experiments.models.conv

experiments/models/conv.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
import copy

import torch
from torch_geometric.nn import MessagePassing
import torch.nn.functional as F
from ogb.graphproppred.mol_encoder import BondEncoder
from torch_geometric.utils import degree


from typing import Optional
from torch_geometric.typing import OptTensor
from torch.nn import Parameter
from torch_geometric.nn.inits import zeros
from torch_geometric.utils import get_laplacian
from torch_geometric.utils import remove_self_loops, add_self_loops, segregate_self_loops
from torch_sparse import coalesce

import pdb


class GINConvNew(MessagePassing):
    def __init__(self, emb_dim, dataset_group):
        '''
            emb_dim (int): node embedding dimensionality
        '''

        super(GINConvNew, self).__init__(aggr = "add")

        self.mlp = torch.nn.Sequential(torch.nn.Linear(emb_dim, 2 * emb_dim), torch.nn.BatchNorm1d(2 * emb_dim),
                                       torch.nn.ReLU(), torch.nn.Linear(2 * emb_dim, emb_dim))
        self.eps = torch.nn.Parameter(torch.Tensor([0]))

        if dataset_group == 'mol' :
            self.edge_encoder = BondEncoder(emb_dim = emb_dim)
        else :
            self.edge_encoder = torch.nn.Linear(7, emb_dim)

    def forward(self, x, edge_index, edge_attr):
        edge_embedding = self.edge_encoder(edge_attr)

        out = self.mlp((1 + self.eps) *x + self.propagate(edge_index, x=x, edge_attr=edge_embedding))

        return out

    def message(self, x_j, edge_attr):
        return F.relu(x_j + edge_attr)

    def update(self, aggr_out):
        return aggr_out


class GCNConvNew(MessagePassing):
    def __init__(self, emb_dim, dataset_group):
        super(GCNConvNew, self).__init__(aggr='add')

        self.linear = torch.nn.Linear(emb_dim, emb_dim)
        self.root_emb = torch.nn.Embedding(1, emb_dim)

        if dataset_group == 'mol' :
            self.edge_encoder = BondEncoder(emb_dim = emb_dim)
        else :
            self.edge_encoder = torch.nn.Linear(7, emb_dim)

    def forward(self, x, edge_index, edge_attr):
        x = self.linear(x)
        edge_embedding = self.edge_encoder(edge_attr)

        row, col = edge_index

        # edge_weight = torch.ones((edge_index.size(1), ), device=edge_index.device)
        deg = degree(row, x.size(0), dtype=x.dtype) + 1
        deg_inv_sqrt = deg.pow(-0.5)
        deg_inv_sqrt[deg_inv_sqrt == float('inf')] = 0

        norm = deg_inv_sqrt[row] * deg_inv_sqrt[col]

        return self.propagate(edge_index, x=x, edge_attr=edge_embedding, norm=norm) + F.relu(
            x + self.root_emb.weight) * 1. / deg.view(-1, 1)

    def message(self, x_j, edge_attr, norm):
        return norm.view(-1, 1) * F.relu(x_j + edge_attr)

    def update(self, aggr_out):
        return aggr_out


class ChebConvNew(MessagePassing):
    def __init__(self, emb_dim: int, K: int, dataset_group: str,
                 normalization: Optional[str] = 'sym', bias: bool = True,
                 **kwargs):
        kwargs.setdefault('aggr', 'add')
        super(ChebConvNew, self).__init__(**kwargs)

        assert K > 0
        assert normalization in [None, 'sym', 'rw'], 'Invalid normalization'

        self.root_emb = torch.nn.Embedding(1, emb_dim)
        if dataset_group == 'mol' :
            self.edge_encoder = BondEncoder(emb_dim = emb_dim)
        else :
            self.edge_encoder = torch.nn.Linear(7, emb_dim)

        self.emb_dim = emb_dim
        self.normalization = normalization
        self.lins = torch.nn.ModuleList([
            torch.nn.Linear(emb_dim, emb_dim, bias=False) for _ in range(K)
        ])

        if bias:
            self.bias = Parameter(torch.Tensor(emb_dim))
        else:
            self.register_parameter('bias', None)

    def __norm__(self, edge_index, num_nodes: Optional[int],
                 edge_weight: OptTensor, normalization: Optional[str],
                 lambda_max, dtype: Optional[int] = None,
                 batch: OptTensor = None):

        edge_index, edge_weight = remove_self_loops(edge_index, edge_weight)
        edge_index, edge_weight = get_laplacian(edge_index, edge_weight,
                                                normalization, dtype,
                                                num_nodes)

        if batch is not None and lambda_max.numel() > 1:
            lambda_max = lambda_max[batch[edge_index[0]]]

        edge_weight = (2.0 * edge_weight) / lambda_max
        edge_weight.masked_fill_(edge_weight == float('inf'), 0)

        edge_index, edge_weight = add_self_loops(edge_index, edge_weight,
                                                 fill_value=-1.,
                                                 num_nodes=num_nodes)
        assert edge_weight is not None
        return edge_index, edge_weight

    def forward(self, x, edge_index, edge_attr: OptTensor = None,
                batch: OptTensor = None, lambda_max: OptTensor = None):
        """"""
        edge_embedding = self.edge_encoder(edge_attr)

        if self.normalization != 'sym' and lambda_max is None:
            raise ValueError('You need to pass `lambda_max` to `forward() in`'
                             'case the normalization is non-symmetric.')

        if lambda_max is None:
            lambda_max = torch.tensor(2.0, dtype=x.dtype, device=x.device)
        if not isinstance(lambda_max, torch.Tensor):
            lambda_max = torch.tensor(lambda_max, dtype=x.dtype,
                                      device=x.device)
        assert lambda_max is not None

        edge_index, norm = self.__norm__(edge_index, x.size(self.node_dim),
                                         None, self.normalization,
                                         lambda_max, dtype=x.dtype,
                                         batch=batch)

        # edge_index, norm = coalesce(edge_index, norm, m=x.shape[0], n=x.shape[0])
        edge_index, norm, loop_edge_index, loop_norm = segregate_self_loops(edge_index, norm)
        loop_edge_index, loop_norm = coalesce(loop_edge_index, loop_norm, m=x.shape[0], n=x.shape[0])

        Tx_0 = x
        Tx_1 = x  # Dummy.
        out = self.lins[0](Tx_0)

        # propagate_type: (x: Tensor, norm: Tensor)
        if len(self.lins) > 1:
            Tx_1 = self.propagate(edge_index, x=x, edge_attr=edge_embedding, norm=norm, size=None)+\
                   loop_norm.view(-1,1)*F.relu(x + self.root_emb.weight)
            out = out + self.lins[1](Tx_1)

        for lin in self.lins[2:]:
            Tx_2 = self.propagate(edge_index, x=Tx_1, edge_attr=edge_embedding, norm=norm, size=None)+\
                   loop_norm.view(-1,1)*F.relu(Tx_1 + self.root_emb.weight)
            Tx_2 = 2. * Tx_2 - Tx_0
            out = out + lin.forward(Tx_2)
            Tx_0, Tx_1 = Tx_1, Tx_2

        if self.bias is not None:
            out += self.bias

        return out

    def message(self, x_j, edge_attr, norm):
        return norm.view(-1, 1) * F.relu(x_j+edge_attr)