logml.stopwatch

Simple stopwatch module to track execution performance.

Classes

NullTracker(*args, **kwargs)	Default time tracker.
Stopwatch([enabled, name, summary_file])	Simple time tracker to check execution time.
TimeTracker(parent_sw, name)	Context manager that tracks time.

class logml.stopwatch.TimeTracker(parent_sw, name: str)

Bases: object

Context manager that tracks time.

class logml.stopwatch.NullTracker(*args, **kwargs)

Bases: object

Default time tracker.

class logml.stopwatch.Stopwatch(enabled: bool = True, name: Optional[str] = None, summary_file=None)

Bases: object

Simple time tracker to check execution time.

Usage: .. code-block:

sw = Stopwatch()

@sw.decorate('init')  # name optional
def fn1():
    pass

# track execution without a function:
with sw.track("render_loop"):
    while(not stop):
        render()

enable()

Enable tracking.

disable()

Disable tracking.

reset_state(profile, log_file: str)

Resets state of SW.

property enabled: bool

Checks whether SW was enabled.

track(name)

Returns context manager to be used with tracked block:

with sw.track("save_to_db"):
    hibernate.save()

on_start(name: str)

Invoke on entering a function.

on_finish(name: str, duration: float)

Invoke when exiting from the function.

decorate(name_or_func)

Decorator for functions being tracked. Can be used as is, or with name as parameter.

get_summary_df(max_threshold: Optional[float] = None) → Optional[pandas.core.frame.DataFrame]

Wraps result of get_summary into pandas dataframe

Parameters

max_threshold – when provided, datafarme is filtered to contain max time >= this value

Returns

pd.DataFrame

get_summary()

Fetch summary for time tracks (sum, count, statistics). :rtype: list or dictionaries, one per name.

dump_summary()

Creates a file with a summary of seen runs.