It would be nice if all of the raw data could be passed back if requested instead of only the minimum (for executions with n_repeats>1)
|
execution_time[i] = np.min(measurements) |
So one could calculate standard deviation for the runs or perform other actions with it.
While at it it might be good idea to pass the raw data back in an extra variable instead adding it to the fitted dictionary:
|
if return_raw_data: |
|
fitted['measures'] = ns |
|
fitted['times'] = time |
|
|
|
return best, fitted |
Idea:
return_values = (best, fitted)
if return_raw_data:
return_values = (best, fitted, {
'measures' = ns,
'times' = time
})
return return_values
Happy to provide a PR!
It would be nice if all of the raw data could be passed back if requested instead of only the minimum (for executions with
n_repeats>1)big_O/big_o/big_o.py
Line 61 in d700425
So one could calculate standard deviation for the runs or perform other actions with it.
While at it it might be good idea to pass the raw data back in an extra variable instead adding it to the fitted dictionary:
big_O/big_o/big_o.py
Lines 171 to 175 in d700425
Idea:
Happy to provide a PR!