rapidsai
diff --git a/‎examples/pytorch/dgi/train.py
+11-11 b/‎examples/pytorch/dgi/train.py
+11-11
diff --git a/‎python/dgl/data/citation_graph.py
-172 b/‎python/dgl/data/citation_graph.py
-172
diff --git a/‎python/dgl/data/gnn_benchmark.py
-50 b/‎python/dgl/data/gnn_benchmark.py
-50
@@ -23,18 +23,18 @@ def main(args):
     # load and preprocess dataset
     data = load_data(args)
     g = data[0]
-    features = torch.FloatTensor(data.features)
-    labels = torch.LongTensor(data.labels)
+    features = torch.FloatTensor(g.ndata['feat'])
+    labels = torch.LongTensor(g.ndata['label'])
     if hasattr(torch, 'BoolTensor'):
-        train_mask = torch.BoolTensor(data.train_mask)
-        val_mask = torch.BoolTensor(data.val_mask)
-        test_mask = torch.BoolTensor(data.test_mask)
+        train_mask = torch.BoolTensor(g.ndata['train_mask'])
+        val_mask = torch.BoolTensor(g.ndata['val_mask'])
+        test_mask = torch.BoolTensor(g.ndata['test_mask'])
     else:
-        train_mask = torch.ByteTensor(data.train_mask)
-        val_mask = torch.ByteTensor(data.val_mask)
-        test_mask = torch.ByteTensor(data.test_mask)
+        train_mask = torch.ByteTensor(g.ndata['train_mask'])
+        val_mask = torch.ByteTensor(g.ndata['val_mask'])
+        test_mask = torch.ByteTensor(g.ndata['test_mask'])
     in_feats = features.shape[1]
-    n_classes = data.num_labels
+    n_classes = data.num_classes
     n_edges = g.number_of_edges()
 
     if args.gpu < 0:
@@ -130,7 +130,7 @@ def main(args):
         loss = F.nll_loss(preds[train_mask], labels[train_mask])
         loss.backward()
         classifier_optimizer.step()
-        
+
         if epoch >= 3:
             dur.append(time.time() - t0)
 
@@ -171,5 +171,5 @@ def main(args):
     parser.set_defaults(self_loop=False)
     args = parser.parse_args()
     print(args)
-    
+
     main(args)
@@ -251,31 +251,6 @@ def num_classes(self):
         We preserve these properties for compatability.
     """
 
-    @property
-    def train_mask(self):
-        deprecate_property('dataset.train_mask', 'g.ndata[\'train_mask\']')
-        return F.asnumpy(self._g.ndata['train_mask'])
-
-    @property
-    def val_mask(self):
-        deprecate_property('dataset.val_mask', 'g.ndata[\'val_mask\']')
-        return F.asnumpy(self._g.ndata['val_mask'])
-
-    @property
-    def test_mask(self):
-        deprecate_property('dataset.test_mask', 'g.ndata[\'test_mask\']')
-        return F.asnumpy(self._g.ndata['test_mask'])
-
-    @property
-    def labels(self):
-        deprecate_property('dataset.label', 'g.ndata[\'label\']')
-        return F.asnumpy(self._g.ndata['label'])
-
-    @property
-    def features(self):
-        deprecate_property('dataset.feat', 'g.ndata[\'feat\']')
-        return self._g.ndata['feat']
-
     @property
     def reverse_edge(self):
         return self._reverse_edge
@@ -306,43 +281,6 @@ def _sample_mask(idx, l):
 class CoraGraphDataset(CitationGraphDataset):
     r""" Cora citation network dataset.
 
-    .. deprecated:: 0.5.0
-
-        - ``graph`` is deprecated, it is replaced by:
-
-            >>> dataset = CoraGraphDataset()
-            >>> graph = dataset[0]
-
-        - ``train_mask`` is deprecated, it is replaced by:
-
-            >>> dataset = CoraGraphDataset()
-            >>> graph = dataset[0]
-            >>> train_mask = graph.ndata['train_mask']
-
-        - ``val_mask`` is deprecated, it is replaced by:
-
-            >>> dataset = CoraGraphDataset()
-            >>> graph = dataset[0]
-            >>> val_mask = graph.ndata['val_mask']
-
-        - ``test_mask`` is deprecated, it is replaced by:
-
-            >>> dataset = CoraGraphDataset()
-            >>> graph = dataset[0]
-            >>> test_mask = graph.ndata['test_mask']
-
-        - ``labels`` is deprecated, it is replaced by:
-
-            >>> dataset = CoraGraphDataset()
-            >>> graph = dataset[0]
-            >>> labels = graph.ndata['label']
-
-        - ``feat`` is deprecated, it is replaced by:
-
-            >>> dataset = CoraGraphDataset()
-            >>> graph = dataset[0]
-            >>> feat = graph.ndata['feat']
-
     Nodes mean paper and edges mean citation
     relationships. Each node has a predefined
     feature with 1433 dimensions. The dataset is
@@ -383,18 +321,6 @@ class CoraGraphDataset(CitationGraphDataset):
     ----------
     num_classes: int
         Number of label classes
-    graph: networkx.DiGraph
-        Graph structure
-    train_mask: numpy.ndarray
-        Mask of training nodes
-    val_mask: numpy.ndarray
-        Mask of validation nodes
-    test_mask: numpy.ndarray
-        Mask of test nodes
-    labels: numpy.ndarray
-        Ground truth labels of each node
-    features: Tensor
-        Node features
 
     Notes
     -----
@@ -454,43 +380,6 @@ def __len__(self):
 class CiteseerGraphDataset(CitationGraphDataset):
     r""" Citeseer citation network dataset.
 
-    .. deprecated:: 0.5.0
-
-        - ``graph`` is deprecated, it is replaced by:
-
-            >>> dataset = CiteseerGraphDataset()
-            >>> graph = dataset[0]
-
-        - ``train_mask`` is deprecated, it is replaced by:
-
-            >>> dataset = CiteseerGraphDataset()
-            >>> graph = dataset[0]
-            >>> train_mask = graph.ndata['train_mask']
-
-        - ``val_mask`` is deprecated, it is replaced by:
-
-            >>> dataset = CiteseerGraphDataset()
-            >>> graph = dataset[0]
-            >>> val_mask = graph.ndata['val_mask']
-
-        - ``test_mask`` is deprecated, it is replaced by:
-
-            >>> dataset = CiteseerGraphDataset()
-            >>> graph = dataset[0]
-            >>> test_mask = graph.ndata['test_mask']
-
-        - ``labels`` is deprecated, it is replaced by:
-
-            >>> dataset = CiteseerGraphDataset()
-            >>> graph = dataset[0]
-            >>> labels = graph.ndata['label']
-
-        - ``feat`` is deprecated, it is replaced by:
-
-            >>> dataset = CiteseerGraphDataset()
-            >>> graph = dataset[0]
-            >>> feat = graph.ndata['feat']
-
     Nodes mean scientific publications and edges
     mean citation relationships. Each node has a
     predefined feature with 3703 dimensions. The
@@ -531,18 +420,6 @@ class CiteseerGraphDataset(CitationGraphDataset):
     ----------
     num_classes: int
         Number of label classes
-    graph: networkx.DiGraph
-        Graph structure
-    train_mask: numpy.ndarray
-        Mask of training nodes
-    val_mask: numpy.ndarray
-        Mask of validation nodes
-    test_mask: numpy.ndarray
-        Mask of test nodes
-    labels: numpy.ndarray
-        Ground truth labels of each node
-    features: Tensor
-        Node features
 
     Notes
     -----
@@ -605,43 +482,6 @@ def __len__(self):
 class PubmedGraphDataset(CitationGraphDataset):
     r""" Pubmed citation network dataset.
 
-    .. deprecated:: 0.5.0
-
-        - ``graph`` is deprecated, it is replaced by:
-
-            >>> dataset = PubmedGraphDataset()
-            >>> graph = dataset[0]
-
-        - ``train_mask`` is deprecated, it is replaced by:
-
-            >>> dataset = PubmedGraphDataset()
-            >>> graph = dataset[0]
-            >>> train_mask = graph.ndata['train_mask']
-
-        - ``val_mask`` is deprecated, it is replaced by:
-
-            >>> dataset = PubmedGraphDataset()
-            >>> graph = dataset[0]
-            >>> val_mask = graph.ndata['val_mask']
-
-        - ``test_mask`` is deprecated, it is replaced by:
-
-            >>> dataset = PubmedGraphDataset()
-            >>> graph = dataset[0]
-            >>> test_mask = graph.ndata['test_mask']
-
-        - ``labels`` is deprecated, it is replaced by:
-
-            >>> dataset = PubmedGraphDataset()
-            >>> graph = dataset[0]
-            >>> labels = graph.ndata['label']
-
-        - ``feat`` is deprecated, it is replaced by:
-
-            >>> dataset = PubmedGraphDataset()
-            >>> graph = dataset[0]
-            >>> feat = graph.ndata['feat']
-
     Nodes mean scientific publications and edges
     mean citation relationships. Each node has a
     predefined feature with 500 dimensions. The
@@ -682,18 +522,6 @@ class PubmedGraphDataset(CitationGraphDataset):
     ----------
     num_classes: int
         Number of label classes
-    graph: networkx.DiGraph
-        Graph structure
-    train_mask: numpy.ndarray
-        Mask of training nodes
-    val_mask: numpy.ndarray
-        Mask of validation nodes
-    test_mask: numpy.ndarray
-        Mask of test nodes
-    labels: numpy.ndarray
-        Ground truth labels of each node
-    features: Tensor
-        Node features
 
     Notes
     -----
 
@@ -106,11 +106,6 @@ def num_classes(self):
         """Number of classes."""
         raise NotImplementedError
 
-    @property
-    def data(self):
-        deprecate_property('dataset.data', 'dataset[0]')
-        return self._data
-
     def __getitem__(self, idx):
         r""" Get graph by index
 
@@ -142,13 +137,6 @@ def __len__(self):
 class CoraFullDataset(GNNBenchmarkDataset):
     r"""CORA-Full dataset for node classification task.
 
-    .. deprecated:: 0.5.0
-
-        - ``data`` is deprecated, it is repalced by:
-
-        >>> dataset = CoraFullDataset()
-        >>> graph = dataset[0]
-
     Extended Cora dataset. Nodes represent paper and edges represent citations.
 
     Reference: `<https://github.com/shchur/gnn-benchmark#datasets>`_
@@ -179,8 +167,6 @@ class CoraFullDataset(GNNBenchmarkDataset):
     ----------
     num_classes : int
         Number of classes for each node.
-    data : list
-        A list of DGLGraph objects
 
     Examples
     --------
@@ -211,13 +197,6 @@ def num_classes(self):
 class CoauthorCSDataset(GNNBenchmarkDataset):
     r""" 'Computer Science (CS)' part of the Coauthor dataset for node classification task.
 
-    .. deprecated:: 0.5.0
-
-        - ``data`` is deprecated, it is repalced by:
-
-        >>> dataset = CoauthorCSDataset()
-        >>> graph = dataset[0]
-
     Coauthor CS and Coauthor Physics are co-authorship graphs based on the Microsoft Academic Graph
     from the KDD Cup 2016 challenge. Here, nodes are authors, that are connected by an edge if they
     co-authored a paper; node features represent paper keywords for each author’s papers, and class
@@ -251,8 +230,6 @@ class CoauthorCSDataset(GNNBenchmarkDataset):
     ----------
     num_classes : int
         Number of classes for each node.
-    data : list
-        A list of DGLGraph objects
 
     Examples
     --------
@@ -283,13 +260,6 @@ def num_classes(self):
 class CoauthorPhysicsDataset(GNNBenchmarkDataset):
     r""" 'Physics' part of the Coauthor dataset for node classification task.
 
-    .. deprecated:: 0.5.0
-
-        - ``data`` is deprecated, it is repalced by:
-
-        >>> dataset = CoauthorPhysicsDataset()
-        >>> graph = dataset[0]
-
     Coauthor CS and Coauthor Physics are co-authorship graphs based on the Microsoft Academic Graph
     from the KDD Cup 2016 challenge. Here, nodes are authors, that are connected by an edge if they
     co-authored a paper; node features represent paper keywords for each author’s papers, and class
@@ -323,8 +293,6 @@ class CoauthorPhysicsDataset(GNNBenchmarkDataset):
     ----------
     num_classes : int
         Number of classes for each node.
-    data : list
-        A list of DGLGraph objects
 
     Examples
     --------
@@ -355,13 +323,6 @@ def num_classes(self):
 class AmazonCoBuyComputerDataset(GNNBenchmarkDataset):
     r""" 'Computer' part of the AmazonCoBuy dataset for node classification task.
 
-    .. deprecated:: 0.5.0
-
-        - ``data`` is deprecated, it is repalced by:
-
-        >>> dataset = AmazonCoBuyComputerDataset()
-        >>> graph = dataset[0]
-
     Amazon Computers and Amazon Photo are segments of the Amazon co-purchase graph [McAuley et al., 2015],
     where nodes represent goods, edges indicate that two goods are frequently bought together, node
     features are bag-of-words encoded product reviews, and class labels are given by the product category.
@@ -394,8 +355,6 @@ class AmazonCoBuyComputerDataset(GNNBenchmarkDataset):
     ----------
     num_classes : int
         Number of classes for each node.
-    data : list
-        A list of DGLGraph objects
 
     Examples
     --------
@@ -426,13 +385,6 @@ def num_classes(self):
 class AmazonCoBuyPhotoDataset(GNNBenchmarkDataset):
     r"""AmazonCoBuy dataset for node classification task.
 
-    .. deprecated:: 0.5.0
-
-        - ``data`` is deprecated, it is repalced by:
-
-        >>> dataset = AmazonCoBuyPhotoDataset()
-        >>> graph = dataset[0]
-
     Amazon Computers and Amazon Photo are segments of the Amazon co-purchase graph [McAuley et al., 2015],
     where nodes represent goods, edges indicate that two goods are frequently bought together, node
     features are bag-of-words encoded product reviews, and class labels are given by the product category.
@@ -465,8 +417,6 @@ class AmazonCoBuyPhotoDataset(GNNBenchmarkDataset):
     ----------
     num_classes : int
         Number of classes for each node.
-    data : list
-        A list of DGLGraph objects
 
     Examples
     --------