This code generates ROC curve and PR curve, and also prints AUCs individually.

ROC_PR.py

import math, random
from matplotlib import pyplot as plt

def normal_cdf(x, mu=0, sigma=1):
    return (1 + math.erf((x - mu) / math.sqrt(2) / sigma)) / 2

def inverse_normal_cdf(p, mu=0, sigma=1, tolerance=0.00001):
    if mu != 0 or sigma != 1:
        return mu + sigma * inverse_normal_cdf(p, tolerance=tolerance)

    low_z, low_p = -10.0, 0
    hi_z, hi_p = 10.0, 1
    while hi_z  - low_z > tolerance:
        mid_z = (low_z + hi_z) / 2
        mid_p = normal_cdf(mid_z)

        if mid_p < p:
            low_z, low_p = mid_z, mid_p
        elif mid_p > p:
            hi_z, hi_p = mid_z, mid_p
        else:
            break

    return mid_z

N = 10000 #総サンプル数
ip = 0.1 #有病率
mu1 = 20 #陽性サンプルの平均
s1 = 5 #陽性サンプルの分散
mu2 = 30 #陰性サンプルの平均
s2 = 5 #陰性サンプルの分散

posi = []
nega = []
posis = []
negas = []
x = []

for i in range(int(N*ip)):
    posi.append(int(inverse_normal_cdf(random.random(), mu1, s1)))
    #posi.append(random.randint(10,30))
for i in range(int(N*(1-ip))):
    nega.append(int(inverse_normal_cdf(random.random(), mu2, s2)))
    #nega.append(random.randint(20,40))
for i in range(max(posi)+1):
    posis.append(0)
for i in range(len(posi)):
    posis[posi[i]] += 1
for i in range(max(nega)+1):
    negas.append(0)
for i in range(len(nega)):
    negas[nega[i]] += 1
for i in range(max(len(posis), len(negas))+1):
    x.append(i)
for i in range(len(x)-len(posis)):
    posis.append(0)
for i in range(len(x)-len(negas)):
    negas.append(0)
plt.title("Datas")
plt.xlabel("values")
plt.ylabel("numbers")
plt.bar(x,posis, color="r")
plt.bar(x,negas, color="b")
plt.show()

rx = []
ry = []
px = []
py = []
rauc = 0
pauc = 0

for i in range(len(x)):
    TP = sum(posis[:i+1])
    FN = sum(posis[i+1:])
    TN = sum(negas[i+1:])
    FP = sum(negas[:i+1])
    if TP+FN != 0:
        Recall = TP / (TP + FN)
    else:
        Recall = 1
    if TP+FP != 0:
        Precision = TP / (TP + FP)
    else:
        Precision = 1
    TPR = TP / (TP + FN)
    FPR = FP / (FP + TN)
    rx.append(FPR)
    ry.append(TPR)
    px.append(Recall)
    py.append(Precision)
    if i >= 1:
        rauc += (ry[i]+ry[i-1])*(rx[i]-rx[i-1])/2
        pauc += (py[i]+py[i-1])*(px[i]-px[i-1])/2
plt.title("ROC Curve")
plt.xlabel("False Positive Rate")
plt.ylabel("True Positive Rate")
plt.plot(rx,ry)
plt.show()
plt.title("PR Curve")
plt.xlabel("Recall")
plt.ylabel("Precision")
plt.plot(px,py)
plt.show()
print("AUC-ROC: " + str(rauc))
print("AUC-PR: " + str(pauc))