small class to quickly profile execution in python without decomposing your code into functions to use "timeit" decorators etc.

ProfilerLog.py

class ProfilerLog(object):
    def __init__(self, profiler_name="", start_ts=None):
        self._profiler_name = profiler_name
        self._start_ts = start_ts if start_ts else time.time()
        self._steps_data = OrderedDict()
        self._aux_data = {}

    def step(self, step_name, begin_ts=None, end_ts=None):
        if step_name not in self._steps_data:
            begin_ts = begin_ts if begin_ts else time.time()
            self._steps_data[step_name] = {'begin_ts': begin_ts, 'end_ts': end_ts}
        else:
            end_ts = end_ts if end_ts else time.time()
            self._steps_data[step_name]['end_ts'] = end_ts

    def get_formatted_step_log(self, step_name):
        if step_name in self._steps_data:
            step_data = self._steps_data.get(step_name)
            if step_data.get('begin_ts'):
                if step_data.get('end_ts'):
                    step_took_sec = step_data.get('end_ts') - step_data.get('begin_ts')
                    since_start_took_sec = step_data.get('end_ts') - self._start_ts
                    step_duration_str = "%.3f" % step_took_sec
                    since_start_duration_str = "%.3f" % since_start_took_sec
                    return f"profiling {self._profiler_name} step: {step_name}: {step_duration_str} s (total: {since_start_duration_str}s)"
                else:
                    return f"profiling {self._profiler_name} step: {step_name} ERROR: missing end ts"
            else:
                return f"profiling {self._profiler_name} step: {step_name} ERROR: missing begin ts"

    def get_formatted_step_aux_log(self, step_name, msg):
        if step_name == 'no_stop_prefix':
            return f"profiling {self._profiler_name} {msg}"
        elif step_name != 'default':
            return f"profiling {self._profiler_name}    |_ {msg}"
        #     return f"profiling {self._profiler_name} step: {step_name} |_detail: {msg}"
        return f"profiling {self._profiler_name} |_ {msg}"

    def log_step(self, step_name=None, include_aux=True):
        for stored_step_name in self._steps_data:
            if step_name:
                if stored_step_name == step_name:
                    step_profile_str = self.get_formatted_step_log(stored_step_name)
                    log.info(step_profile_str)
                    break
            else:
                step_profile_str = self.get_formatted_step_log(stored_step_name)
                log.info(step_profile_str)

            if include_aux:
                for msg in self._aux_data.get(stored_step_name, []):
                    log.info(self.get_formatted_step_aux_log(stored_step_name, msg))

        if include_aux:
            if self._aux_data.get('default', []):
                log.info(self.get_formatted_step_aux_log('no_stop_prefix', "[default details]"))
            for msg in self._aux_data.get('default', []):
                log.info(self.get_formatted_step_aux_log('default', msg))

    def log_detail(self, msg, step_name='default'):
        try:
            self._aux_data[step_name].append(msg)
        except KeyError:
            self._aux_data[step_name] = []
            self._aux_data[step_name].append(msg)

Use by invoking "step" method twice around code to profile:

step_prof = ProfilerLog(profiler_name="0000001.csv", start_ts=time.time())

step_prof.step('csv_load')
data = get_current_data_part_csv_as_dataframe("0000001.csv", skiprows=lines_to_skip)
step_prof.step('csv_load')

at the end call "log_step" with optional step name. With no step name passed all registered steps are logged:

step_prof.log_step()

which will produce output similar to this:

2023-06-15 15:33:30,311 [INFO ]: profiling 0000001.csv step: getting_lines_to_skip: 2.210 s (total: 2.210s)
2023-06-15 15:33:30,312 [INFO ]: profiling 0000001.csv step: csv_load: 1.077 s (total: 3.288s)
2023-06-15 15:33:30,312 [INFO ]: profiling 0000001.csv step: next_part_csv_load: 0.041 s (total: 3.329s)
2023-06-15 15:33:30,312 [INFO ]: profiling 0000001.csv step: 1s_downsampling: 18.821 s (total: 22.358s)
2023-06-15 15:33:30,312 [INFO ]: profiling 0000001.csv step: 1s_downsampling - analog re-write, parsing datetime: 16.542 s (total: 20.079s)
2023-06-15 15:33:30,312 [INFO ]: profiling 0000001.csv step: 1s_downsampling - analog re-write, filtering from: 0.053 s (total: 20.132s)
2023-06-15 15:33:30,312 [INFO ]: profiling 0000001.csv step: 1s_downsampling - analog re-write, converting epoch: 2.226 s (total: 22.358s)
2023-06-15 15:33:30,312 [INFO ]: profiling 0000001.csv step: head_tail_trimming: 0.009 s (total: 22.368s)
2023-06-15 15:33:30,312 [INFO ]: profiling 0000001.csv step: interpolation: 11.747 s (total: 34.115s)