编写pytorch代码，df读取数据集dataset，通过滑动窗口划分法划分为训练集、验证集和测试集,窗口为30，步长为15，比例为0.7：0.2：0.1

下面是一个示例代码，实现了通过滑动窗口划分法将数据集划分为训练集、验证集和测试集：

import torch
from torch.utils.data import Dataset, DataLoader

class MyDataset(Dataset):
    def __init__(self, data):
        self.data = data

    def __len__(self):
        return len(self.data)

    def __getitem__(self, index):
        # 返回数据的特征和标签
        return self.data[index][:-1], self.data[index][-1]

def sliding_window_split(data, window_size=30, stride=15, train_ratio=0.7, val_ratio=0.2, test_ratio=0.1):
    # 计算数据集的总长度、训练集长度、验证集长度和测试集长度
    data_len = len(data)
    train_len = int((data_len - window_size) * train_ratio // stride) * stride + window_size
    val_len = int((data_len - window_size) * val_ratio // stride) * stride + window_size
    test_len = data_len - train_len - val_len + window_size

    # 划分数据集为训练集、验证集和测试集
    train_data = data[:train_len]
    val_data = data[train_len:train_len+val_len]
    test_data = data[train_len+val_len:]

    # 将训练集、验证集和测试集分别按照窗口大小和步长划分为多个样本
    train_samples = [train_data[i:i+window_size+1] for i in range(0, train_len-window_size+1, stride)]
    val_samples = [val_data[i:i+window_size+1] for i in range(0, val_len-window_size+1, stride)]
    test_samples = [test_data[i:i+window_size+1] for i in range(0, test_len-window_size+1, stride)]

    # 将样本转化为Dataset格式
    train_dataset = MyDataset(train_samples)
    val_dataset = MyDataset(val_samples)
    test_dataset = MyDataset(test_samples)

    return train_dataset, val_dataset, test_dataset

# 读取数据集
df = ...

# 将数据集转化为numpy数组格式
data = df.values.astype(float)

# 划分数据集为训练集、验证集和测试集
train_dataset, val_dataset, test_dataset = sliding_window_split(data, window_size=30, stride=15, train_ratio=0.7, val_ratio=0.2, test_ratio=0.1)

# 创建DataLoader
train_loader = DataLoader(train_dataset, batch_size=128, shuffle=True)
val_loader = DataLoader(val_dataset, batch_size=128, shuffle=False)
test_loader = DataLoader(test_dataset, batch_size=128, shuffle=False)

上述代码首先定义了一个`MyDataset`类，用于将数据集转化为Dataset格式。然后定义了`sliding_window_split`函数，用于按照滑动窗口划分法将数据集划分为训练集、验证集和测试集，并将每个样本转化为Dataset格式。最后，使用PyTorch的DataLoader将训练集、验证集和测试集转化为可迭代的数据加载器。