saving cProfile results to readable external file

Question

I am using cProfile try to profile my codes:

pr = cProfile.Profile()
pr.enable()
my_func()   # the code I want to profile
pr.disable()
pr.print_stats()

However, the results are too long and cannot be fully displayed in the Spyder terminal (the function calls which take the longest time to run cannot be seen…). I also tried saving the results using

 cProfile.run('my_func()','profile_results')

but the output file is not in human readable format (tried with and without .txt suffix).

So my question is how can I save the profiling results to an external file which is human readable (like in a .txt format with all the words properly displayed)?

Asked By: Physicist

||

Source

Answer 1

Updated. You can get output of profiler using io.StringIO() and save it into file.
Here is an example:

import cProfile
import pstats
import io


def my_func():
    result = []
    for i in range(10000):
        result.append(i)

    return result

pr = cProfile.Profile()
pr.enable()

my_result = my_func()

pr.disable()
s = io.StringIO()
ps = pstats.Stats(pr, stream=s).sort_stats('tottime')
ps.print_stats()

with open('test.txt', 'w+') as f:
    f.write(s.getvalue())

Run our script and open test.txt. You will see readable result:

   10002 function calls in 0.003 seconds

   Ordered by: internal time

   ncalls  tottime  percall  cumtime  percall filename:lineno(function)
        1    0.002    0.002    0.003    0.003 /path_to_script.py:26(my_func)
    10000    0.001    0.000    0.001    0.000 {method 'append' of 'list' objects}
        1    0.000    0.000    0.000    0.000 {method 'disable' of '_lsprof.Profiler' objects}

Also I can recommend to use dump_stats + pstats.Stats. Here is an example how to use it. Structure of files:

# test_ex.py - just a small web app
import cProfile
import json
from functools import wraps
from flask import Flask

from example.mod1 import func1
from example.mod2 import func2

app = Flask(__name__)

# profiling decorator
def profiling():

    def _profiling(f):
        @wraps(f)
        def __profiling(*rgs, **kwargs):
            pr = cProfile.Profile()
            pr.enable()

            result = f(*rgs, **kwargs)

            pr.disable()
            # save stats into file
            pr.dump_stats('profile_dump')

            return result
        return __profiling
    return _profiling

# demonstration route with profiler
@app.route('/test')
@profiling()
def test():
    counter = func1()
    dict_data = func2()
    result = dict()

    for key, val in dict_data.items():
        result[key] = val + counter

    return json.dumps(result)


if __name__ == '__main__':
    app.run(debug=True, port=8083)

example package – let’s imagine that this is some kind of application logic.

# example.mod1
def func1():
    counter = 0

    for i in range(100000):
        counter += i

    return counter

# example.mod2
def func2():
    res = dict()

    for i in range(300000):
        res['key_' + str(i)] = i

    return res

Now let’s run server(python3 test_ex.py) and open http://localhost:8083/test. After a few seconds you will see long json. After that you will see profile_dump file in project folder. Now run python live interpreter in project folder and print our dump using pstats:

import pstats
p = pstats.Stats('profile_dump')
# skip strip_dirs() if you want to see full path's
p.strip_dirs().print_stats()

Also you can easy sorting results:

p.strip_dirs().sort_stats('tottime').print_stats()
p.strip_dirs().sort_stats('cumulative').print_stats()
p.strip_dirs().sort_stats().print_stats('mod1')

Hope this helps.

Answered By: Danila Ganchar

Answer 2

Expanding upon the previous answer, you can dump everything out to a .csv file to sort and play around with in your favorite spreadsheet application.

import pstats,StringIO

# print stats to a string
result=StringIO.StringIO()
pstats.Stats(filename,stream=result).print_stats()
result=result.getvalue()

# chop the string into a csv-like buffer
result='ncalls'+result.split('ncalls')[-1]
result='n'.join([','.join(line.rstrip().split(None,6)) for line in result.split('n')])

# save it to disk
f=open(filename.rsplit('.')[0]+'.csv','w')
f.write(result)
f.close()

Answered By: TheHeadlessSourceMan

Answer 3

You don’t really need StringIO, as a file qualifies as a stream.

import pstats

with open("profilingStatsAsText.txt", "w") as f:
    ps = pstats.Stats("profilingResults.cprof", stream=f)
    ps.sort_stats('cumulative')
    ps.print_stats()

Answered By: Migdalin HasNoLastName

Answer 4

you can use dump_stats. In Python 3.10:

with cProfile.Profile() as pr:
    my_func()  

pr.dump_stats('/path/to/filename.prof')

Answered By: Jay

Answer 5

You can run the profiler saving the output into a file, as you do:

import cProfile

cProfile.run('my_func()', 'profile_results')

And then format that result using the class pstats.Stats (https://docs.python.org/3/library/profile.html#the-stats-class):

import pstats

file = open('formatted_profile.txt', 'w')
profile = pstats.Stats('.profile_results', stream=file)
profile.sort_stats('cumulative') # Sorts the result according to the supplied criteria
profile.print_stats(15) # Prints the first 15 lines of the sorted report
file.close()

Answered By: Anto Barbero

saving cProfile results to readable external file

Question:

Answers: