def forward(self, gl, gr):
gl = gl.local_var()
gr = gr.local_var()
xl = gl.ndata["feat"]
xr = gr.ndata["feat"]
for idx in range(self.depth):
mu_lr, mu_rl = self.attention(xl, xr)
gl.ndata["x"] = xl
gr.ndata["x"] = xr
gl.update_all(dgl.function.copy_src(src='x', out='m'),
dgl.function.sum(msg='m', out='x'))
gr.update_all(dgl.function.copy_src(src='x', out='m'),
dgl.function.sum(msg='m', out='x'))
xl = gl.ndata["x"]
xr = gr.ndata["x"]
xl = torch.cat([xl, mu_rl], dim=-1)
xr = torch.cat([xr, mu_lr], dim=-1)
xl = getattr(self, "ff%s" % idx)(xl)
xr = getattr(self, "ff%s" % idx)(xr)
gl.ndata["x"] = xl
gr.ndata["x"] = xr
xl = dgl.sum_nodes(gl, "x")
xr = dgl.sum_nodes(gr, "x")
x = self.ff(torch.cat([xl, xr], dim=-1))
return x
def forward(self, graph, node_feature, qs, ally_node_type_index=NODE_ALLY):
assert isinstance(graph, dgl.BatchedDGLGraph)
w_emb = self.w_gn(graph, node_feature) # [# nodes x # node_dim]
w = torch.abs(self.w_ff(graph, w_emb)) # [# nodes x # 1]
ally_node_indices = get_filtered_node_index_by_type(
graph, ally_node_type_index)
device = w_emb.device
_qs = torch.zeros(size=(graph.number_of_nodes(), 1), device=device)
w = w[ally_node_indices, :] # [# allies x 1]
_qs[ally_node_indices, :] = w * qs.view(-1, 1)
graph.ndata['node_feature'] = _qs
q_tot = dgl.sum_nodes(graph, 'node_feature')
_ = graph.ndata.pop('node_feature')
v_emb = self.v_gn(graph, node_feature) # [# nodes x # node_dim]
v = self.v_ff(graph, v_emb) # [# nodes x # 1]
v = v[ally_node_indices, :] # [# allies x 1]
_v = torch.zeros(size=(graph.number_of_nodes(), 1), device=device)
_v[ally_node_indices, :] = v
graph.ndata['node_feature'] = _v
v = dgl.sum_nodes(graph, 'node_feature')
_ = graph.ndata.pop('node_feature')
q_tot = q_tot + v
return q_tot.view(-1)
def forward(self, g, h, e, snorm_n, snorm_e):
# modified dtype for new dataset
h = h.float()
h = self.embedding_lin(h.cuda())
h_in = h # for residual connection
# list of hidden representation at each layer (including input)
hidden_rep = [h]
for i in range(self.n_layers):
h = self.ginlayers[i](g, h, snorm_n)
# Residual Connection
if self.residual:
if self.residual == "gated":
z = torch.sigmoid(self.W_g(torch.cat([h, h_in], dim=1)))
h = z * h + (torch.ones_like(z) - z) * h_in
else:
h += h_in
g.ndata['h'] = self.linear_ro(h)
if self.readout == "sum":
hg = dgl.sum_nodes(g, 'h')
elif self.readout == "max":
hg = dgl.max_nodes(g, 'h')
elif self.readout == "mean":
hg = dgl.mean_nodes(g, 'h')
else:
hg = dgl.sum_nodes(g, 'h') # default readout is summation
score = self.linear_prediction(hg)
return score
def forward(self, g, h, e, snorm_n, snorm_e):
# modified dtype for new dataset
h = h.float()
h = self.embedding_lin(h)
h = self.in_feat_dropout(h)
for conv in self.layers:
h_in = h
h = conv(g, h, snorm_n)
if self.residual:
if self.residual == "gated":
z = torch.sigmoid(self.W_g(torch.cat([h, h_in], dim=1)))
h = z * h + (torch.ones_like(z) - z) * h_in
else:
h += h_in
g.ndata['h'] = self.linear_ro(h)
if self.readout == "sum":
hg = dgl.sum_nodes(g, 'h')
elif self.readout == "max":
hg = dgl.max_nodes(g, 'h')
elif self.readout == "mean":
hg = dgl.mean_nodes(g, 'h')
else:
hg = dgl.sum_nodes(g, 'h') # default readout is summation
return self.linear_predict(hg)
def forward(self, g):
g.nodes['word'].data['feat'] = self.dropout(
self.word_embedding(g.nodes['word'].data['x']))
g.nodes['concept'].data['feat'] = self.dropout(
self.concept_embedding(g.nodes['concept'].data['x']))
g.edges['A'].data['weight'] = self.dropout(
self.w_w_embedding(g.edges['A'].data['h']))
g.edges['B'].data['weight'] = self.dropout(
self.w_c_embedding(g.edges['B'].data['h']))
g.edges['C'].data['weight'] = self.dropout(
self.c_w_embedding(g.edges['C'].data['h']))
# g.nodes['word'].data['feat'] = self.word_embedding(g.nodes['word'].data['x'])
# g.nodes['concept'].data['feat'] = self.concept_embedding(g.nodes['concept'].data['x'])
# g.edges['A'].data['weight'] = self.w_w_embedding(g.edges['A'].data['h'])
# g.edges['B'].data['weight'] = self.w_c_embedding(g.edges['B'].data['h'])
# g.edges['C'].data['weight'] = self.c_w_embedding(g.edges['C'].data['h'])
h = g.ndata['feat']
h = self.rgcn(g, h)
with g.local_scope():
g.ndata['h'] = h
# Calculate graph representation by average readout.
# hg = 0
hg = torch.cat((dgl.sum_nodes(
g, 'h', ntype='word'), dgl.sum_nodes(g, 'h', ntype='concept')),
-1)
# for ntype in g.ntypes:
# hg = hg + dgl.mean_nodes(g, 'h', ntype=ntype)
# r = torch.cat(dgl.sum_nodes(g, 'h', ntype=ntype))
return self.classify(hg)
def get_q(self, graph, node_feature, qs, ws=None):
device = node_feature.device
ally_indices = get_filtered_node_index_by_type(graph, NODE_ALLY)
# compute weighted sum of qs
if ws is None:
ws = self.get_w(graph, node_feature) # [#. allies x #. clusters]
weighted_q = qs.view(-1, 1) * ws # [#. allies x #. clusters]
qs = torch.zeros(size=(graph.number_of_nodes(), self.num_clusters), device=device)
qs[ally_indices, :] = weighted_q
graph.ndata['q'] = qs
q_aggregated = dgl.sum_nodes(graph, 'q') # [#. graph x #. clusters]
# compute state_dependent_bias
graph.ndata['node_feature'] = node_feature
sum_node_feature = dgl.sum_nodes(graph, 'node_feature') # [#. graph x feature dim]
q_v = self.q_b_net(sum_node_feature) # [#. graph x #. clusters]
_ = graph.ndata.pop('node_feature')
_ = graph.ndata.pop('q')
q_aggregated = q_aggregated + q_v
return q_aggregated # [#. graph x #. clusters]
def test_simple_readout():
g1 = dgl.DGLGraph()
g1.add_nodes(3)
g2 = dgl.DGLGraph()
g2.add_nodes(4) # no edges
g1.add_edges([0, 1, 2], [2, 0, 1])
n1 = F.randn((3, 5))
n2 = F.randn((4, 5))
e1 = F.randn((3, 5))
s1 = F.sum(n1, 0) # node sums
s2 = F.sum(n2, 0)
se1 = F.sum(e1, 0) # edge sums
m1 = F.mean(n1, 0) # node means
m2 = F.mean(n2, 0)
me1 = F.mean(e1, 0) # edge means
w1 = F.randn((3, ))
w2 = F.randn((4, ))
max1 = F.max(n1, 0)
max2 = F.max(n2, 0)
maxe1 = F.max(e1, 0)
ws1 = F.sum(n1 * F.unsqueeze(w1, 1), 0)
ws2 = F.sum(n2 * F.unsqueeze(w2, 1), 0)
wm1 = F.sum(n1 * F.unsqueeze(w1, 1), 0) / F.sum(F.unsqueeze(w1, 1), 0)
wm2 = F.sum(n2 * F.unsqueeze(w2, 1), 0) / F.sum(F.unsqueeze(w2, 1), 0)
g1.ndata['x'] = n1
g2.ndata['x'] = n2
g1.ndata['w'] = w1
g2.ndata['w'] = w2
g1.edata['x'] = e1
assert F.allclose(dgl.sum_nodes(g1, 'x'), s1)
assert F.allclose(dgl.sum_nodes(g1, 'x', 'w'), ws1)
assert F.allclose(dgl.sum_edges(g1, 'x'), se1)
assert F.allclose(dgl.mean_nodes(g1, 'x'), m1)
assert F.allclose(dgl.mean_nodes(g1, 'x', 'w'), wm1)
assert F.allclose(dgl.mean_edges(g1, 'x'), me1)
assert F.allclose(dgl.max_nodes(g1, 'x'), max1)
assert F.allclose(dgl.max_edges(g1, 'x'), maxe1)
g = dgl.batch([g1, g2])
s = dgl.sum_nodes(g, 'x')
m = dgl.mean_nodes(g, 'x')
max_bg = dgl.max_nodes(g, 'x')
assert F.allclose(s, F.stack([s1, s2], 0))
assert F.allclose(m, F.stack([m1, m2], 0))
assert F.allclose(max_bg, F.stack([max1, max2], 0))
ws = dgl.sum_nodes(g, 'x', 'w')
wm = dgl.mean_nodes(g, 'x', 'w')
assert F.allclose(ws, F.stack([ws1, ws2], 0))
assert F.allclose(wm, F.stack([wm1, wm2], 0))
s = dgl.sum_edges(g, 'x')
m = dgl.mean_edges(g, 'x')
max_bg_e = dgl.max_edges(g, 'x')
assert F.allclose(s, F.stack([se1, F.zeros(5)], 0))
assert F.allclose(m, F.stack([me1, F.zeros(5)], 0))
assert F.allclose(max_bg_e, F.stack([maxe1, F.zeros(5)], 0))
def forward(self, graph, node_feature, sub_q_tots):
graph.ndata['node_feature'] = node_feature
device = node_feature.device
len_groups = []
q_tot = torch.zeros(graph.batch_size, device=device)
for i, sub_q_tot in enumerate(sub_q_tots):
node_indices = get_filtered_node_index_by_assignment(graph, i)
len_groups.append(len(node_indices))
mask = torch.zeros(size=(node_feature.shape[0], 1), device=device)
mask[node_indices, :] = 1
graph.ndata[
'masked_node_feature'] = graph.ndata['node_feature'] * mask
w_input = dgl.sum_nodes(graph, 'masked_node_feature')
if self.rectifier == 'abs':
q_tot = q_tot + torch.abs(self.w(w_input)).view(-1) * sub_q_tot
elif self.rectifier == 'softplus':
q_tot = q_tot + F.softplus(
self.w(w_input)).view(-1) * sub_q_tot
else:
raise RuntimeError("Not implemented rectifier")
_ = graph.ndata.pop('masked_node_feature')
# testing
_ = graph.ndata.pop('node_feature')
ally_indices = get_filtered_node_index_by_type(graph, NODE_ALLY)
_v = torch.zeros(size=(graph.number_of_nodes(), node_feature.shape[1]),
device=device)
_v[ally_indices, :] = node_feature[ally_indices, :]
graph.ndata['node_feature'] = _v
# testing
v = self.v(dgl.sum_nodes(graph, 'node_feature')).view(-1)
q_tot = q_tot + v
_ = graph.ndata.pop('node_feature')
len_groups = np.array(len_groups)
ratio_groups = len_groups / np.sum(len_groups)
print("Num elements in groups {}".format(len_groups))
print("Num elements ratio {}".format(ratio_groups))
ally_indices = get_filtered_node_index_by_type(graph, NODE_ALLY)
target_assignment_weight = graph.ndata['normalized_score'][
ally_indices]
print("Average normalized scores {}".format(
target_assignment_weight.mean(0)))
return q_tot
def _take_action(self, action):
undecided = self.x == 2
self.x[undecided] = action[undecided]
self.t += 1
x1 = (self.x == 1)
self.g = self.g.to(self.device)
self.g.ndata['h'] = x1.float()
self.g.update_all(fn.copy_src(src='h', out='m'),
fn.sum(msg='m', out='h'))
x1_deg = self.g.ndata.pop('h')
## forgive clashing
clashed = x1 & (x1_deg > 0)
self.x[clashed] = 2
x1_deg[clashed] = 0
# graph clean up
still_undecided = (self.x == 2)
self.x[still_undecided & (x1_deg > 0)] = 0
# fill timeout with zeros
still_undecided = (self.x == 2)
timeout = (self.t == self.max_epi_t)
self.x[still_undecided & timeout] = 0
done = self._check_done()
self.epi_t[~done] += 1
# compute reward and solution
x1 = (self.x == 1).float()
node_sol = x1
h = node_sol
self.g.ndata['h'] = h
next_sol = dgl.sum_nodes(self.g, 'h')
self.g.ndata.pop('h')
reward = (next_sol - self.sol)
if self.hamming_reward_coef > 0.0 and self.num_samples == 2:
xl, xr = self.x.split(1, dim=1)
undecidedl, undecidedr = undecided.split(1, dim=1)
hamming_d = torch.abs(xl.float() - xr.float())
hamming_d[(xl == 2) | (xr == 2)] = 0.0
hamming_d[~undecidedl & ~undecidedr] = 0.0
self.g.ndata['h'] = hamming_d
hamming_reward = dgl.sum_nodes(self.g, 'h').expand_as(reward)
self.g.ndata.pop('h')
reward += self.hamming_reward_coef * hamming_reward
reward /= self.max_num_nodes
return reward, next_sol, done
def forward(self, g, pos):
normalizer = torch.tensor(g.batch_num_nodes).unsqueeze_(1).float().to(
pos.device)
g.ndata['a_gp'] = (pos == 0).float()
gp_embed = dgl.sum_nodes(g, 'h', 'a_gp') / normalizer
g.ndata['a_p'] = (pos == 1).float()
p_embed = dgl.mean_nodes(g, 'h', 'a_p')
g.ndata['a_sib'] = (pos == 2).float()
sib_embed = dgl.sum_nodes(g, 'h', 'a_sib') / normalizer
return torch.cat((gp_embed, p_embed, sib_embed), 1)
def test_sum_case1(idtype):
# NOTE: If you want to update this test case, remember to update the docstring
# example too!!!
g1 = dgl.graph(([0, 1], [1, 0]), idtype=idtype, device=F.ctx())
g1.ndata['h'] = F.tensor([1., 2.])
g2 = dgl.graph(([0, 1], [1, 2]), idtype=idtype, device=F.ctx())
g2.ndata['h'] = F.tensor([1., 2., 3.])
bg = dgl.batch([g1, g2])
bg.ndata['w'] = F.tensor([.1, .2, .1, .5, .2])
assert F.allclose(F.tensor([3.]), dgl.sum_nodes(g1, 'h'))
assert F.allclose(F.tensor([3., 6.]), dgl.sum_nodes(bg, 'h'))
assert F.allclose(F.tensor([.5, 1.7]), dgl.sum_nodes(bg, 'h', 'w'))
def test_simple_readout():
g1 = dgl.DGLGraph()
g1.add_nodes(3)
g2 = dgl.DGLGraph()
g2.add_nodes(4) # no edges
g1.add_edges([0, 1, 2], [2, 0, 1])
n1 = th.randn(3, 5)
n2 = th.randn(4, 5)
e1 = th.randn(3, 5)
s1 = n1.sum(0) # node sums
s2 = n2.sum(0)
se1 = e1.sum(0) # edge sums
m1 = n1.mean(0) # node means
m2 = n2.mean(0)
me1 = e1.mean(0) # edge means
w1 = th.randn(3)
w2 = th.randn(4)
ws1 = (n1 * w1[:, None]).sum(0) # weighted node sums
ws2 = (n2 * w2[:, None]).sum(0)
wm1 = (n1 * w1[:, None]).sum(0) / w1[:, None].sum(0) # weighted node means
wm2 = (n2 * w2[:, None]).sum(0) / w2[:, None].sum(0)
g1.ndata['x'] = n1
g2.ndata['x'] = n2
g1.ndata['w'] = w1
g2.ndata['w'] = w2
g1.edata['x'] = e1
assert U.allclose(dgl.sum_nodes(g1, 'x'), s1)
assert U.allclose(dgl.sum_nodes(g1, 'x', 'w'), ws1)
assert U.allclose(dgl.sum_edges(g1, 'x'), se1)
assert U.allclose(dgl.mean_nodes(g1, 'x'), m1)
assert U.allclose(dgl.mean_nodes(g1, 'x', 'w'), wm1)
assert U.allclose(dgl.mean_edges(g1, 'x'), me1)
g = dgl.batch([g1, g2])
s = dgl.sum_nodes(g, 'x')
m = dgl.mean_nodes(g, 'x')
assert U.allclose(s, th.stack([s1, s2], 0))
assert U.allclose(m, th.stack([m1, m2], 0))
ws = dgl.sum_nodes(g, 'x', 'w')
wm = dgl.mean_nodes(g, 'x', 'w')
assert U.allclose(ws, th.stack([ws1, ws2], 0))
assert U.allclose(wm, th.stack([wm1, wm2], 0))
s = dgl.sum_edges(g, 'x')
m = dgl.mean_edges(g, 'x')
assert U.allclose(s, th.stack([se1, th.zeros(5)], 0))
assert U.allclose(m, th.stack([me1, th.zeros(5)], 0))
def euclidean_matrix(graphs, dims, readout='sum'):
'''Returns the pairwise euclidean distance between readout feature from all graphs.
graphs : list of dgl graphs
dims : graph features are concatenation of features obtained from all iterations, and this variable has
the individual feature dimensions for the iterations.
'''
graphs = dgl.batch(graphs)
if readout == 'sum':
graph_reprs = dgl.sum_nodes(graphs, 'h')
elif readout == 'mean':
graph_reprs = dgl.mean_nodes(graphs, 'h')
else:
raise ValueError('Readout for gram_matrix shall be either "mean" or "sum"')
distances = []
dims = np.cumsum([0] + dims)
with torch.no_grad():
for dim_start, dim_end in zip(dims, dims[1:]):
features = graph_reprs[:, dim_start:dim_end]
matrix = dist_matrix(features, features)
distances.append(matrix.cpu().numpy())
return distances
def gram_matrix(graphs, dims, dist_fn, readout='sum'):
'''Wrapper function to compute the gram matrix on graphs in batch, returns list of gram matrices as numpy.array
graphs : list of dgl graphs
dims : graph features are concatenation of features obtained from all iterations, and this variable has
the individual feature dimensions for the iterations.
'''
graphs = dgl.batch(graphs)
if readout == 'sum':
graph_reprs = dgl.sum_nodes(graphs, 'h')
elif readout == 'mean':
graph_reprs = dgl.mean_nodes(graphs, 'h')
else:
raise ValueError('Readout for gram_matrix shall be either "mean" or "sum"')
distances = []
dims = np.cumsum([0] + dims)
with torch.no_grad():
for dim_start, dim_end in zip(dims, dims[1:]):
features = graph_reprs[:, dim_start:dim_end]
gram_matrix = dist_fn(features, features)
distances.append(gram_matrix.cpu().numpy())
return distances
def forward(self, graph, edge_feat, node_feat, g_repr):
node_trf_func = lambda x: self.compute_node_repr(
nodes=x, graph=graph, g_repr=g_repr)
graph.edata['edge_feat'] = edge_feat
graph.ndata['node_feat'] = node_feat
edge_trf_func = lambda x: self.compute_edge_repr(
edges=x, graph=graph, g_repr=g_repr)
graph.apply_edges(edge_trf_func)
graph.update_all(self.graph_message_func, self.graph_reduce_func,
node_trf_func)
e_comb = dgl.sum_edges(graph, 'edge_feat')
n_comb = dgl.sum_nodes(graph, 'node_feat')
e_out = graph.edata['edge_feat']
n_out = graph.ndata['node_feat']
e_keys = list(graph.edata.keys())
n_keys = list(graph.ndata.keys())
for key in e_keys:
graph.edata.pop(key)
for key in n_keys:
graph.ndata.pop(key)
return e_out, n_out, self.compute_u_repr(n_comb, e_comb, g_repr)
def forward(self, g, node_feats):
r"""Computes graph representations out of node representations.
Parameters
----------
g : DGLGraph
DGLGraph for a batch of graphs.
node_feats : float32 tensor of shape (V, node_in_feats)
Input node features. V for the number of nodes in the batch of graphs.
Returns
-------
g_feats : float32 tensor of shape (G, node_in_feats)
Output graph representations. G for the number of graphs in the batch.
"""
if self.gaussian_expand:
node_feats = self.gaussian_histogram(node_feats)
with g.local_scope():
g.ndata['h'] = node_feats
g_feats = dgl.sum_nodes(g, 'h')
if self.gaussian_expand:
g_feats = self.to_out(g_feats)
if self.activation is not None:
g_feats = self.activation(g_feats)
return g_feats
def forward(self, bg, feat):
batch_size = bg.batch_size
x = bg.ndata[feat]
batch = tensor([], dtype=torch.int64)
batch_num_nodes = bg.batch_num_nodes
for index, num in enumerate(batch_num_nodes):
batch = torch.cat((batch, tensor(index).expand(num)))
h = (x.new_zeros((self.num_layers, batch_size, self.in_channels)),
x.new_zeros((self.num_layers, batch_size, self.in_channels)))
q_star = x.new_zeros(batch_size, self.out_channels)
for i in range(self.processing_steps):
q, h = self.lstm(q_star.unsqueeze(0), h)
q = q.view(batch_size, self.in_channels)
e = (x * q[batch]).sum(dim=-1, keepdim=True)
a = torch.cat(list(
map(lambda x: softmax(x, dim=0),
list(torch.split(e, batch_num_nodes)))),
dim=0)
bg.ndata['w'] = a
r = dgl.sum_nodes(bg, feat, 'w')
q_star = torch.cat([q, r], dim=-1)
return q_star
def forward(self,
g,
h,
e,
snorm_n,
snorm_e,
mlp=True,
head=False,
return_graph=False):
h = self.embedding_h(h)
h = self.in_feat_dropout(h)
for conv in self.layers:
h = conv(g, h, snorm_n)
g.ndata['h'] = h
if return_graph:
return g
if self.readout == "sum":
hg = dgl.sum_nodes(g, 'h')
elif self.readout == "max":
hg = dgl.max_nodes(g, 'h')
elif self.readout == "mean":
hg = dgl.mean_nodes(g, 'h')
else:
hg = dgl.mean_nodes(g, 'h') # default readout is mean nodes
if mlp:
return self.MLP_layer(hg)
else:
if head:
return self.projection_head(hg)
else:
return hg
def forward(self, g, node_feats):
"""Computes graph representations out of node features.
Parameters
----------
g : DGLGraph
DGLGraph for a batch of graphs.
node_feats : float32 tensor of shape (V, node_feats)
Input node features, V for the number of nodes.
Returns
-------
graph_feats : float32 tensor of shape (G, graph_feats)
Graph representations computed. G for the number of graphs.
"""
node_feats = self.in_project(node_feats)
if self.activation is not None:
node_feats = self.activation(node_feats)
node_feats = self.out_project(node_feats)
with g.local_scope():
g.ndata['h'] = node_feats
if self.mode == 'max':
graph_feats = dgl.max_nodes(g, 'h')
elif self.mode == 'mean':
graph_feats = dgl.mean_nodes(g, 'h')
elif self.mode == 'sum':
graph_feats = dgl.sum_nodes(g, 'h')
return graph_feats
def forward(self, g, h, e):
h = self.embedding_h(h)
h = self.in_feat_dropout(h)
for conv in self.layers:
# For reduced graphs
h = conv(g, h, e)
# For original graphs
# h = conv(g, h)
g.ndata['h'] = h
if self.readout == "sum":
# For reduced graphs
hg = dgl.sum_nodes(g, feat='h', weight='weight')
# For original graphs
# hg = dgl.sum_nodes(g, feat= 'h')
elif self.readout == "max":
# For reduced graphs
hg = dgl.max_nodes(g, feat='h', weight='weight')
# For original graphs
# hg = dgl.max_nodes(g, feat= 'h')
elif self.readout == "mean":
# For reduced graphs
hg = dgl.mean_nodes(g, feat='h', weight='weight')
# For original graphs
# hg = dgl.mean_nodes(g, feat= 'h')
else:
# For reduced graphs
hg = dgl.mean_nodes(
g, feat='h', weight='weight') # default readout is mean nodes
# For original graphs
# hg = dgl.mean_nodes(g, feat= 'h')
return self.MLP_layer(hg)
def forward(self, g, h, e, pos_enc=None):
# input embedding
if self.pos_enc:
h = self.embedding_pos_enc(pos_enc)
else:
h = self.embedding_h(h)
# computing the 'pseudo' named tensor which depends on node degrees
g.ndata['deg'] = g.in_degrees()
g.apply_edges(self.compute_pseudo)
pseudo = g.edata['pseudo'].to(self.device).float()
for i in range(len(self.layers)):
h = self.layers[i](g, h, self.pseudo_proj[i](pseudo))
g.ndata['h'] = h
if self.readout == "sum":
hg = dgl.sum_nodes(g, 'h')
elif self.readout == "max":
hg = dgl.max_nodes(g, 'h')
elif self.readout == "mean":
hg = dgl.mean_nodes(g, 'h')
else:
hg = dgl.mean_nodes(g, 'h') # default readout is mean nodes
return self.MLP_layer(hg)
def forward(self, g, h, e, snorm_n, snorm_e):
h = self.embedding_h(h)
h = self.in_feat_dropout(h)
if self.edge_feat:
e = self.embedding_e(e)
# Loop all layers
for i, conv in enumerate(self.layers):
# Graph conv layers
h_t = conv(g, h, e, snorm_n)
h = h_t
# Virtual node layer
if self.virtual_node_layers is not None:
if i == 0:
vn_h = 0
if i < len(self.virtual_node_layers):
vn_h, h = self.virtual_node_layers[i].forward(g, h, vn_h)
g.ndata['h'] = h
# Readout layer
if self.readout == "sum":
hg = dgl.sum_nodes(g, 'h')
elif self.readout == "max":
hg = dgl.max_nodes(g, 'h')
elif self.readout == "mean":
hg = dgl.mean_nodes(g, 'h')
else:
hg = dgl.mean_nodes(g, 'h') # default readout is mean nodes
return self.MLP_layer(hg)
def forward(self, g):
self.embedding_layer(g, "node_0")
if self.atom_ref is not None:
self.e0(g, "e0")
self.rbf_layer(g)
self.edge_embedding_layer(g)
for idx in range(self.n_conv):
self.conv_layers[idx](g, idx + 1)
node_embeddings = tuple(g.ndata["node_%d" % (i)]
for i in range(self.n_conv + 1))
g.ndata["node"] = th.cat(node_embeddings, 1)
# concat multilevel representations
node = self.node_dense_layer1(g.ndata["node"])
node = self.activation(node)
res = self.node_dense_layer2(node)
g.ndata["res"] = res
if self.atom_ref is not None:
g.ndata["res"] = g.ndata["res"] + g.ndata["e0"]
if self.norm:
g.ndata["res"] = g.ndata[
"res"] * self.std_per_node + self.mean_per_node
res = dgl.sum_nodes(g, "res")
return res
def forward(self, g, node_feats, g_feats, get_node_weight=False):
"""
Parameters
----------
g : DGLGraph or BatchedDGLGraph
Constructed DGLGraphs.
node_feats : float32 tensor of shape (V, N1)
Input node features. V for the number of nodes and N1 for the feature size.
g_feats : float32 tensor of shape (G, N2)
Input graph features. G for the number of graphs and N2 for the feature size.
get_node_weight : bool
Whether to get the weights of atoms during readout.
Returns
-------
float32 tensor of shape (G, N2)
Updated graph features.
float32 tensor of shape (V, 1)
The weights of nodes in readout.
"""
with g.local_scope():
g.ndata['z'] = self.compute_logits(
torch.cat([dgl.broadcast_nodes(g, F.relu(g_feats)), node_feats], dim=1))
g.ndata['a'] = dgl.softmax_nodes(g, 'z')
g.ndata['hv'] = self.project_nodes(node_feats)
context = F.elu(dgl.sum_nodes(g, 'hv', 'a'))
if get_node_weight:
return self.gru(context, g_feats), g.ndata['a']
else:
return self.gru(context, g_feats)
def forward(self, graph, edge_feat, node_feat, g_repr, edge_hidden, node_hidden, graph_hidden):
graph.edata['edge_feat'] = edge_feat
graph.ndata['node_feat'] = node_feat
graph.edata['hidden1'] = edge_hidden[0][0]
graph.ndata['hidden1'] = node_hidden[0][0]
graph.edata['hidden2'] = edge_hidden[1][0]
graph.ndata['hidden2'] = node_hidden[1][0]
node_trf_func = lambda x : self.compute_node_repr(nodes=x, graph=graph, g_repr=g_repr)
edge_trf_func = lambda x: self.compute_edge_repr(edges=x, graph=graph, g_repr=g_repr)
graph.apply_edges(edge_trf_func)
graph.update_all(self.graph_message_func, self.graph_reduce_func, node_trf_func)
e_comb = dgl.sum_edges(graph, 'edge_feat')
n_comb = dgl.sum_nodes(graph, 'node_feat')
u_out, u_hidden = self.compute_u_repr(n_comb, e_comb, g_repr, graph_hidden)
e_feat = graph.edata['edge_feat']
n_feat = graph.ndata['node_feat']
h_e = (torch.unsqueeze(graph.edata['hidden1'],0),torch.unsqueeze(graph.edata['hidden2'],0))
h_n = (torch.unsqueeze(graph.ndata['hidden1'],0),torch.unsqueeze(graph.ndata['hidden2'],0))
e_keys = list(graph.edata.keys())
n_keys = list(graph.ndata.keys())
for key in e_keys:
graph.edata.pop(key)
for key in n_keys:
graph.ndata.pop(key)
return e_feat, h_e, n_feat, h_n, u_out, u_hidden
def forward(self, g):
h = g.ndata['attr']
h = h.to(self.device)
# list of hidden representation at each layer (including input)
hidden_rep = [h]
for layer in range(self.num_layers - 1):
h = self.ginlayers[layer](g, h)
hidden_rep.append(h)
score_over_layer = 0
# perform pooling over all nodes in each graph in every layer
for layer, h in enumerate(hidden_rep):
g.ndata['h'] = h
if self.graph_pooling_type == 'sum':
pooled_h = dgl.sum_nodes(g, 'h')
elif self.graph_pooling_type == 'mean':
pooled_h = dgl.mean_nodes(g, 'h')
elif self.graph_pooling_type == 'max':
pooled_h = dgl.max_nodes(g, 'h')
else:
raise NotImplementedError()
score_over_layer += F.dropout(
self.linears_prediction[layer](pooled_h),
self.final_dropout,
training=self.training)
return score_over_layer
def forward(self, g, h, e, h_lap_pos_enc=None, h_wl_pos_enc=None):
# input embedding
h = self.embedding_h(h)
h = self.in_feat_dropout(h)
if self.lap_pos_enc:
h_lap_pos_enc = self.embedding_lap_pos_enc(h_lap_pos_enc.float())
h = h + h_lap_pos_enc
if self.wl_pos_enc:
h_wl_pos_enc = self.embedding_wl_pos_enc(h_wl_pos_enc)
h = h + h_wl_pos_enc
if not self.edge_feat: # edge feature set to 1
e = torch.ones(e.size(0), 1).to(self.device)
e = self.embedding_e(e)
# convnets
for conv in self.layers:
h, e = conv(g, h, e)
g.ndata['h'] = h
if self.readout == "sum":
hg = dgl.sum_nodes(g, 'h')
elif self.readout == "max":
hg = dgl.max_nodes(g, 'h')
elif self.readout == "mean":
hg = dgl.mean_nodes(g, 'h')
else:
hg = dgl.mean_nodes(g, 'h') # default readout is mean nodes
return self.MLP_layer(hg)
def forward(self, g):
# g_list list of molecules
g.edata['distance'] = g.edata['distance'].reshape(-1, 1)
self.embedding_layer(g)
if self.atom_ref is not None:
self.e0(g, "e0")
self.rbf_layer(g)
for idx in range(self.n_conv):
self.conv_layers[idx](g)
atom = self.atom_dense_layer1(g.ndata["node"])
atom = self.activation(atom)
atom = self.activation(self.atom_dense_layer2(atom))
res = self.regressor(atom)
g.ndata["res"] = res
if self.atom_ref is not None:
g.ndata["res"] = g.ndata["res"] + g.ndata["e0"]
if self.norm:
g.ndata["res"] = g.ndata[
"res"] * self.std_per_atom + self.mean_per_atom
res = dgl.sum_nodes(g, "res")
return res
def forward(self, g, h, e, snorm_n, snorm_e):
h = self.embedding_h(h)
h = self.in_feat_dropout(h)
h_init = h
'''for conv in self.layers:
h = conv(g, h, snorm_n)
h = self.joining_layer(h_init + h)'''
for i in range(self.layer_count):
conv = self.layers[i]
joint = self.joining_layers[i]
h = conv(g, h, snorm_n)
h = joint(h_init + h)
g.ndata['h'] = h
if self.readout == "sum":
hg = dgl.sum_nodes(g, 'h')
elif self.readout == "max":
hg = dgl.max_nodes(g, 'h')
elif self.readout == "mean":
hg = dgl.mean_nodes(g, 'h')
else:
hg = dgl.mean_nodes(g, 'h') # default readout is mean nodes
return self.MLP_layer(hg)
def forward(self, g, h, e, snorm_n, snorm_e):
h = self.embedding_h(h)
# computing the 'pseudo' named tensor which depends on node degrees
us, vs = g.edges()
# to avoid zero division in case in_degree is 0, we add constant '1' in all node degrees denoting self-loop
pseudo = [[
1 / np.sqrt(g.in_degree(us[i]) + 1),
1 / np.sqrt(g.in_degree(vs[i]) + 1)
] for i in range(g.number_of_edges())]
pseudo = torch.Tensor(pseudo).to(self.device)
for i in range(len(self.layers)):
h = self.layers[i](g, h, self.pseudo_proj[i](pseudo), snorm_n)
g.ndata['h'] = h
if self.readout == "sum":
hg = dgl.sum_nodes(g, 'h')
elif self.readout == "max":
hg = dgl.max_nodes(g, 'h')
elif self.readout == "mean":
hg = dgl.mean_nodes(g, 'h')
else:
hg = dgl.mean_nodes(g, 'h') # default readout is mean nodes
return self.MLP_layer(hg)
