Shulou(Shulou.com)06/03 Report--

Timeit is a built-in gadget in the Python standard library that can quickly test the performance of small pieces of code.

Get to know timeit

Timeit function:

Timeit.timeit (stmt, setup,timer, number)

Parameter description:

Stmt: abbreviation for statement, the code or statement you want to test, plain text, the default value is "pass"

Setup: configuration statement before running stmt, plain text, default value is also "pass"

Timer: timer. This parameter is generally ignored.

Number: the number of times stmt is executed. Default is 1000000, 1 million.

Repeat function:

Timeit.repeat (stmt, setup, timer, repeat, number)

Is the repeat version of timeit, you can specify the number of times to repeat timeit, the default is 3 times, and then return an array.

Give a simple example to illustrate the usage:

Import timeit

Print (timeit.timeit ('output = 1055')) # 0.014560436829924583

Print (timeit.repeat ('output = 1055')) # [0.01492984383367002,0.01342877489514649, 0.013638464966788888]

Well, it looks fine, but in fact, no one is going to test meaningless addition, subtraction, multiplication and division. We need to test our own code.

Test multiple lines of code

Testing multiple lines of code can use semicolons to concatenate statements.

Print (timeit.timeit (stmt='a=10;b=10;sum=a+b'))

You can also use three quotation marks to write stmt.

Import timeit

Import_module = "import random"

Testcode ='

Def test ():

Return random.randint (10,100)

''

Print (timeit.repeat (stmt=testcode, setup=import_module))

But in fact, it is quite ridiculous, their own code will be so simple? We are modular programming.

Test the functions in the module

If the functions you want to test are all in the same module, you can write timeit like this.

Import timeit

Import random

Import arrow

Local function

Def stupid1 ():

Return random.randint (1,10) depends on other functions

Def stupid2 ():

Return stupid1 () depends on other packages or modules

Def stupid3 ():

Return arrow.now ()

Print (timeit.timeit ('stupid1 ()', setup='from main import stupid1'))

Print (timeit.timeit ('stupid2 ()', setup='from main import stupid2'))

Print (timeit.timeit ('stupid3 ()', setup='from main import stupid3', number=100))

What is written like this is actually an one-line pattern.

Borrow default_timer

The default_timer that comes with timeit can be borrowed.

Import timeit

Import random

Def test ():

Return random.randint (10,100)

Starttime = timeit.default_timer ()

Print ("The starttime is:", starttime)

Test ()

Print ("The time difference is:", timeit.default_timer ()-starttime)

Usage of the command line

Timeit also supports command-line invocation, but I feel too tired to try.

C:\ pythontest > python-m timeit-s' text= "hello world"'

20000000 loops, best of 5: 13.1 nsec per loop

Share a case

On February 29, I think this year is a leap year. How many algorithms are there to calculate a leap year? Kong Yiji said that there are three kinds:

Def is_leap_year_0 (year):

If year% 4 = 0: if year% 100 = 0: if year% 400 = 0: return Trueelse: return False else: return Trueelse: return False

Def is_leap_year_1 (year):

Return year% 4 = 0 and (year% 100! = 0 or year% 400 = = 0)

Def is_leap_year_2 (year):

If year% 400 = 0: return Trueif year% 100 = 0: return Falseif year% 4 = = 0: return Truereturn False

Which of these three methods is the best? This can not be generalized, because it depends on what your parameters are, for example, 1991 is not a leap year, method 0 and method 1 return directly, but method 2 needs to go to the last if to know that it is not a leap year. In 2020, for example, method 2 returns directly, but methods 0 and 1 need to go to the innermost if to get the result. So we need sampling and testing to be fair, for example, from 1900 to 2900, each function ran 10000 times.

Timeit is not very convenient, the parameters it accepts can not be so complex, we need to wrap it.

Def perf_test (method):

Years = range (1900, 2900) if method = = 0: for y in years: is_leap_year_0 (y) if method = = 1: for y in years: is_leap_year_1 (y) if method = 2: for y in years: is_leap_year_2 (y)

Print (timeit ('perf_test (0)', setup='from main import perf_test', number=10000))

Print (timeit ('perf_test (1)', setup='from main import perf_test', number=10000))

Print (timeit ('perf_test (2)', setup='from main import perf_test', number=10000))

Guess which method has the best result? You have no idea.

1.6432780250906944

1.7527272370643914

0.0023122059646993876

Other ideas

In fact, timeit is still too weak, casual testing is OK, if you really want to check performance problems, you still need to use more professional means. For example:

PyCharm Profiler (Pro version feature)

CProfile

Pycallgraph

Memory_profiler

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.