import numpy as np

"""
实验说明:
- 自行编写实现数据加载与预测病马死亡的对率回归算法。 

注意事项:
- 在 function.py 中实现对率回归算法。
- 可以自行导入其他库，需要在 requirements.txt 添加依赖，若未在 requirements.txt 中添加使用到的依赖，则无法自动评测，成绩为零。
- 自动评测脚本仅会加载 train_logreg、predict_logreg 和 load_data 函数，请勿修改函数名。

测评环境:
- CPU: Intel Xeon(R) CPU E3-1231 v3 @ 3.40GHz
- OS: Ubuntu 22.04.4 LTS
- Python: 3.10.12
- numpy: 2.1.1
"""

# 请在这里填写学号和姓名，不填写即成绩无效。
id = "202210423118"
name = "刘国庆"


def load_data(filepath):
    """
    从文件中加载数据，若不自行正确实现，则无法通过评测，不记录成绩。
    
    Args:
    - filepath (str): 数据文件路径。
    
    Returns:
    - np.array: 特征数据。
    - np.array: 标签数据。
    """
    for delimiter in [',', ' ', '\t']:  # 尝试不同的分隔符：逗号、空格、制表符
        try:
            # Load data, filling missing values with NaN
            data = np.genfromtxt(filepath, delimiter=delimiter, missing_values='?', filling_values=np.nan)
            
            # Check if the data has been loaded correctly with at least two dimensions
            if data.ndim < 2 or data.shape[1] < 2:  # 至少需要两列：特征和标签
                continue  # 尝试下一个分隔符

            # Remove rows with NaN values
            data = data[~np.isnan(data).any(axis=1)]
            
            # Ensure there's still data left after removing NaN
            if data.size == 0:
                print(f"Error: No valid data after removing NaN rows in file '{filepath}'.")
                return None, None

            # Split into features (X) and labels (y)
            X = data[:, :-1]  # Features
            y = data[:, -1]   # Labels

            # Check if shapes of X and y match
            if X.shape[0] != y.shape[0]:
                print(f"Error: Feature and label count mismatch in file '{filepath}'")
                return None, None
            
            # 输出数据形状供检查
            print(f"Loaded data from '{filepath}': {X.shape[0]} samples, {X.shape[1]} features.")
            
            return X, y
        except Exception as e:
            print(f"Error loading data with delimiter '{delimiter}': {e}")

    print(f"Error: File '{filepath}' may be empty or improperly formatted.")
    return None, None


def sigmoid(z):
    """
    Sigmoid activation function.
    """
    return 1 / (1 + np.exp(-z))


def train_logreg(X, y, learning_rate=0.01, epochs=1000):
    """
    训练逻辑回归模型。
    
    Args:
    - X (np.array): 特征数据。
    - y (np.array): 标签数据。
    - learning_rate (float): 学习率。
    - epochs (int): 训练的迭代次数。
    
    Returns:
    - np.array: 训练后的权重。
    """
    # Initialize weights
    X = np.c_[np.ones(X.shape[0]), X]  # Add intercept term
    weights = np.zeros(X.shape[1])
    
    # Gradient descent for logistic regression
    for epoch in range(epochs):
        z = np.dot(X, weights)
        predictions = sigmoid(z)
        
        # Calculate the gradient
        gradient = np.dot(X.T, (predictions - y)) / y.size
        weights -= learning_rate * gradient
        
    return weights


def predict_logreg(X, weights):
    """
    使用逻辑回归模型进行预测。
    
    Args:
    - X (np.array): 特征数据。
    - weights (np.array): 训练得到的权重。
    
    Returns:
    - np.array: 预测的标签（0 或 1）。
    """
    X = np.c_[np.ones(X.shape[0]), X]  # Add intercept term
    probabilities = sigmoid(np.dot(X, weights))
    return (probabilities >= 0.5).astype(int)