@bybyte #15 下载不能测太多，要不太占资源，也太占 cf 的流量，CloudflareST 这个程序是默认只测延迟最低的 10 个 ip 的下载速度，排序要综合考虑，而不是单纯按下载速度或延迟排序。

CloudflareST 生成测速结果的 result.csv 文件的格式如下
IP 地址,已发送,已接收,丢包率,平均延迟,下载速度 (MB/s)
104.18.17.65,4,4,0.00,173.89,15.90
104.19.139.124,4,4,0.00,156.60,15.80
104.19.244.185,4,4,0.00,156.82,14.50
104.16.120.31,4,4,0.00,157.34,14.30
104.21.233.134,4,4,0.00,157.53,14.21
104.18.27.152,4,4,0.00,158.02,13.20
104.18.29.225,4,4,0.00,158.45,13.12
104.21.228.41,4,4,0.00,159.05,12.50
104.18.125.115,4,4,0.00,159.28,11.50
104.19.132.67,4,4,0.00,160.25,11.00

# 找出上述 ip 的最优 ip 的 python 代码
import pandas as pd
df = pd.read_csv('result.csv') # 读取 result.csv 文件
total_latency = df['平均延迟'].sum()
total_download_speed = df['下载速度 (MB/s)'].sum()
df['Latency_Percentage'] = df['平均延迟'] / total_latency
df['Download_Speed_Percentage'] = -(df['下载速度 (MB/s)'] / total_download_speed)
df['Composite_Score'] = df['Latency_Percentage'] + df['Download_Speed_Percentage']
best_ip_row = df.loc[df['Composite_Score'].idxmin()]
first_ip = best_ip_row['IP 地址']
print(first_ip) # 输出 104.19.139.124