[Feature] Change the Default Value of transpose in DGLGraph.adj (#2046)

* Update adj

* Fix

examples/mxnet/sampling/dis_sampling/dis_gcn_cv_sc.py

@@ -67,7 +67,7 @@ def gcn_cv_train(g, ctx, args, n_classes, train_nid, test_nid, n_test_samples, d
 
     # initialize graph
     dur = []
-    adj = g.adjacency_matrix().as_in_context(g_ctx)
+    adj = g.adjacency_matrix(transpose=False).as_in_context(g_ctx)
     for epoch in range(args.n_epochs):
         start = time.time()
         if distributed:

examples/mxnet/sampling/dis_sampling/dis_graphsage_cv.py

@@ -65,7 +65,7 @@ def graphsage_cv_train(g, ctx, args, n_classes, train_nid, test_nid, n_test_samp
     # initialize graph
     dur = []
 
-    adj = g.adjacency_matrix().as_in_context(g_ctx)
+    adj = g.adjacency_matrix(transpose=False).as_in_context(g_ctx)
     for epoch in range(args.n_epochs):
         start = time.time()
         if distributed:

examples/mxnet/sampling/gcn_cv_sc.py

@@ -191,7 +191,7 @@ def gcn_cv_train(g, ctx, args, n_classes, train_nid, test_nid, n_test_samples, d
 
     # initialize graph
     dur = []
-    adj = g.adjacency_matrix().as_in_context(g_ctx)
+    adj = g.adjacency_matrix(transpose=False).as_in_context(g_ctx)
     for epoch in range(args.n_epochs):
         start = time.time()
         if distributed:

examples/mxnet/sampling/graphsage_cv.py

@@ -229,7 +229,7 @@ def graphsage_cv_train(g, ctx, args, n_classes, train_nid, test_nid, n_test_samp
     # initialize graph
     dur = []
 
-    adj = g.adjacency_matrix().as_in_context(g_ctx)
+    adj = g.adjacency_matrix(transpose=False).as_in_context(g_ctx)
     for epoch in range(args.n_epochs):
         start = time.time()
         if distributed:

examples/pytorch/_deprecated/adaptive_sampling/adaptive_sampling.py

@@ -106,7 +106,7 @@ class AdaptGenerator(object):
     def __init__(self, graph, num_blocks, node_feature=None, sampler=None, num_workers=0, coalesce=False,
                  sampler_weights=None, layer_nodes=None):
         self.node_feature = node_feature
-        adj = graph.adjacency_matrix_scipy()
+        adj = graph.adjacency_matrix_scipy(transpose=False)
         adj.data = np.ones(adj.nnz)
         self.norm_adj = normalize_adj(adj).tocsr()
         self.layer_nodes = layer_nodes
@@ -370,7 +370,7 @@ class AdaptGraphSAGENet(nn.Module):
         ])
         self.sample_weights = sample_weights
         self.node_feature = node_feature
-        self.norm_adj = normalize_adj(trainG.adjacency_matrix_scipy())
+        self.norm_adj = normalize_adj(trainG.adjacency_matrix_scipy(transpose=False))
 
     def forward(self, nf, h, is_test=False):
         for i, layer in enumerate(self.layers):

examples/pytorch/diffpool/model/dgl_layers/gnn.py

@@ -123,7 +123,7 @@ class DiffPoolBatchedGraphLayer(nn.Module):
         assign_tensor = masked_softmax(assign_tensor, mask,
                                        memory_efficient=False)
         h = torch.matmul(torch.t(assign_tensor), feat)
-        adj = g.adjacency_matrix(ctx=device)
+        adj = g.adjacency_matrix(transpose=False, ctx=device)
         adj_new = torch.sparse.mm(adj, assign_tensor)
         adj_new = torch.mm(torch.t(assign_tensor), adj_new)
 

python/dgl/heterograph.py

@@ -3329,11 +3329,11 @@ class DGLHeteroGraph(object):
         else:
             return deg
 
-    def adjacency_matrix(self, transpose=None, ctx=F.cpu(), scipy_fmt=None, etype=None):
+    def adjacency_matrix(self, transpose=True, ctx=F.cpu(), scipy_fmt=None, etype=None):
         """Alias of :func:`adj`"""
         return self.adj(transpose, ctx, scipy_fmt, etype)
 
-    def adj(self, transpose=None, ctx=F.cpu(), scipy_fmt=None, etype=None):
+    def adj(self, transpose=True, ctx=F.cpu(), scipy_fmt=None, etype=None):
         """Return the adjacency matrix of edges of the given edge type.
 
         By default, a row of returned adjacency matrix represents the
@@ -3345,7 +3345,7 @@ class DGLHeteroGraph(object):
         Parameters
         ----------
         transpose : bool, optional
-            A flag to transpose the returned adjacency matrix. (Default: False)
+            A flag to transpose the returned adjacency matrix. (Default: True)
         ctx : context, optional
             The context of returned adjacency matrix. (Default: cpu)
         scipy_fmt : str, optional
@@ -3381,31 +3381,24 @@ class DGLHeteroGraph(object):
         Get a backend dependent sparse tensor. Here we use PyTorch for example.
 
         >>> g.adj(etype='develops')
-        tensor(indices=tensor([[0, 2],
-                               [0, 1]]),
+        tensor(indices=tensor([[0, 1],
+                               [0, 2]]),
                values=tensor([1., 1.]),
-               size=(3, 2), nnz=2, layout=torch.sparse_coo)
+               size=(2, 3), nnz=2, layout=torch.sparse_coo)
 
         Get a scipy coo sparse matrix.
 
         >>> g.adj(scipy_fmt='coo', etype='develops')
-        <3x2 sparse matrix of type '<class 'numpy.int64'>'
+        <2x3 sparse matrix of type '<class 'numpy.int64'>'
 	    with 2 stored elements in COOrdinate format>
         """
-        if transpose is None:
-            dgl_warning(
-                "Currently adjacency_matrix() returns a matrix with destination as rows"
-                " by default.\n\tIn 0.5 the result will have source as rows"
-                " (i.e. transpose=True)")
-            transpose = False
-
         etid = self.get_etype_id(etype)
         if scipy_fmt is None:
             return self._graph.adjacency_matrix(etid, transpose, ctx)[0]
         else:
             return self._graph.adjacency_matrix_scipy(etid, transpose, scipy_fmt, False)
 
-    def adjacency_matrix_scipy(self, transpose=None, fmt='csr', return_edge_ids=None):
+    def adjacency_matrix_scipy(self, transpose=True, fmt='csr', return_edge_ids=None):
         """DEPRECATED: please use ``dgl.adjacency_matrix(transpose, scipy_fmt=fmt)``.
         """
         dgl_warning('DGLGraph.adjacency_matrix_scipy is deprecated. '

tests/compute/test_heterograph.py

@@ -453,7 +453,7 @@ def test_edge_ids():
 @parametrize_dtype
 def test_adj(idtype):
     g = create_test_heterograph(idtype)
-    adj = F.sparse_to_numpy(g.adj(etype='follows'))
+    adj = F.sparse_to_numpy(g.adj(transpose=False, etype='follows'))
     assert np.allclose(
             adj,
             np.array([[0., 0., 0.],
@@ -465,7 +465,7 @@ def test_adj(idtype):
             np.array([[0., 1., 0.],
                       [0., 0., 1.],
                       [0., 0., 0.]]))
-    adj = F.sparse_to_numpy(g.adj(etype='plays'))
+    adj = F.sparse_to_numpy(g.adj(transpose=False, etype='plays'))
     assert np.allclose(
             adj,
             np.array([[1., 1., 0.],
@@ -477,29 +477,29 @@ def test_adj(idtype):
                       [1., 1.],
                       [0., 1.]]))
 
-    adj = g.adj(scipy_fmt='csr', etype='follows')
+    adj = g.adj(transpose=False, scipy_fmt='csr', etype='follows')
     assert np.allclose(
             adj.todense(),
             np.array([[0., 0., 0.],
                       [1., 0., 0.],
                       [0., 1., 0.]]))
-    adj = g.adj(scipy_fmt='coo', etype='follows')
+    adj = g.adj(transpose=False, scipy_fmt='coo', etype='follows')
     assert np.allclose(
             adj.todense(),
             np.array([[0., 0., 0.],
                       [1., 0., 0.],
                       [0., 1., 0.]]))
-    adj = g.adj(scipy_fmt='csr', etype='plays')
+    adj = g.adj(transpose=False, scipy_fmt='csr', etype='plays')
     assert np.allclose(
             adj.todense(),
             np.array([[1., 1., 0.],
                       [0., 1., 1.]]))
-    adj = g.adj(scipy_fmt='coo', etype='plays')
+    adj = g.adj(transpose=False, scipy_fmt='coo', etype='plays')
     assert np.allclose(
             adj.todense(),
             np.array([[1., 1., 0.],
                       [0., 1., 1.]]))
-    adj = F.sparse_to_numpy(g['follows'].adj())
+    adj = F.sparse_to_numpy(g['follows'].adj(transpose=False))
     assert np.allclose(
             adj,
             np.array([[0., 0., 0.],

tests/compute/test_traversal.py

@@ -69,7 +69,7 @@ def test_topological_nodes(idtype, n=100):
 
     layers_dgl = dgl.topological_nodes_generator(g)
 
-    adjmat = g.adjacency_matrix()
+    adjmat = g.adjacency_matrix(transpose=False)
     def tensor_topo_traverse():
         n = g.number_of_nodes()
         mask = F.copy_to(F.ones((n, 1)), F.cpu())

tests/mxnet/test_nn.py

@@ -24,7 +24,7 @@ def test_graph_conv(idtype):
     g = dgl.from_networkx(nx.path_graph(3))
     g = g.astype(idtype).to(F.ctx())
     ctx = F.ctx()
-    adj = g.adjacency_matrix(ctx=ctx)
+    adj = g.adjacency_matrix(transpose=False, ctx=ctx)
 
     conv = nn.GraphConv(5, 2, norm='none', bias=True)
     conv.initialize(ctx=ctx)
@@ -131,7 +131,7 @@ def _S2AXWb(A, N, X, W, b):
 def test_tagconv():
     g = dgl.from_networkx(nx.path_graph(3)).to(F.ctx())
     ctx = F.ctx()
-    adj = g.adjacency_matrix(ctx=ctx)
+    adj = g.adjacency_matrix(transpose=False, ctx=ctx)
     norm = mx.nd.power(g.in_degrees().astype('float32'), -0.5)
 
     conv = nn.TAGConv(5, 2, bias=True)
@@ -294,7 +294,7 @@ def test_dense_cheb_conv():
     for k in range(1, 4):
         ctx = F.ctx()
         g = dgl.from_scipy(sp.sparse.random(100, 100, density=0.3)).to(F.ctx())
-        adj = g.adjacency_matrix(ctx=ctx).tostype('default')
+        adj = g.adjacency_matrix(transpose=False, ctx=ctx).tostype('default')
         cheb = nn.ChebConv(5, 2, k)
         dense_cheb = nn.DenseChebConv(5, 2, k)
         cheb.initialize(ctx=ctx)
@@ -318,7 +318,7 @@ def test_dense_cheb_conv():
 def test_dense_graph_conv(idtype, g, norm_type):
     g = g.astype(idtype).to(F.ctx())
     ctx = F.ctx()
-    adj = g.adjacency_matrix(ctx=ctx).tostype('default')
+    adj = g.adjacency_matrix(transpose=False, ctx=ctx).tostype('default')
     conv = nn.GraphConv(5, 2, norm=norm_type, bias=True)
     dense_conv = nn.DenseGraphConv(5, 2, norm=norm_type, bias=True)
     conv.initialize(ctx=ctx)
@@ -337,7 +337,7 @@ def test_dense_graph_conv(idtype, g, norm_type):
 def test_dense_sage_conv(idtype, g):
     g = g.astype(idtype).to(F.ctx())
     ctx = F.ctx()
-    adj = g.adjacency_matrix(ctx=ctx).tostype('default')
+    adj = g.adjacency_matrix(transpose=False, ctx=ctx).tostype('default')
     sage = nn.SAGEConv(5, 2, 'gcn')
     dense_sage = nn.DenseSAGEConv(5, 2)
     sage.initialize(ctx=ctx)

tests/pytorch/test_nn.py

@@ -20,7 +20,7 @@ def _AXWb(A, X, W, b):
 def test_graph_conv0():
     g = dgl.DGLGraph(nx.path_graph(3)).to(F.ctx())
     ctx = F.ctx()
-    adj = g.adjacency_matrix(ctx=ctx)
+    adj = g.adjacency_matrix(transpose=False, ctx=ctx)
 
     conv = nn.GraphConv(5, 2, norm='none', bias=True)
     conv = conv.to(ctx)
@@ -125,7 +125,7 @@ def test_tagconv():
     g = dgl.DGLGraph(nx.path_graph(3))
     g = g.to(F.ctx())
     ctx = F.ctx()
-    adj = g.adjacency_matrix(ctx=ctx)
+    adj = g.adjacency_matrix(transpose=False, ctx=ctx)
     norm = th.pow(g.in_degrees().float(), -0.5)
 
     conv = nn.TAGConv(5, 2, bias=True)
@@ -599,7 +599,7 @@ def test_dense_graph_conv(norm_type, g, idtype):
     g = g.astype(idtype).to(F.ctx())
     ctx = F.ctx()
     # TODO(minjie): enable the following option after #1385
-    adj = g.adjacency_matrix(ctx=ctx).to_dense()
+    adj = g.adjacency_matrix(transpose=False, ctx=ctx).to_dense()
     conv = nn.GraphConv(5, 2, norm=norm_type, bias=True)
     dense_conv = nn.DenseGraphConv(5, 2, norm=norm_type, bias=True)
     dense_conv.weight.data = conv.weight.data
@@ -616,7 +616,7 @@ def test_dense_graph_conv(norm_type, g, idtype):
 def test_dense_sage_conv(g, idtype):
     g = g.astype(idtype).to(F.ctx())
     ctx = F.ctx()
-    adj = g.adjacency_matrix(ctx=ctx).to_dense()
+    adj = g.adjacency_matrix(transpose=False, ctx=ctx).to_dense()
     sage = nn.SAGEConv(5, 2, 'gcn')
     dense_sage = nn.DenseSAGEConv(5, 2)
     dense_sage.fc.weight.data = sage.fc_neigh.weight.data
@@ -662,7 +662,7 @@ def test_dense_cheb_conv():
         ctx = F.ctx()
         g = dgl.DGLGraph(sp.sparse.random(100, 100, density=0.1), readonly=True)
         g = g.to(F.ctx())
-        adj = g.adjacency_matrix(ctx=ctx).to_dense()
+        adj = g.adjacency_matrix(transpose=False, ctx=ctx).to_dense()
         cheb = nn.ChebConv(5, 2, k, None)
         dense_cheb = nn.DenseChebConv(5, 2, k)
         #for i in range(len(cheb.fc)):

tests/tensorflow/test_nn.py

@@ -21,7 +21,7 @@ def _AXWb(A, X, W, b):
 def test_graph_conv():
     g = dgl.DGLGraph(nx.path_graph(3)).to(F.ctx())
     ctx = F.ctx()
-    adj = tf.sparse.to_dense(tf.sparse.reorder(g.adjacency_matrix(ctx=ctx)))
+    adj = tf.sparse.to_dense(tf.sparse.reorder(g.adjacency_matrix(transpose=False, ctx=ctx)))
 
     conv = nn.GraphConv(5, 2, norm='none', bias=True)
     # conv = conv
@@ -488,7 +488,7 @@ def test_dense_cheb_conv():
         g = dgl.DGLGraph(sp.sparse.random(100, 100, density=0.1, random_state=42))
         g = g.to(ctx)
 
-        adj = tf.sparse.to_dense(tf.sparse.reorder(g.adjacency_matrix(ctx=ctx)))
+        adj = tf.sparse.to_dense(tf.sparse.reorder(g.adjacency_matrix(transpose=False, ctx=ctx)))
         cheb = nn.ChebConv(5, 2, k, None, bias=True)
         dense_cheb = nn.DenseChebConv(5, 2, k, bias=True)