Welcome, guest | Sign In | My Account | Store | Cart

Two convenience classes for applying different scenarios (data) to different functions and measuring timing as a practical framework for optimizing performance. NOTE: time.clock() and time.time() resolution depend on system time.clock() may be more appropriate.

Python, 104 lines
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
"""classes for timing during development

by wolfgang resch, 070722
"""

from time import clock, time

#--- timer decorator -----------------------------------
class Timer(object):
    def __init__(self, f):
        self.__f=f
        self.__name__=f.__name__
        self.__doc__=f.__doc__
        self.times=[]
    def __call__(self, *args, **kwargs):
        start=time()
        result=self.__f(*args, **kwargs)
        stop=time()
        self.times.append(stop-start)
    def report(self):
        results = {'min':min(self.times)*1000, 'max':max(self.times)*1000, 'mean':sum(self.times)/len(self.times)*1000}
        return self.__doc__.ljust(70) + "\t%(mean)6.5fms [%(min)6.5f - %(max)6.5f]" % results
    def reset(self):
        self.times=[]

#--- scenarios -------------------------------------------
class Scenario(object):
    def __init__(self, **kw):
        """
        pass each scenario as a kw={'desc':"", data:(), functions:()}
        
        for each scenario each function is applied to the data and results are printed
        """
        self.scenarios=kw
    def run(self, nr_iter):
        for title, scenario in self.scenarios.items():
            print 50*'-' + "\n" + title.upper()
            print scenario['desc']
            for f in scenario['functions']:
                for x in xrange(nr_iter):
                    f(*scenario['data'])
                print "\t" + f.report()
                f.reset()


#--- example: validators------------------------------------------
# validator takes a iterable object and checks if each element
# in the object is part of a defined alphabet
@Timer
def validate1(s, a):
    """validate1(s, a): list comprehension with a.index"""
    try:
        [a.index(x) for x in s]
        return True
    except:
        return False

@Timer
def validate2(s,a):
    """validate2(s, a): for loop with a.index"""
    for l in s:
        try:
            a.index(l)
        except:
            return False
    return True

@Timer
def validate3(s,a):
    """validate3(s, a): list comprehension with (l in a)"""
    return min([(l in a) for l in s])

@Timer
def validate4(s,a):
    """validate4(s, a): for loop with generator and (l in a)"""
    for x in ((l in a) for l in s):
        if not x:
            return False
    return True    

@Timer
def validate5(s,a):
    """validate5(s,a): convert s to set and compare to set alphabet"""
    return set(s).issubset(a)

#--- main-------------------------------------------------

if __name__=='__main__':
    replications=1000
    s_false="GATCTTGACGATGCGATGCGATG"*40 + "==" + "GATCTTGACGATGCGATGCGATG"*40
    s_true=80*"GATCTTGACGATGCGATGCGATG" + "GA"
    a=['G','A','T','C','R','W','B','C','M','T','Q','X','.','-','~','U']
   
    testValidator = Scenario(scene_1={'desc':"Validator evaluates to true; alphabet is a list",
                                      'data':(s_true, a),
                                      'functions':(validate1, validate2, validate3, validate4, validate5)},
                             scene_2= {'desc':"Validator evaluates to false; alphabet is a list",
                                      'data':(s_false, a),
                                      'functions':(validate1, validate2, validate3, validate4, validate5)},
                             scene_3={'desc':"Validator evaluates to true; alphabet is a tuple",
                                      'data':(s_true, tuple(a)),
                                      'functions':( validate3, validate4, validate5)})
    testValidator.run(1000)                         
-----------------------------

I like to systematically test different algorithms for some projects to help performance. This is mostly because I'm not a particularly good programmer, so i usually come up with slow ways of doing things. This is a very simple framework for putting functions through their paces, consisting of a timing decorator class and a class that applies different scenarios to different functions and measures timing. I did this b/c for some algorithms it's important to test different data as is the case for the validators in this example - these validators check if elements of a sequence s conform to an alphabet a. The performance of some algorithms depends on weather the validation fails or not. Please let me know any suggestions/corrections you may have.

Created by wolfgang resch on Sat, 21 Jul 2007 (PSF)
Python recipes (4591)
wolfgang resch's recipes (1)

Required Modules

  • (none specified)

Other Information and Tasks