diff --git a/python/dgllife/__init__.py b/python/dgllife/__init__.py
index 504a563e..8bf06f3e 100644
--- a/python/dgllife/__init__.py
+++ b/python/dgllife/__init__.py
@@ -10,7 +10,8 @@
 
 try:
     import rdkit
-    from . import data
-    from . import utils
 except ImportError:
     print('RDKit is not installed, which is required for utils related to cheminformatics')
+
+from . import data
+from . import utils
diff --git a/python/dgllife/model/gnn/attentivefp.py b/python/dgllife/model/gnn/attentivefp.py
index 6708e7d4..ec5ddc20 100644
--- a/python/dgllife/model/gnn/attentivefp.py
+++ b/python/dgllife/model/gnn/attentivefp.py
@@ -69,7 +69,7 @@ def forward(self, g, edge_logits, edge_feats, node_feats):
         """
         g = g.local_var()
         g.edata['e'] = edge_softmax(g, edge_logits) * self.edge_transform(edge_feats)
-        g.update_all(fn.copy_edge('e', 'm'), fn.sum('m', 'c'))
+        g.update_all(fn.copy_e('e', 'm'), fn.sum('m', 'c'))
         context = F.elu(g.ndata['c'])
         return F.relu(self.gru(context, node_feats))
 
@@ -123,7 +123,7 @@ def forward(self, g, edge_logits, node_feats):
         g.edata['a'] = edge_softmax(g, edge_logits)
         g.ndata['hv'] = self.project_node(node_feats)
 
-        g.update_all(fn.src_mul_edge('hv', 'a', 'm'), fn.sum('m', 'c'))
+        g.update_all(fn.u_mul_e('hv', 'a', 'm'), fn.sum('m', 'c'))
         context = F.elu(g.ndata['c'])
         return F.relu(self.gru(context, node_feats))
 
diff --git a/python/dgllife/model/gnn/weave.py b/python/dgllife/model/gnn/weave.py
index 0b00248b..091cebb0 100644
--- a/python/dgllife/model/gnn/weave.py
+++ b/python/dgllife/model/gnn/weave.py
@@ -107,7 +107,7 @@ def forward(self, g, node_feats, edge_feats, node_only=False):
         # Update node features
         node_node_feats = self.activation(self.node_to_node(node_feats))
         g.edata['e2n'] = self.activation(self.edge_to_node(edge_feats))
-        g.update_all(fn.copy_edge('e2n', 'm'), fn.sum('m', 'e2n'))
+        g.update_all(fn.copy_e('e2n', 'm'), fn.sum('m', 'e2n'))
         edge_node_feats = g.ndata.pop('e2n')
         new_node_feats = self.activation(self.update_node(
             torch.cat([node_node_feats, edge_node_feats], dim=1)))
diff --git a/python/dgllife/model/gnn/wln.py b/python/dgllife/model/gnn/wln.py
index 680a99e4..87b478a8 100644
--- a/python/dgllife/model/gnn/wln.py
+++ b/python/dgllife/model/gnn/wln.py
@@ -173,10 +173,10 @@ def forward(self, g, node_feats, edge_feats):
             if g.num_edges() > 0:
                 # The following lines do not work for a graph without edges.
                 g.ndata['hv'] = node_feats
-                g.apply_edges(fn.copy_src('hv', 'he_src'))
+                g.apply_edges(fn.copy_u('hv', 'he_src'))
                 concat_edge_feats = torch.cat([g.edata['he_src'], edge_feats], dim=1)
                 g.edata['he'] = self.project_concatenated_messages(concat_edge_feats)
-                g.update_all(fn.copy_edge('he', 'm'), fn.sum('m', 'hv_new'))
+                g.update_all(fn.copy_e('he', 'm'), fn.sum('m', 'hv_new'))
                 node_feats = self.get_new_node_feats(
                     torch.cat([node_feats, g.ndata['hv_new']], dim=1))
             else:
diff --git a/python/dgllife/model/model_zoo/wln_reaction_center.py b/python/dgllife/model/model_zoo/wln_reaction_center.py
index a3196a0f..c2b2f417 100644
--- a/python/dgllife/model/model_zoo/wln_reaction_center.py
+++ b/python/dgllife/model/model_zoo/wln_reaction_center.py
@@ -67,7 +67,7 @@ def forward(self, batch_complete_graphs, node_feats, feat_sum, node_pair_feat):
                 self.project_node_pair_feature(node_pair_feat)
             )
             batch_complete_graphs.update_all(
-                fn.src_mul_edge('hv', 'a', 'm'), fn.sum('m', 'context'))
+                fn.u_mul_e('hv', 'a', 'm'), fn.sum('m', 'context'))
             node_contexts = batch_complete_graphs.ndata.pop('context')
 
         return node_contexts