example6- code python

tes1

Codice per la generazione delle tuple, **6_example.py**  

>Si trova in https://www.pythonanywhere.com/user/cristiano74/files/home/cristiano74/mysite

#### <i class="icon-share"></i> **Results samples**
> In  questo [file][1] si possono vedere i risultati delle elaborazioni. Con vari campioni e differenti popolazioni prodotte.

#### <i class="icon-share"></i> 6_example.py

> Codice python:

```
import itertools;
v=[0,5,10,15,20,25,30,35,40,45,50,55,60,65,70,75,80,85,90,95,100];


testmode=0 # testmode =1 significa VERO

target=100      #somma da raggiungere - test

if testmode==1:
    n_rep=3
    tot= 1000       #numero casi da generare - test
    size=30         #dimensione campione estratto - test
else:
    n_rep=7
    tot= 100000         # numero casi da generare
    size=3000           #dimensione campione estratto

combos = [a for a in itertools.product(v,repeat=n_rep) if sum(a)==target]
#combos = itertools.product(v,repeat=3)
usable_combos = []
lista = []
results= []

#log = open("filename.txt","w")
#import scipy
from scipy.stats import mannwhitneyu
import random
import collections
tuple_tot = 0
acc=0
high = 0
low = 0
med = 0

for e in combos:
        print e     #indica la tupla
        #log = open("filename.txt","w")
        #log = open("filename.txt","a")
        #print >> log, "1m\n"*perc1

        #print "--->tupla:", e[0],e[1],e[2]  #   stampa della tupla
        tuple_tot += 1                      #   per sapere numero totale tuple

        if testmode==1:
            a1=[1]*(e[0]*tot/100)+[2]*(e[1]*tot/100)+[3]*(e[2]*tot/100)  #3 repl
        else:
            a1=[1]*(e[0]*tot/100)+[2]*(e[1]*tot/100)+[3]*(e[2]*tot/100)+[4]*(e[3]*tot/100)+[5]*(e[4]*tot/100)+[6]*(e[5]*tot/100)+[7]*(e[6]*tot/100)  # 7 rep


        #print a1
        #print "lung=",len(a1)

        #print "sum=",sum(a1)
        #print "mean=",numpy.mean(a1)


        random.shuffle(a1)
        a2=random.sample(a1, size)              #estrazione campione
        #print a1
        #print a2
        #print "lung_sample=",len(a2)
        #print "sum_sample=",sum(a2)
        #print "mean_sample=",numpy.mean(a2)


        counter_orig=collections.Counter(a1)
        counter_sample=collections.Counter(a2)
        #print "pop:",counter_orig[1],counter_orig[2],counter_orig[3]
        #print "sample",counter_sample[1],counter_sample[2],counter_sample[3]

        #print "pop%:",float(float(counter_orig[1])/tot*100),float(float(counter_orig[2])/tot*100),float(float(counter_orig[3])/tot*100)
        #print "sample%",float(float(counter_sample[1])/size*100),float(float(counter_sample[2])/size*100),float(float(counter_sample[3])/size*100)
        if testmode==1:
            p1 = round(float(float(counter_orig[1])/tot*100),2)
            p2 = round(float(float(counter_orig[2])/tot*100),2)
            p3 = round(float(float(counter_orig[3])/tot*100),2)
        else:
            p1 = round(float(float(counter_orig[1])/tot*100),2)
            p2 = round(float(float(counter_orig[2])/tot*100),2)
            p3 = round(float(float(counter_orig[3])/tot*100),2)
            p4 = round(float(float(counter_orig[4])/tot*100),2)
            p5 = round(float(float(counter_orig[5])/tot*100),2)
            p6 = round(float(float(counter_orig[6])/tot*100),2)
            p7 = round(float(float(counter_orig[7])/tot*100),2)

        if testmode==1:
            s1 = round(float(float(counter_sample[1])/size*100),2)
            s2 = round(float(float(counter_sample[2])/size*100),2)
            s3 = round(float(float(counter_sample[3])/size*100),2)
        else:
            s1 = round(float(float(counter_sample[1])/size*100),2)
            s2 = round(float(float(counter_sample[2])/size*100),2)
            s3 = round(float(float(counter_sample[3])/size*100),2)
            s4 = round(float(float(counter_sample[4])/size*100),2)
            s5 = round(float(float(counter_sample[5])/size*100),2)
            s6 = round(float(float(counter_sample[6])/size*100),2)
            s7 = round(float(float(counter_sample[7])/size*100),2)

        #TEST Mann-Whitney ######################
        #p_value = 1
        try:
            u, p_value = mannwhitneyu(a1, a2)

            #u, p_value = mannwhitneyu(a1, a2)
            #print "two-sample wilcoxon-test", p_value
            if p_value > 0.40:
                acc += 1
                test = "HIGH precision"
                high += 1
            else:
                if p_value < 0.30:
                    test = "Low precision"
                    low += 1
                else:
                    test = "Medium precision"
                    med += 1
            ###################STANMPA RISULTATI
            #results = p1,p2,p3,tot,s1,s2,s3,len(a2),round(float(p_value),2),test

            #results = p1,p2,p3,p4,p5,p6,p7,tot,s1,s2,s3,s4,s5,s6,s7,len(a2),round(float(p_value),2),test
            # print results
            #####################################



            #print acc
            #print test

            #c1 = counter[0]/size*target
            #c2 = counter[-1]/size*target
            # print c1,c2

            #final = e[0],e[1],e[2],tot,round(numpy.mean(a1),2),len(a2),round(numpy.mean(a2),2)
            #final = e[0],e[1],e[2],e[3],e[4],e[5],e[6],tot,round(numpy.mean(a1),2),len(a2),round(numpy.mean(a2),2)

            #log = open("filename.txt","a")
            #print >> log,final
            #log.close()




            #print "mean=",sum(a1)/tot*1.0
            #lista.append(perc1)
            #print >> log, ("1m\n"*(e[0]*tot/100))[:-1]
            #print >> log, ("2m\n"*(e[1]*tot/100))[:-1]
            #print >> log, ("3m\n"*(e[2]*tot/100))[:-1]
            #print >> log, ("4m\n"*(e[3]*tot/100))[:-1]
            # print >> log, ("5m\n"*(e[4]*tot/100))[:-1]
            #print >> log, ("6m\n"*(e[5]*tot/100))[:-1]
            #print >> log, ("7m\n"*(e[6]*tot/100))[:-1]

            #log.close()
            # perc1 = ("1m\n"*(e[1]*tot/100))[:-1]
            #lista.append(perc1)
            #print lista
            #print perc1
            #usable_combos.append(e)
            #print usable_combos
            #log = open("out_1_test.txt","a")
                #print >> log, usable_combos
            #log.close()
        except:
            pass
#log = open("SAMPLERESULTS.txt","w")
log = open("SAMPLERESULTS.txt","a")
final = "<<<Tuple cha passano il test in high>>>", round(float(float(acc)/tuple_tot*100),2),"%",'pop=', tot, "; campione=", size , "High:",high," Med:",med," Low:",low
print final
print >> log,final
log.close()

```


[1]: https://docs.google.com/a/dalfarra.it/spreadsheet/ccc?key=0AnsGgPg-nkhEdDg5dWFNY2ZONjNtVFo4WWxlWmFuWnc&usp=sharing