1  #!/usr/bin/env python 
 2   
 3  #  corestats.py (COREy STATS) - Copyright (c) 2006, Corey Goldberg 
 4  #     
 5  #    statistical calculation class 
 6  #    for processing numeric sequences 
 7  # 
 8  #  license: GNU LGPL 
 9  # 
10  #  This library is free software; you can redistribute it and/or 
11  #  modify it under the terms of the GNU Lesser General Public 
12  #  License as published by the Free Software Foundation; either 
13  #  version 2.1 of the License, or (at your option) any later version. 
14   
15   
16  # get fastselect here: http://aspn.activestate.com/ASPN/Cookbook/Python/Recipe/466330 
17  import fastselect 
18  import sys 
19   
20   
21 -class Stats: 
22           
23 -    def __init__(self, sequence): 
24          # a sequence of numbers we will process 
25          # we convert all items to floats for numerical processing         
26          self.sequence = [float(item) for item in sequence] 
27       
28       
29 -    def print_sequence(self): 
30          # print the sequence (newline delimited) 
31          for item in self.sequence: 
32              print item 
33       
34       
35 -    def sum(self): 
36          if len(self.sequence) < 1:  
37              return None 
38          else: 
39              return sum(self.sequence) 
40       
41       
42 -    def count(self): 
43          return len(self.sequence) 
44           
45       
46 -    def min(self): 
47          if len(self.sequence) < 1:  
48              return None 
49          else: 
50              return min(self.sequence) 
51       
52       
53 -    def max(self): 
54          if len(self.sequence) < 1:  
55              return None 
56          else: 
57              return max(self.sequence) 
58       
59   
60 -    def avg(self): 
61          if len(self.sequence) < 1:  
62              return None 
63          else:  
64              return sum(self.sequence) / len(self.sequence)     
65       
66       
67 -    def median(self): 
68          if len(self.sequence) < 1:  
69              return None 
70          else:  
71              return self.sum() // 2 
72               
73       
74 -    def stdev(self): 
75          if len(self.sequence) < 1:  
76              return None 
77          else: 
78              avg = self.avg() 
79              sdsq = sum([(i - avg) ** 2 for i in self.sequence]) 
80              stdev = (sdsq / (len(self.sequence) - 1 or 1)) ** .5 
81              return stdev 
82       
83       
84 -    def percentile(self, percentile): 
85          if len(self.sequence) < 1:  
86              return None 
87          else:  
88              percent = percentile / 100.0 
89              if (percentile >= 100): 
90                  sys.stderr.write('ERROR: percentile must be < 100.  you supplied: %s\n'% percentile) 
91                  sys.exit(1) 
92              # using fastselect algorithm.  see fastselect.py for performance details 
93              return fastselect.select(int(len(self.sequence) * percent), self.sequence) 
94