gm.logging')
        logger.debug('state of %s', instance)
        for attr in [ a for a in dir(instance) if not a.startswith('__') ]:
                try:
                        val = getattr(instance, attr)
                except AttributeError:
                        val = '<cannot access>'
                logger.debug('  %s: %s', attr, val)
def log_multiline(level, message: str = None, line_prefix: str = None, text: str = None)

Log multi-line text in a standard format.

Args

level
a log level
message
an arbitrary message to add in
line_prefix
a string to prefix lines with
text
the multi-line text to log
Expand source code
def log_multiline(level, message:str=None, line_prefix:str=None, text:str=None):
        """Log multi-line text in a standard format.

        Args:
                level: a log level
                message: an arbitrary message to add in
                line_prefix: a string to prefix lines with
                text: the multi-line text to log
        """
        if text is None:
                return

        if message is None:
                message = 'multiline text:'
        if line_prefix is None:
                line_template = '  > %s'
        else:
                line_template = '%s: %%s' % line_prefix
        lines2log = [message]
        lines2log.extend([ line_template % line for line in text.split('\n') ])
        logger = logging.getLogger('gm.logging')
        logger.log(level, '\n'.join(lines2log))
def log_stack_trace(message: str = None, t=None, v=None, tb=None)

Log exception details and stack trace.

(t,v,tb) are what sys.exc_info() returns.

If any of (t,v,tb) is None it is attempted to be retrieved from sys.exc_info().

Args

message
arbitrary message to add in
t
an exception type
v
an exception value
tb
a traceback object
Expand source code
def log_stack_trace(message:str=None, t=None, v=None, tb=None):
        """Log exception details and stack trace.

        (t,v,tb) are what sys.exc_info() returns.

        If any of (t,v,tb) is None it is attempted to be
        retrieved from sys.exc_info().

        Args:
                message: arbitrary message to add in
                t: an exception type
                v: an exception value
                tb: a traceback object
        """

        logger = logging.getLogger('gm.logging')

        if t is None:
                t = sys.exc_info()[0]
        if v is None:
                v = sys.exc_info()[1]
        if tb is None:
                tb = sys.exc_info()[2]
        if tb is None:
                logger.debug('sys.exc_info() did not return a traceback object, trying sys.last_traceback')
                try:
                        tb = sys.last_traceback
                except AttributeError:
                        logger.debug('no stack to trace (no exception information available)')
                        return

        # log exception details
        logger.debug('exception: %s', v)
        logger.debug('type: %s', t)
        logger.debug('list of attributes:')
        for attr in [ a for a in dir(v) if not a.startswith('__') ]:
                try:
                        val = getattr(v, attr)
                except AttributeError:
                        val = '<cannot access>'
                logger.debug('  %s: %s', attr, val)

        # make sure we don't leave behind a binding
        # to the traceback as warned against in
        # sys.exc_info() documentation
        try:
                # recurse back to root caller
                while 1:
                        if not tb.tb_next:
                                break
                        tb = tb.tb_next
                # put the frames on a stack
                stack_of_frames = []
                frame = tb.tb_frame
                while frame:
                        stack_of_frames.append(frame)
                        frame = frame.f_back
        finally:
                del tb
        stack_of_frames.reverse()

        if message is not None:
                logger.debug(message)
        logger.debug('stack trace follows:')
        logger.debug('(locals by frame, outmost frame first)')
        for frame in stack_of_frames:
                logger.debug (
                        '--- frame [%s]: #%s, %s -------------------',
                        frame.f_code.co_name,
                        frame.f_lineno,
                        frame.f_code.co_filename
                )
                for varname, value in frame.f_locals.items():
                        if varname == '__doc__':
                                continue
                        logger.debug('%20s = %s', varname, value)
def log_step(level: int = 10, message: str = None, restart: bool = False)
Expand source code
def log_step(level:int=logging.DEBUG, message:str=None, restart:bool=False):
        if restart:
                global __current_log_step
                __current_log_step = 1
        logger = logging.getLogger('gm.logging')
        if message:
                logger.log(level, '%s - %s', __current_log_step, message)
        else:
                logger.log(level, '%s', __current_log_step)
        __current_log_step += 1
def print_logfile_name()
Expand source code
def print_logfile_name():
        global _logfile_name_printed
        if _logfile_name_printed:
                return

        print('Log file:', _logfile_name)
        _logfile_name_printed = True