可能没有说清楚，我用networkx构蛋白质图，每个残基作为一个节点，并给每个节点添加了一个19维的特征，将图保存为了pkl文件，每个蛋白质为一张图，现在我想用pyg将图导入，用于下一步图神经网络的训练，请给出导入数据的代码以及模型搭建、训练的示例代码

在PyTorch Geometric (PyG)中处理这种类型的蛋白质图数据通常需要先加载pkl文件，然后创建`DataList`或`Data`对象。这里是一个基本的步骤：

import torch
from torch_geometric.data import Data, DataLoader
from torch.nn import Linear
from torch_geometric.nn import GCNConv
from networkx.readwrite import pickle

# 加载pkl文件
def load_protein_graphs(pkl_file_path):
    graphs = []
    with open(pkl_file_path, 'rb') as f:
        for protein_data in pickle.load(f):
            nodes = protein_data['nodes']  # 节点列表
            edges = protein_data['edges']   # 边列表
            features = torch.tensor(protein_data['features'], dtype=torch.float)  # 19维特征
            graph = Data(x=features, edge_index=edges)  # 创建Data对象
            graphs.append(graph)
    return graphs

# 读取并加载数据
protein_graphs = load_protein_graphs('protein_graphs.pkl')
train_data, val_data, test_data = train_val_test_split(protein_graphs, split_ratio=[0.8, 0.1, 0.1])  # 分割数据集

# 数据加载器
train_loader = DataLoader(train_data, batch_size=32, shuffle=True)
val_loader = DataLoader(val_data, batch_size=32, shuffle=False)
test_loader = DataLoader(test_data, batch_size=32, shuffle=False)

# 模型搭建
class ProteinGraphModel(torch.nn.Module):
    def __init__(self):
        super(ProteinGraphModel, self).__init__()
        self.conv1 = GCNConv(protein_features_dim, 64)  # 假设蛋白特征维度为protein_features_dim
        self.conv2 = GCNConv(64, num_classes)  # num_classes为目标类别数

    def forward(self, data):
        x, edge_index = data.x, data.edge_index
        x = F.relu(self.conv1(x, edge_index))
        x = F.dropout(x, training=self.training)
        x = self.conv2(x, edge_index)
        return F.log_softmax(x, dim=1)

model = ProteinGraphModel()

# 训练示例
optimizer = torch.optim.Adam(model.parameters(), lr=0.01)
criterion = torch.nn.NLLLoss()

for epoch in range(num_epochs):  # 设定总迭代次数num_epochs
    for data in train_loader:
        optimizer.zero_grad()
        out = model(data)
        loss = criterion(out[data.train_mask], data.y[data.train_mask])
        loss.backward()
        optimizer.step()

    # 记录学习过程
    print(f"Epoch {epoch+1}/{num_epochs}, Loss: {loss.item()}")

# 验证和测试阶段
with torch.no_grad():
    model.eval()
    train_acc = accuracy(model, train_loader)
    val_acc = accuracy(model, val_loader)
    test_acc = accuracy(model, test_loader)

print(f"Train Acc: {train_acc}, Val Acc: {val_acc}, Test Acc: {test_acc}")

在这个例子中，你需要替换`protein_features_dim`为实际的19维特征维度。此外，记得根据你的需求调整模型结构、损失函数和优化器参数。别忘了运行完训练后评估模型性能。