Code Lab - 2

pip install torch-scatter -f https://pytorch-geometric.com/whl/torch-1.10.2+cu102.html
pip install torch-sparse -f https://pytorch-geometric.com/whl/torch-1.10.2+cu102.html
pip install torch-geometric
pip install ogb

---

1. PyG Datasets

PyG有两个类，用于存储图以及将图转换为Tensor格式
torch_geometry.datasets 包含各种常见的图形数据集
torch_geometric.data 提供Tensor的图数据处理

1.1 从torch_geometric.datasets中读取数据集

# 每一个dataset都是多张图的list，单张图的类型为torch_geometric.data.Data
import torch
import os
from torch_geometric.datasets import TUDataset

root = './enzymes'
name = 'ENZYMES'

# The ENZYMES dataset
pyg_dataset= TUDataset('./enzymes', 'ENZYMES')

# You can find that there are 600 graphs in this dataset
print(pyg_dataset)
print(type(pyg_dataset))

# 对于第一张图
# print(pyg_dataset[0])
# print(pyg_dataset[0].num_nodes)
# print(pyg_dataset[0].edge_index)
# print(pyg_dataset[0].x)
# print(pyg_dataset[0].y)

1.2 ENZYMES数据集中类别数量和特征维度

# num_classes
def get_num_classes(pyg_dataset):
    num_classes = pyg_dataset.num_classes
    return num_classes
# num_features
def get_num_features(pyg_dataset):
    num_features = pyg_dataset.num_node_features
    return num_features

num_classes = get_num_classes(pyg_dataset)
num_features = get_num_features(pyg_dataset)
print("{} dataset has {} classes".format(name, num_classes))
print("{} dataset has {} features".format(name, num_features))

1.3 ENZYMES数据集中第idx张图的label和边的数量

def get_graph_class(pyg_dataset, idx):
    # y就是这张图的类别
    label = pyg_dataset[idx].y.item()
    return label

graph_0 = pyg_dataset[0]
print(graph_0)
idx = 100
label = get_graph_class(pyg_dataset, idx)
print('Graph with index {} has label {}'.format(idx, label))


def get_graph_num_edges(pyg_dataset, idx):
    num_edges = pyg_dataset[idx].num_edges/2 # 无向图
    return num_edges

idx = 200
num_edges = get_graph_num_edges(pyg_dataset, idx)
print('Graph with index {} has {} edges'.format(idx, num_edges))

---

2. Open Graph Benchmark (OGB)

OGB是基准数据集的集合，使用OGB数据加载器自动下载、处理和拆分，通过OGB Evaluator以统一的方式来评估模型性能

2.1 读取OBG的数据集（以ogbn-arxiv为例）

import ogb
import torch_geometric.transforms as T
from ogb.nodeproppred import PygNodePropPredDataset

dataset_name = 'ogbn-arxiv'
# 加载ogbn-arxiv数据集，并使用ToSparseTensor转换成Tensor格式
dataset = PygNodePropPredDataset(name=dataset_name,
                                 transform=T.ToSparseTensor())
print('The {} dataset has {} graph'.format(dataset_name, len(dataset)))

# 第一张图
data = dataset[0]
print(data)

def graph_num_features(data):
    num_features=data.num_features
    return num_features

num_features = graph_num_features(data)
print('The graph has {} features'.format(num_features))

---

3. GNN节点属性预测（节点分类）（以ogbn-arxiv数据集为例）

3.1 加载并预处理数据集

import torch
import pandas as pd
import torch.nn.functional as F
from torch_geometric.nn import GCNConv
import torch_geometric.transforms as T
from ogb.nodeproppred import PygNodePropPredDataset, Evaluator

device = 'cuda' if torch.cuda.is_available() else 'cpu'
dataset_name = 'ogbn-arxiv'
# PygNodePropPredDataset读取节点分类的数据集
dataset = PygNodePropPredDataset(name=dataset_name,
                              transform=T.ToSparseTensor())
data = dataset[0]

# 转换成稀疏矩阵
data.adj_t = data.adj_t.to_symmetric()
print(type(data.adj_t))

data = data.to(device)
# 用get_idx_split划分数据集为train，valid，test三部分
split_idx = dataset.get_idx_split()
print(split_idx)
train_idx = split_idx['train'].to(device)

3.2 GCN Model

class GCN(torch.nn.Module):
    
## Note:
## 1. You should use torch.nn.ModuleList for self.convs and self.bns
## 2. self.convs has num_layers GCNConv layers
## 3. self.bns has num_layers - 1 BatchNorm1d layers
## 4. You should use torch.nn.LogSoftmax for self.softmax
## 5. The parameters you can set for GCNConv include 'in_channels' and 
## 'out_channels'. For more information please refer to the documentation:
## https://pytorch-geometric.readthedocs.io/en/latest/modules/nn.html#torch_geometric.nn.conv.GCNConv
## 6. The only parameter you need to set for BatchNorm1d is 'num_features'

    def __init__(self, input_dim, hidden_dim, output_dim, num_layers,
                 dropout, return_embeds=False):
        
        super(GCN, self).__init__()

        # A list of GCNConv layers
        self.convs = torch.nn.ModuleList()
        for i in range(num_layers - 1):
            self.convs.append(GCNConv(input_dim, hidden_dim))
            input_dim = hidden_dim
        self.convs.append(GCNConv(hidden_dim, output_dim))
        
        # A list of 1D batch normalization layers
        self.bns = torch.nn.ModuleList()
        for i in range(num_layers - 1):
            self.convs.append(torch.nn.BatchNorm1d(hidden_dim))

        # The log softmax layer
        self.softmax = torch.nn.LogSoftmax()

        # Probability of an element getting zeroed
        self.dropout = dropout

        # Skip classification layer and return node embeddings
        self.return_embeds = return_embeds

    def reset_parameters(self):
        for conv in self.convs:
            conv.reset_parameters()
        for bn in self.bns:
            bn.reset_parameters()

## Note:
## 1. Construct the network as shown in the figure
## 2. torch.nn.functional.relu and torch.nn.functional.dropout are useful
## For more information please refer to the documentation:
## https://pytorch.org/docs/stable/nn.functional.html
## 3. Don't forget to set F.dropout training to self.training
## 4. If return_embeds is True, then skip the last softmax layer

    def forward(self, x, adj_t):
        for layer in range(len(self.convs)-1):
            x=self.convs[layer](x,adj_t)
            x=self.bns[layer](x)
            x=F.relu(x)
            x=F.dropout(x,self.dropout,self.training)
        out=self.convs[-1](x,adj_t)
        if not self.return_embeds:
            out=self.softmax(out)
        return out

3.3 训练和评估

def train(model, data, train_idx, optimizer, loss_fn):
    model.train()
    optimizer.zero_grad()
    out = model(data.x,data.adj_t)
    # 计算训练部分的loss
    train_output = out[train_idx]
    train_label = data.y[train_idx,0]
    loss = loss_fn(train_output,train_label)
    loss.backward()
    optimizer.step()

    return loss.item()

# 注意data.y[train_idx]和data.y[train_idx,0]的区别
print(data.y[train_idx])
print(data.y[train_idx,0])

def test(model, data, split_idx, evaluator, save_model_results=False):
    model.eval()
    out = model(data.x,data.adj_t)
    y_pred = out.argmax(dim=-1, keepdim=True)
    # 使用OGB Evaluator进行评估
    train_acc = evaluator.eval({
        'y_true': data.y[split_idx['train']],
        'y_pred': y_pred[split_idx['train']],
    })['acc']
    valid_acc = evaluator.eval({
        'y_true': data.y[split_idx['valid']],
        'y_pred': y_pred[split_idx['valid']],
    })['acc']
    test_acc = evaluator.eval({
        'y_true': data.y[split_idx['test']],
        'y_pred': y_pred[split_idx['test']],
    })['acc']

    return train_acc, valid_acc, test_acc

args = {
    'device': device,
    'num_layers': 3,
    'hidden_dim': 256,
    'dropout': 0.5,
    'lr': 0.01,
    'epochs': 100,
}
model = GCN(data.num_features, args['hidden_dim'],
            dataset.num_classes, args['num_layers'],
            args['dropout']).to(device)
evaluator = Evaluator(name='ogbn-arxiv')


# 初始化模型参数
model.reset_parameters()

optimizer = torch.optim.Adam(model.parameters(), lr=args['lr'])
loss_fn = F.nll_loss

for epoch in range(1, args["epochs"]+1):
    loss = train(model, data, train_idx, optimizer, loss_fn)
    train_acc, valid_acc, test_acc = test(model, data, split_idx, evaluator)
    print(f'Epoch: {epoch:02d}, '
        f'Loss: {loss:.4f}, '
        f'Train: {100 * train_acc:.2f}%, '
        f'Valid: {100 * valid_acc:.2f}% '
        f'Test: {100 * test_acc:.2f}%')

---

4. GNN图属性预测（图分类）（以ogbn-arxiv数据集为例）

4.1 加载并预处理数据集

from ogb.graphproppred import PygGraphPropPredDataset, Evaluator
from torch_geometric.data import DataLoader
from tqdm.notebook import tqdm

# PygGraphPropPredDataset读取图分类的数据集
dataset = PygGraphPropPredDataset(name='ogbg-molhiv')
split_idx = dataset.get_idx_split()

# DataLoader
train_loader = DataLoader(dataset[split_idx["train"]], batch_size=32, shuffle=True, num_workers=0)
valid_loader = DataLoader(dataset[split_idx["valid"]], batch_size=32, shuffle=False, num_workers=0)
test_loader = DataLoader(dataset[split_idx["test"]], batch_size=32, shuffle=False, num_workers=0)

4.2 GCN Model

from ogb.graphproppred.mol_encoder import AtomEncoder
from torch_geometric.nn import global_add_pool, global_mean_pool

class GCN_Graph(torch.nn.Module):
    def __init__(self, hidden_dim, output_dim, num_layers, dropout):
        super(GCN_Graph, self).__init__()

        # encoders
        self.node_encoder = AtomEncoder(hidden_dim)

        # 通过GCN
        self.gnn_node = GCN(hidden_dim, hidden_dim,
            hidden_dim, num_layers, dropout, return_embeds=True)
        
        # 全局池化
        self.pool = global_mean_pool
        self.linear = torch.nn.Linear(hidden_dim, output_dim)


    def reset_parameters(self):
        self.gnn_node.reset_parameters()
        self.linear.reset_parameters()

    def forward(self, batched_data):
        x, edge_index, batch = batched_data.x, batched_data.edge_index, batched_data.batch
        embed = self.node_encoder(x)
        
        out = self.gnn_node(embed,edge_index)
        out = self.pool(out,batch)
        out = self.linear(out)
        return out

4.3 训练和评估

def train(model, device, data_loader, optimizer, loss_fn):
    model.train()
    for step, batch in enumerate(tqdm(data_loader, desc="Iteration")):
        batch = batch.to(device)
        # 跳过不完整的batch
        if batch.x.shape[0] == 1 or batch.batch[-1] == 0:
            pass
        else:
            # 过滤掉无标签的数据
            is_labeled = (batch.y == batch.y)
            optimizer.zero_grad()
            op = model(batch)
            train_op = op[is_labeled]
            train_labels = batch.y[is_labeled].view(-1)
            loss = loss_fn(train_op.float(),train_labels.float())
            loss.backward()
            optimizer.step()

    return loss.item()

def eval(model, device, loader, evaluator):
    model.eval()
    y_true = []
    y_pred = []
    for step, batch in enumerate(tqdm(loader, desc="Iteration")):
        batch = batch.to(device)
        if batch.x.shape[0] == 1:
            pass
        else:
            with torch.no_grad():
                pred = model(batch)
            y_true.append(batch.y.view(pred.shape).detach().cpu())
            y_pred.append(pred.detach().cpu())

    y_true = torch.cat(y_true, dim = 0).numpy()
    y_pred = torch.cat(y_pred, dim = 0).numpy()
    input_dict = {"y_true": y_true, "y_pred": y_pred}
    result = evaluator.eval(input_dict)
    return result

args = {
    'device': device,
    'num_layers': 5,
    'hidden_dim': 256,
    'dropout': 0.5,
    'lr': 0.001,
    'epochs': 30,
}

model = GCN_Graph(args['hidden_dim'],
            dataset.num_tasks, args['num_layers'],
            args['dropout']).to(device)
evaluator = Evaluator(name='ogbg-molhiv')

model.reset_parameters()
optimizer = torch.optim.Adam(model.parameters(), lr=args['lr'])
loss_fn = torch.nn.BCEWithLogitsLoss()


for epoch in range(1, 1 + args["epochs"]):
    print('Training...')
    loss = train(model, device, train_loader, optimizer, loss_fn)

    print('Evaluating...')
    train_result = eval(model, device, train_loader, evaluator)
    val_result = eval(model, device, valid_loader, evaluator)
    test_result = eval(model, device, test_loader, evaluator)

    train_acc, valid_acc, test_acc = train_result[dataset.eval_metric], val_result[dataset.eval_metric], test_result[dataset.eval_metric]
    print(f'Epoch: {epoch:02d}, '
        f'Loss: {loss:.4f}, '
        f'Train: {100 * train_acc:.2f}%, '
        f'Valid: {100 * valid_acc:.2f}% '
        f'Test: {100 * test_acc:.2f}%')