import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
from sklearn.model_selection import train_test_split
from sklearn.ensemble import RandomForestClassifier
from sklearn.metrics import classification_report, accuracy_score
from pandas.plotting import register_matplotlib_converters
register_matplotlib_converters()
# 设置随机种子，确保结果可重复
np.random.seed(42)

# 1. 生成时间戳数据
date_range = pd.date_range(start='2024-01-01', periods=60*24*30, freq='T')  # 1周内每分钟的数据

# 2. 生成CPU、内存、电源、温度的随机数据
cpu_usage = np.random.uniform(50, 80, size=len(date_range))  # 50% - 80%
memory_usage = np.random.uniform(4, 16, size=len(date_range))  # 4GB - 16GB
power_voltage = np.random.uniform(215, 225, size=len(date_range))  # 215V - 225V
temperature = np.random.uniform(40, 75, size=len(date_range))  # 40°C - 75°C

# 3. 模拟系统崩溃
# 随机选取一些时间点作为崩溃点
crash_indices = np.random.choice(len(date_range), size=10, replace=False)
is_crash = np.zeros(len(date_range))
is_crash[crash_indices] = 1

# 4. 在崩溃前后，增加CPU、内存、温度的数值，以模拟异常
for crash_index in crash_indices:
    if crash_index > 5:  # 确保索引安全
        cpu_usage[crash_index-5:crash_index+1] = np.random.uniform(90, 100, size=6)  # 崩溃前CPU升高
        memory_usage[crash_index-5:crash_index+1] = np.random.uniform(14, 16, size=6)  # 崩溃前内存占用接近极限
        temperature[crash_index-5:crash_index+1] = np.random.uniform(75, 90, size=6)  # 温度异常升高
        power_voltage[crash_index-5:crash_index+1] = np.random.uniform(210, 230, size=6)  # 电压波动

# 5. 将数据整理成DataFrame
data = pd.DataFrame({
    'timestamp': date_range,
    'cpu_usage': cpu_usage,
    'memory_usage': memory_usage,
    'power_voltage': power_voltage,
    'temperature': temperature,
    'is_crash': is_crash
})

# 将timestamp设置为索引
data.set_index('timestamp', inplace=True)

# 6. 数据预览
print(data.head(10))

# 7. 可视化数据，观察是否有异常
plt.figure(figsize=(15, 10))

plt.subplot(4, 1, 1)
plt.plot(data.index, data['cpu_usage'], label='CPU Usage')
plt.title('CPU Usage Over Time')
plt.ylabel('CPU (%)')

plt.subplot(4, 1, 2)
plt.plot(data.index, data['memory_usage'], label='Memory Usage', color='g')
plt.title('Memory Usage Over Time')
plt.ylabel('Memory (GB)')

plt.subplot(4, 1, 3)
plt.plot(data.index, data['power_voltage'], label='Power Voltage', color='orange')
plt.title('Power Voltage Over Time')
plt.ylabel('Voltage (V)')

plt.subplot(4, 1, 4)
plt.plot(data.index, data['temperature'], label='Temperature', color='red')
plt.title('Temperature Over Time')
plt.ylabel('Temperature (°C)')

plt.tight_layout()
plt.savefig('plot111.png')

# 数据保存为csv文件，方便后续处理
data.to_csv('simulated_system_data.csv')

# 8. 使用随机森林分类器进行系统崩溃分析
# 分离特征和目标变量
X = data[['cpu_usage', 'memory_usage', 'power_voltage', 'temperature']]  # 特征
y = data['is_crash']  # 目标变量

# 划分训练集和测试集
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=42)

# 初始化并训练随机森林分类器
model = RandomForestClassifier(n_estimators=100, random_state=42, class_weight='balanced')
model.fit(X_train, y_train)

# 使用测试集进行预测
y_pred = model.predict(X_test)

# 打印模型评估结果
print("模型准确率:", accuracy_score(y_test, y_pred))
print("分类报告:\n", classification_report(y_test, y_pred, zero_division=1))

# 分析特征重要性
importances = model.feature_importances_
feature_importances = pd.Series(importances, index=X.columns)

# 可视化特征重要性
plt.figure(figsize=(10, 6))
sns.barplot(x=feature_importances, y=feature_importances.index)
plt.title('Feature Importance for System Crash Prediction')
plt.xlabel('Importance')
plt.ylabel('Feature')
plt.savefig('plot.png')