Edited autocorrelation.py to add function to plot correlogram

Edit keytestsuite.py to add new plotcorr function from autocorrelation.py
Edited montecarlopitest.py to split normalization stage from simulation stage
Added file graphs.py to plot graphs of statistical test results

Author: AAK05

autocorrelation.py

@@ -31,6 +31,18 @@ def min_autocorr(data,maxk=100):
     p = graph[kmin]
     return (kmin,p)
 
+def plot_correlogram(data,maxk=100):
+    import numpy as np
+    import matplotlib.pyplot as plt
+    graph = correlogram(data,maxk)
+    x = np.linspace(1,maxk,maxk)
+    y = np.array([graph[i] for i in range(1,maxk+1)])
+    plt.grid()
+    plt.plot(x,y,marker=".",color="b")
+    plt.xlabel("Lag (k)")
+    plt.ylabel("Autocorrelation")
+    #plt.legend()
+    plt.show()
 
 #data = [23.4,24.1,44.4,25.5,29.2,37.1,45.0,61.8,56.7,62.3,30.9,69.6,11.0,56.6,83.4,33.9,75.5,87.3,55.4,95.2]
 #print(autocorrelation(data,2))

graphs.py

@@ -0,0 +1,75 @@
+from autocorrelation import *
+from dig_transition import *
+from freq_analysis import *
+from montecarlopitest import *
+from difftests import *
+import sys
+import json
+import numpy as np
+import matplotlib.pyplot as plt
+from matplotlib.ticker import PercentFormatter
+
+with open("Amber-Spyglass_Encrypted-Enumerated.json","r") as f:
+    cipher = json.load(f)
+
+with open("randorgTRNG3.json","r") as f:
+    trng = json.load(f)
+trng = trng[0:len(cipher)]
+
+def plotcorr(data,maxk=100):
+    plot_correlogram(data,maxk=maxk)
+
+def plotcorrdouble(data1,data2,maxk=100):
+    #import numpy as np
+    #import matplotlib.pyplot as plt
+    graph1 = correlogram(data1,maxk)
+    graph2 = correlogram(data2,maxk)
+    x1 = np.linspace(1,maxk,maxk)
+    y1 = np.array([graph1[i] for i in range(1,maxk+1)])
+    x2 = np.linspace(1,maxk,maxk)
+    y2 = np.array([graph2[i] for i in range(1,maxk+1)])
+    plt.grid()
+    plt.plot(x1,y1,marker=".",color="b",label="Ciphertext Correlogram")
+    plt.plot(x2,y2,marker=".",color="r",label="TRNG Correlogram")
+    plt.xlabel("Lag (k)")
+    plt.ylabel("Autocorrelation")
+    plt.legend()
+    plt.show()
+
+def plotfreq(data):
+    x = np.array(data)
+    plt.hist(x,bins=[0.5+i for i in range(27)],color="b",edgecolor="black",linewidth="1.2",density=True)
+    plt.gca().yaxis.set_major_formatter(PercentFormatter(1))
+    plt.ylabel("% Frequency")
+    plt.xlabel("Character")
+    plt.show()
+
+def plotcarlo(data):
+    data1 = data
+    data2 = data1[::-1]
+    datax1 = []
+    datax2 = []
+    for i in range(3*(len(data1)//3)):
+        if i%3 == 0:
+            n = 26*26*data[i] + 26*data[i+1] + data[i+2] #+ data[i+3] + data[i+4]
+            datax1.append(int(n))
+    for i in range(3*(len(data2)//3)):
+        if i%3 == 0:
+            n = 26*26*(data2[i]) + 26*(data2[i+1]) + (data2[i+2]) #+ (data2[i+3]) + (data2[i+4])
+            datax2.append(int(n))
+    x,y = normalizedcarlo(datax1,datax2)
+    figure,axes = plt.subplots()
+    plt.scatter(x,y,s=0.15,color="r",linewidths=0)
+    plt.xlim(0,1)
+    plt.ylim(0,1)
+    c = plt.Circle((0.5,0.5),0.5,fill=False,edgecolor="black",linewidth=0.5,visible=True)
+    axes.add_artist(c)
+    plt.show()
+    print(MonteCarloPi(x,y))
+
+
+if __name__ == "__main__":
+    #plotcorr(trng)
+    #plotcorrdouble(cipher,trng)
+    #plotfreq(cipher)
+    plotcarlo(trng)

keytestsuite.py

@@ -15,7 +15,7 @@
         data = [x for x in str(f.readline())]
     print(len(data))
 
-data1 = data#[0:500000]
+data1 = data[0:664044]
 data2 = data1[::-1]
 
 #datax and datax2 adds together 1st and 2nd, 3rd and 4th, etc
@@ -71,6 +71,7 @@
 
 print("Maximum autocorrelation (k,autocorrelation): " + str(max_autocorr(data1,20)))
 print("Minimum autocorrelation (k,autocorrelation): " + str(min_autocorr(data1,20)))
+#plot_correlogram(data1,100)
 #print("Maximum pattern error from expected (pattern,%): " + str(max_pattern(data1)))
 #print("Minimum pattern error from expected (pattern,%): " + str(min_pattern(data1)))
 print("Max diff error (diff,%): " + str(max_diff_error(data1)))

montecarlopitest.py

@@ -5,7 +5,7 @@ def distancetocentre(x,y,centerx=0.5,centery=0.5):
     dist = ((x-centerx)**2 + (y-centery)**2)**0.5
     return dist
 
-def MonteCarloPi(datax,datay=None):
+def normalizedcarlo(datax,datay=None):
     x = np.array(datax).reshape(-1,1)
     scaler = preprocessing.MinMaxScaler()
     normalizedx = scaler.fit_transform(x)
@@ -21,17 +21,22 @@ def MonteCarloPi(datax,datay=None):
     normalizedx = normalizedx[0]
     normalizedy = normalizedy.tolist()
     normalizedy = normalizedy[0]
+    return normalizedx,normalizedy
+
+def MonteCarloPi(datax,datay=None):
+    x,y = normalizedcarlo(datax,datay)
     n=0
     inside = 0
     outside = 0
-    for i in normalizedx:
-        dist = distancetocentre(i,normalizedy[n])
+    for i in x:
+        dist = distancetocentre(i,y[n])
         n+=1
         if dist <= 0.5:
             inside += 1
         else:
             outside += 1
     pi = 4*(inside/(inside+outside))
+    print(inside,outside+inside)
     return pi
 
 """from RNG_div_all_Coprime import DivAllCoPrimeRNG