Logging is packed in these 5 functions
reset_logger("trace", console_width=80)
Trace("Testing {1,2,3}")
Debug("TESTING {1,2,3}")
Info("TESTING {1,2,3}")
Warn("TESTING {1,2,3}")
Excep("TESTING {1,2,3}")[08/24/24 12:16:05] TRACE Testing {1,2,3} 1838145784.py:<module>:2
DEBUG TESTING {1,2,3} 1838145784.py:<module>:3
INFO TESTING {1,2,3} 1838145784.py:<module>:4
WARNING TESTING {1,2,3} 1838145784.py:<module>:5
ERROR TESTING {1,2,3} 1838145784.py:<module>:6 NoneType: None
Excep will also print a traceback
def do():
try:
1 / 0
except Exception as e:
Excep(e)
def do2():
do()
do2()[08/24/24 12:16:07] ERROR division by zero 1923207781.py:do:5 ZeroDivisionError: division by zero ╭─ Traceback (most recent call─╮ │ in do:3 │ │ │ │ 1 def do(): │ │ 2 │ try: │ │ ❱ 3 │ │ 1 / 0 │ │ 4 │ except Exception as │ │ 5 │ │ Excep(e) │ │ 6 │ ╰──────────────────────────────╯ ZeroDivisionError: division by zero
All logging functions have an optional depth that will be helpful to raise the context level to outer functions
def do():
Info("In do")
do()[08/24/24 12:16:15] INFO In do 809811040.py:do:2
def do2():
Info("Log will still say it is from `do2` now, to the right of log print")
def do():
do2()
do()[08/24/24 12:17:13] INFO Log will still say it is from 2099333876.py:do2:2 `do2` now, to the right of log print
def do2():
Info("But now, log will still say it is from do", depth=1)
def do():
do2()
do()[08/24/24 12:17:25] INFO But now, log will still say it 4054019042.py:do:5 is from do
in_logger_mode (level:str)
return’s T/F, checking if logger is in a specific mode or not
logger_mode (level)
temporarily, using with context, set the level to something else
get_logger_level ()
get the current logger’s level
Let’s log every level in the do function below. We can control what we need to log from outside the function’s context by
using with <level>_model():
def do():
Trace(0)
Debug(1)
Info(2)
Warn(3)
Excep(4)
def line(x):
sep = "=" * 20
print(f"{sep}{x}{sep}")
print(f"{in_excep_mode()=}")
print(f"{in_warn_mode()=}")
print(f"{in_info_mode()=}")
print(f"{in_debug_mode()=}")
print(f"{in_trace_mode()=}")
reset_logger()
with excep_mode():
line("Excep mode")
do()
with warn_mode():
line("Warn mode")
do()
with info_mode():
line("Info mode")
do()
with debug_mode():
line("Debug mode")
do()
with trace_mode():
line("Trace mode")
do()====================Excep mode====================
in_excep_mode()=True
in_warn_mode()=False
in_info_mode()=False
in_debug_mode()=False
in_trace_mode()=False[08/24/24 12:13:08] ERROR 4 3432048509.py:do:6 NoneType: None
====================Warn mode====================
in_excep_mode()=False
in_warn_mode()=True
in_info_mode()=False
in_debug_mode()=False
in_trace_mode()=False[08/24/24 12:13:08] WARNING 3 3432048509.py:do:5
ERROR 4 3432048509.py:do:6 NoneType: None
====================Info mode====================
in_excep_mode()=False
in_warn_mode()=False
in_info_mode()=True
in_debug_mode()=False
in_trace_mode()=False[08/24/24 12:13:08] INFO 2 3432048509.py:do:4
WARNING 3 3432048509.py:do:5
ERROR 4 3432048509.py:do:6 NoneType: None
====================Debug mode====================
in_excep_mode()=False
in_warn_mode()=False
in_info_mode()=False
in_debug_mode()=True
in_trace_mode()=False[08/24/24 12:13:08] DEBUG 1 3432048509.py:do:3
INFO 2 3432048509.py:do:4
WARNING 3 3432048509.py:do:5
ERROR 4 3432048509.py:do:6 NoneType: None
====================Trace mode====================
in_excep_mode()=False
in_warn_mode()=False
in_info_mode()=False
in_debug_mode()=False
in_trace_mode()=True[08/24/24 12:13:08] TRACE 0 3432048509.py:do:2
DEBUG 1 3432048509.py:do:3
INFO 2 3432048509.py:do:4
WARNING 3 3432048509.py:do:5
ERROR 4 3432048509.py:do:6 NoneType: None
The in_<level>_mode gives an additional layer of control, to be used for debugging dynamically. Let’s say, you want to show an image (for the sake of debugging)
def do(im_path):
from torch_snippets import show, read
im = read(im_path)
show(im, sz=3)
print(im.mean())
do("assets/Preamble.png")
145.5542982442515But now you are happy with your code and don’t want the show, say the code is going to production. A common way out is to just comment that line
def do(im_path):
from torch_snippets import show, read
im = read(im_path)
# show(im, sz=3) # line is commented, but will need to be re-uncommented any time it needs debugging
print(im.mean())
do("assets/Preamble.png")145.5542982442515But if you want to re-check, it’s a pain to again uncomment. Not to mention this method is not scalable to 100s of lines of code. The simple way to deal with such transient code that needs to activate only when you want it to, is to enclose in an if in_<level>_mode conditional like so
def do(im_path):
from torch_snippets import show, read
im = read(im_path)
if in_debug_mode():
show(im, sz=3)
print(im.mean())
do("assets/Preamble.png")145.5542982442515This way, you can always activate the show by calling do with a temporary with_debug_mode context
with debug_mode():
do("assets/Preamble.png")
145.5542982442515reset_logger(console_width=100)
with trace_mode():
with notify_waiting("Downloading"):
time.sleep(10)
with notify_waiting("Downloading", spinner="guess_the_movie"):
time.sleep(10)
with notify_waiting("One more message", spinner="earth"):
time.sleep(3)[08/24/24 12:36:51] TRACE Available Spinners: dict_keys(['dots', 'dots2', 'dots3', 'dots4', 'dots5', 'dots6', 'dots7', 'dots8', 'dots9', 3340474699.py:notify_waiting:41 'dots10', 'dots11', 'dots12', 'dots8Bit', 'line', 'line2', 'pipe', 'simpleDots', 'simpleDotsScrolling', 'star', 'star2', 'flip', 'hamburger', 'growVertical', 'growHorizontal', 'balloon', 'balloon2', 'noise', 'bounce', 'boxBounce', 'boxBounce2', 'triangle', 'arc', 'circle', 'squareCorners', 'circleQuarters', 'circleHalves', 'squish', 'toggle', 'toggle2', 'toggle3', 'toggle4', 'toggle5', 'toggle6', 'toggle7', 'toggle8', 'toggle9', 'toggle10', 'toggle11', 'toggle12', 'toggle13', 'arrow', 'arrow2', 'arrow3', 'bouncingBar', 'bouncingBall', 'smiley', 'monkey', 'hearts', 'clock', 'earth', 'material', 'moon', 'runner', 'pong', 'shark', 'dqpb', 'weather', 'christmas', 'grenade', 'point', 'layer', 'betaWave', 'aesthetic', 'guess', 'guess_the_movie']) Using: clock
[08/24/24 12:37:01] INFO Downloading - Completed in 10.01 s contextlib.py:__exit__:144
[08/24/24 12:37:11] INFO Downloading - Completed in 10.01 s contextlib.py:__exit__:144
[08/24/24 12:37:14] INFO One more message - Completed in 3.00 s contextlib.py:__exit__:144