[Doc/Feature] Refactor, doc update and behavior fix for graphs (#1983)

* Update graph

* Fix for dgl.graph

* from_scipy

* Replace canonical_etypes with relations

* from_networkx

* Update for hetero_from_relations

* Roll back the change of canonical_etypes to relations

* heterograph

* bipartite

* Update doc

* Fix lint

* Fix lint

* Fix test cases

* Fix

* Fix

* Fix

* Fix

* Fix

* Fix

* Update

* Fix test

* Fix

* Update

* Use DGLError

* Update

* Update

* Update

* Update

* Fix

* Fix

* Fix

* Fix

* Fix

* Fix

* Fix

* Fix

* Update

* Fix

* Update

* Fix

* Fix

* Fix

* Update

* Fix

* Update

* Fix

* Update

* Update

* Update

* Update

* Update

* Update

* Update

* Fix

* Fix

* Update

* Update

* Update

* Update

* Update

* Update

* rewrite sanity checks

* delete unnecessary checks

* Update

* Update

* Update

* Update

* Update

* Update

* Update

* Update

* Fix

* Update

* Update

* Update

* Fix

* Fix

* Fix

* Update

* Fix

* Update

* Fix

* Fix

* Update

* Fix

* Update

* Fix
Co-authored-by: xiang song(charlie.song) <classicxsong@gmail.com>
Co-authored-by: Minjie Wang <wmjlyjemaine@gmail.com>
Co-authored-by: Quan Gan <coin2028@hotmail.com>

python/dgl/nn/mxnet/conv/tagconv.py

@@ -106,7 +106,7 @@ class TAGConv(gluon.Block):
             is size of output feature.
         """
         with graph.local_scope():
-            assert graph.is_homogeneous(), 'Graph is not homogeneous'
+            assert graph.is_homogeneous, 'Graph is not homogeneous'
 
             degs = graph.in_degrees().astype('float32')
             norm = mx.nd.power(mx.nd.clip(degs, a_min=1, a_max=float("inf")), -0.5)

python/dgl/nn/pytorch/conv/gatedgraphconv.py

@@ -136,8 +136,9 @@ class GatedGraphConv(nn.Module):
             is the output feature size.
         """
         with graph.local_scope():
-            assert graph.is_homogeneous(), \
-                "not a homograph; convert it with to_homo and pass in the edge type as argument"
+            assert graph.is_homogeneous, \
+                "not a homogeneous graph; convert it with to_homogeneous " \
+                "and pass in the edge type as argument"
             assert etypes.min() >= 0 and etypes.max() < self._n_etypes, \
                 "edge type indices out of range [0, {})".format(self._n_etypes)
             zero_pad = feat.new_zeros((feat.shape[0], self._out_feats - feat.shape[1]))

python/dgl/nn/pytorch/conv/tagconv.py

@@ -111,7 +111,7 @@ class TAGConv(nn.Module):
             is size of output feature.
         """
         with graph.local_scope():
-            assert graph.is_homogeneous(), 'Graph is not homogeneous'
+            assert graph.is_homogeneous, 'Graph is not homogeneous'
 
             norm = th.pow(graph.in_degrees().float().clamp(min=1), -0.5)
             shp = norm.shape + (1,) * (feat.dim() - 1)

python/dgl/nn/tensorflow/conv/relgraphconv.py

@@ -270,8 +270,9 @@ class RelGraphConv(layers.Layer):
         tf.Tensor
             New node features.
         """
-        assert g.is_homogeneous(), \
-            "not a homograph; convert it with to_homo and pass in the edge type as argument"
+        assert g.is_homogeneous, \
+            "not a homogeneous graph; convert it with to_homogeneous " \
+            "and pass in the edge type as argument"
         with g.local_scope():
             g.ndata['h'] = x
             g.edata['type'] = tf.cast(etypes, tf.int64)

python/dgl/ops/segment.py

@@ -50,7 +50,7 @@ def segment_reduce(seglen, value, reducer='sum'):
     if len(u) != len(v):
         raise DGLError("Invalid seglen array:", seglen,
                        ". Its summation must be equal to value.shape[0].")
-    g = convert.bipartite((u, v))
+    g = convert.heterograph({('_U', '_E', '_V'): (u, v)})
     g.srcdata['h'] = value
     g.update_all(fn.copy_u('h', 'm'), getattr(fn, reducer)('m', 'h'))
     return g.dstdata['h']

python/dgl/sampling/pinsage.py

@@ -114,8 +114,10 @@ class RandomWalkNeighborSampler(object):
         dst = F.boolean_mask(dst, src_mask)
 
         # count the number of visits and pick the K-most frequent neighbors for each node
-        neighbor_graph = convert.graph(
-            (src, dst), num_nodes=self.G.number_of_nodes(self.ntype), ntype=self.ntype)
+        neighbor_graph = convert.heterograph(
+            {(self.ntype, '_E', self.ntype): (src, dst)},
+            {self.ntype: self.G.number_of_nodes(self.ntype)}
+        )
         neighbor_graph = transform.to_simple(neighbor_graph, return_counts=self.weight_column)
         counts = neighbor_graph.edata[self.weight_column]
         neighbor_graph = select_topk(neighbor_graph, self.num_neighbors, self.weight_column)
@@ -176,8 +178,8 @@ class PinSAGESampler(RandomWalkNeighborSampler):
 
     >>> g = scipy.sparse.random(3000, 5000, 0.003)
     >>> G = dgl.heterograph({
-    ...     ('A', 'AB', 'B'): g,
-    ...     ('B', 'BA', 'A'): g.T})
+    ...     ('A', 'AB', 'B'): g.nonzero(),
+    ...     ('B', 'BA', 'A'): g.T.nonzero()})
 
     Then we create a PinSage neighbor sampler that samples a graph of node type "A".  Each
     node would have (a maximum of) 10 neighbors.

python/dgl/sampling/randomwalks.py

@@ -80,8 +80,7 @@ def random_walk(g, nodes, *, metapath=None, length=None, prob=None, restart_prob
     Examples
     --------
     The following creates a homogeneous graph:
-
-    >>> g1 = dgl.graph([(0, 1), (1, 2), (1, 3), (2, 0), (3, 0)], 'user', 'follow')
+    >>> g1 = dgl.graph(([0, 1, 1, 2, 3], [1, 2, 3, 0, 0]))
 
     Normal random walk:
 
@@ -93,7 +92,7 @@ def random_walk(g, nodes, *, metapath=None, length=None, prob=None, restart_prob
 
     The first tensor indicates the random walk path for each seed node.
     The j-th element in the second tensor indicates the node type ID of the j-th node
-    in every path.  In this case, it is returning all 0 (``user``).
+    in every path.  In this case, it is returning all 0.
 
     Random walk with restart:
 
@@ -115,9 +114,9 @@ def random_walk(g, nodes, *, metapath=None, length=None, prob=None, restart_prob
     Metapath-based random walk:
 
     >>> g2 = dgl.heterograph({
-    ...     ('user', 'follow', 'user'): [(0, 1), (1, 2), (1, 3), (2, 0), (3, 0)],
-    ...     ('user', 'view', 'item'): [(0, 0), (0, 1), (1, 1), (2, 2), (3, 2), (3, 1)],
-    ...     ('item', 'viewed-by', 'user'): [(0, 0), (1, 0), (1, 1), (2, 2), (2, 3), (1, 3)]})
+    ...     ('user', 'follow', 'user'): ([0, 1, 1, 2, 3], [1, 2, 3, 0, 0]),
+    ...     ('user', 'view', 'item'): ([0, 0, 1, 2, 3, 3], [0, 1, 1, 2, 2, 1]),
+    ...     ('item', 'viewed-by', 'user'): ([0, 1, 1, 2, 2, 1], [0, 0, 1, 2, 3, 3])
     >>> dgl.sampling.random_walk(
     ...     g2, [0, 1, 2, 0], metapath=['follow', 'view', 'viewed-by'] * 2)
     (tensor([[0, 1, 1, 1, 2, 2, 3],
@@ -215,9 +214,9 @@ def pack_traces(traces, types):
     Examples
     --------
     >>> g2 = dgl.heterograph({
-    ...     ('user', 'follow', 'user'): [(0, 1), (1, 2), (1, 3), (2, 0), (3, 0)],
-    ...     ('user', 'view', 'item'): [(0, 0), (0, 1), (1, 1), (2, 2), (3, 2), (3, 1)],
-    ...     ('item', 'viewed-by', 'user'): [(0, 0), (1, 0), (1, 1), (2, 2), (2, 3), (1, 3)]})
+    ...     ('user', 'follow', 'user'): ([0, 1, 1, 2, 3], [1, 2, 3, 0, 0]),
+    ...     ('user', 'view', 'item'): ([0, 0, 1, 2, 3, 3], [0, 1, 1, 2, 2, 1]),
+    ...     ('item', 'viewed-by', 'user'): ([0, 1, 1, 2, 2, 1], [0, 0, 1, 2, 3, 3])
     >>> traces, types = dgl.sampling.random_walk(
     ...     g2, [0, 0], metapath=['follow', 'view', 'viewed-by'] * 2,
     ...     restart_prob=torch.FloatTensor([0, 0.5, 0, 0, 0.5, 0]))

python/dgl/subgraph.py

@@ -60,8 +60,9 @@ def node_subgraph(graph, nodes):
     Instantiate a heterograph.
 
     >>> g = dgl.heterograph({
-    ...     ('user', 'plays', 'game'): ([0, 1, 1, 2], [0, 0, 2, 1]),
-    ...     ('user', 'follows', 'user'): ([0, 1, 1], [1, 2, 2])})
+    >>>     ('user', 'plays', 'game'): ([0, 1, 1, 2], [0, 0, 2, 1]),
+    >>>     ('user', 'follows', 'user'): ([0, 1, 1], [1, 2, 2])
+    >>> })
     >>> # Set node features
     >>> g.nodes['user'].data['h'] = torch.tensor([[0.], [1.], [2.]])
 
@@ -119,7 +120,11 @@ def node_subgraph(graph, nodes):
             return F.astype(F.nonzero_1d(F.copy_to(v, graph.device)), graph.idtype)
         else:
             return utils.prepare_tensor(graph, v, 'nodes["{}"]'.format(ntype))
-    induced_nodes = [_process_nodes(ntype, nodes.get(ntype, [])) for ntype in graph.ntypes]
+
+    induced_nodes = []
+    for ntype in graph.ntypes:
+        nids = nodes.get(ntype, F.copy_to(F.tensor([], graph.idtype), graph.device))
+        induced_nodes.append(_process_nodes(ntype, nids))
     sgi = graph._graph.node_subgraph(induced_nodes)
     induced_edges = sgi.induced_edges
     return _create_hetero_subgraph(graph, sgi, induced_nodes, induced_edges)
@@ -176,8 +181,9 @@ def edge_subgraph(graph, edges, preserve_nodes=False):
     Instantiate a heterograph.
 
     >>> g = dgl.heterograph({
-    ...     ('user', 'plays', 'game'): ([0, 1, 1, 2], [0, 0, 2, 1]),
-    ...     ('user', 'follows', 'user'): ([0, 1, 1], [1, 2, 2])})
+    >>>     ('user', 'plays', 'game'): ([0, 1, 1, 2], [0, 0, 2, 1]),
+    >>>     ('user', 'follows', 'user'): ([0, 1, 1], [1, 2, 2])
+    >>> })
     >>> # Set edge features
     >>> g.edges['follows'].data['h'] = torch.tensor([[0.], [1.], [2.]])
 
@@ -238,9 +244,10 @@ def edge_subgraph(graph, edges, preserve_nodes=False):
             return utils.prepare_tensor(graph, e, 'edges["{}"]'.format(etype))
 
     edges = {graph.to_canonical_etype(etype): e for etype, e in edges.items()}
-    induced_edges = [
-        _process_edges(cetype, edges.get(cetype, []))
-        for cetype in graph.canonical_etypes]
+    induced_edges = []
+    for cetype in graph.canonical_etypes:
+        eids = edges.get(cetype, F.copy_to(F.tensor([], graph.idtype), graph.device))
+        induced_edges.append(_process_edges(cetype, eids))
     sgi = graph._graph.edge_subgraph(induced_edges, preserve_nodes)
     induced_nodes = sgi.induced_nodes
     return _create_hetero_subgraph(graph, sgi, induced_nodes, induced_edges)
@@ -452,8 +459,9 @@ def node_type_subgraph(graph, ntypes):
     Instantiate a heterograph.
 
     >>> g = dgl.heterograph({
-    ...     ('user', 'plays', 'game'): ([0, 1, 1, 2], [0, 0, 2, 1]),
-    ...     ('user', 'follows', 'user'): ([0, 1, 1], [1, 2, 2])})
+    >>>     ('user', 'plays', 'game'): ([0, 1, 1, 2], [0, 0, 2, 1]),
+    >>>     ('user', 'follows', 'user'): ([0, 1, 1], [1, 2, 2])
+    >>> })
     >>> # Set node features
     >>> g.nodes['user'].data['h'] = torch.tensor([[0.], [1.], [2.]])
 
@@ -519,8 +527,9 @@ def edge_type_subgraph(graph, etypes):
     Instantiate a heterograph.
 
     >>> g = dgl.heterograph({
-    ...     ('user', 'plays', 'game'): ([0, 1, 1, 2], [0, 0, 2, 1]),
-    ...     ('user', 'follows', 'user'): ([0, 1, 1], [1, 2, 2])})
+    >>>     ('user', 'plays', 'game'): ([0, 1, 1, 2], [0, 0, 2, 1]),
+    >>>     ('user', 'follows', 'user'): ([0, 1, 1], [1, 2, 2])
+    >>> })
     >>> # Set edge features
     >>> g.edges['follows'].data['h'] = torch.tensor([[0.], [1.], [2.]])
 
@@ -549,7 +558,7 @@ def edge_type_subgraph(graph, etypes):
     node_type_subgraph
     """
     etype_ids = [graph.get_etype_id(etype) for etype in etypes]
-    # meta graph is homograph, still using int64
+    # meta graph is homogeneous graph, still using int64
     meta_src, meta_dst, _ = graph._graph.metagraph.find_edges(utils.toindex(etype_ids, "int64"))
     rel_graphs = [graph._graph.get_relation_graph(i) for i in etype_ids]
     meta_src = meta_src.tonumpy()

python/dgl/transform.py

@@ -148,7 +148,7 @@ def knn_graph(x, k):
         (F.asnumpy(F.zeros_like(dst) + 1), (F.asnumpy(dst), F.asnumpy(src))),
         shape=(n_samples * n_points, n_samples * n_points))
 
-    return convert.graph(adj)
+    return convert.from_scipy(adj)
 
 #pylint: disable=invalid-name
 def segmented_knn_graph(x, k, segs):
@@ -220,8 +220,7 @@ def segmented_knn_graph(x, k, segs):
     src = F.reshape(src, (-1,))
     adj = sparse.csr_matrix((F.asnumpy(F.zeros_like(dst) + 1), (F.asnumpy(dst), F.asnumpy(src))))
 
-    g = convert.graph(adj)
-    return g
+    return convert.from_scipy(adj)
 
 def to_bidirected(g, readonly=None, copy_ndata=False):
     r"""Convert the graph to a bidirectional simple graph, adding reverse edges and
@@ -373,7 +372,7 @@ def add_reverse_edges(g, readonly=None, copy_ndata=True,
     --------
     **Homogeneous graphs**
 
-    >>> g = dgl.graph(th.tensor([0, 0]), th.tensor([0, 1]))
+    >>> g = dgl.graph((th.tensor([0, 0]), th.tensor([0, 1])))
     >>> bg1 = dgl.add_reverse_edges(g)
     >>> bg1.edges()
     (tensor([0, 0, 0, 1]), tensor([0, 1, 0, 0]))
@@ -381,10 +380,10 @@ def add_reverse_edges(g, readonly=None, copy_ndata=True,
     **Heterogeneous graphs with Multiple Edge Types**
 
     >>> g = dgl.heterograph({
-    ...     ('user', 'wins', 'user'): (th.tensor([0, 2, 0, 2, 2]), th.tensor([1, 1, 2, 1, 0])),
-    ...     ('user', 'plays', 'game'): (th.tensor([1, 2, 1]), th.tensor([2, 1, 1])),
-    ...     ('user', 'follows', 'user'): (th.tensor([1, 2, 1), th.tensor([0, 0, 0]))
-    ... })
+    >>>     ('user', 'wins', 'user'): (th.tensor([0, 2, 0, 2, 2]), th.tensor([1, 1, 2, 1, 0])),
+    >>>     ('user', 'plays', 'game'): (th.tensor([1, 2, 1]), th.tensor([2, 1, 1])),
+    >>>     ('user', 'follows', 'user'): (th.tensor([1, 2, 1), th.tensor([0, 0, 0]))
+    >>> })
     >>> g.nodes['game'].data['hv'] = th.ones(3, 1)
     >>> g.edges['wins'].data['h'] = th.tensor([0, 1, 2, 3, 4])
 
@@ -521,7 +520,7 @@ def line_graph(g, backtracking=True, shared=False):
     >>> lg.edges()
     (tensor([0, 1, 2, 4]), tensor([4, 0, 3, 1]))
     """
-    assert g.is_homogeneous(), \
+    assert g.is_homogeneous, \
         'only homogeneous graph is supported'
 
     dev = g.device
@@ -572,7 +571,7 @@ def khop_adj(g, k):
             [1., 3., 3., 1., 0.],
             [0., 1., 3., 3., 1.]])
     """
-    assert g.is_homogeneous(), \
+    assert g.is_homogeneous, \
         'only homogeneous graph is supported'
     adj_k = g.adj(scipy_fmt=g.formats()['created'][0]) ** k
     return F.tensor(adj_k.todense().astype(np.float32))
@@ -637,7 +636,7 @@ def khop_graph(g, k, copy_ndata=True):
              ndata_schemes={}
              edata_schemes={})
     """
-    assert g.is_homogeneous(), \
+    assert g.is_homogeneous, \
         'only homogeneous graph is supported'
     n = g.number_of_nodes()
     adj_k = g.adj(transpose=True, scipy_fmt=g.formats()['created'][0]) ** k
@@ -958,12 +957,9 @@ def metapath_reachable_graph(g, metapath):
     adj = (adj != 0).tocsr()
     srctype = g.to_canonical_etype(metapath[0])[0]
     dsttype = g.to_canonical_etype(metapath[-1])[2]
-    if srctype == dsttype:
-        assert adj.shape[0] == adj.shape[1]
-        new_g = convert.graph(adj, ntype=srctype, idtype=g.idtype, device=g.device)
-    else:
-        new_g = convert.bipartite(adj, utype=srctype, vtype=dsttype,
-                                  idtype=g.idtype, device=g.device)
+    new_g = convert.heterograph({(srctype, '_E', dsttype): adj.nonzero()},
+                                {srctype: adj.shape[0], dsttype: adj.shape[1]},
+                                idtype=g.idtype, device=g.device)
 
     # copy srcnode features
     new_g.nodes[srctype].data.update(g.nodes[srctype].data)
@@ -1516,7 +1512,8 @@ def compact_graphs(graphs, always_preserve=None, copy_ndata=True, copy_edata=Tru
     The following code constructs a bipartite graph with 20 users and 10 games, but
     only user #1 and #3, as well as game #3 and #5, have connections:
 
-    >>> g = dgl.bipartite([(1, 3), (3, 5)], 'user', 'plays', 'game', num_nodes=(20, 10))
+    >>> g = dgl.heterograph({('user', 'plays', 'game'): ([1, 3], [3, 5])},
+    >>>                      {'user': 20, 'game': 10})
 
     The following would compact the graph above to another bipartite graph with only
     two users and two games.
@@ -1538,7 +1535,8 @@ def compact_graphs(graphs, always_preserve=None, copy_ndata=True, copy_edata=Tru
     of the given graphs are removed.  So if you compact ``g`` and the following ``g2``
     graphs together:
 
-    >>> g2 = dgl.bipartite([(1, 6), (6, 8)], 'user', 'plays', 'game', num_nodes=(20, 10))
+    >>> g2 = dgl.heterograph({('user', 'plays', 'game'): ([1, 6], [6, 8])},
+    >>>                      {'user': 20, 'game': 10})
     >>> (new_g, new_g2), induced_nodes = dgl.compact_graphs([g, g2])
     >>> induced_nodes
     {'user': tensor([1, 3, 6]), 'game': tensor([3, 5, 6, 8])}
@@ -1671,8 +1669,7 @@ def to_block(g, dst_nodes=None, include_dst_in_src=True):
     Examples
     --------
     Converting a homogeneous graph to a block as described above:
-
-    >>> g = dgl.graph([(0, 1), (1, 2), (2, 3)])
+    >>> g = dgl.graph(([0, 1, 2], [1, 2, 3]))
     >>> block = dgl.to_block(g, torch.LongTensor([3, 2]))
 
     The output nodes would be exactly the same as the ones given: [3, 2].
@@ -1708,7 +1705,7 @@ def to_block(g, dst_nodes=None, include_dst_in_src=True):
     Converting a heterogeneous graph to a block is similar, except that when specifying
     the output nodes, you have to give a dict:
 
-    >>> g = dgl.bipartite([(0, 1), (1, 2), (2, 3)], utype='A', vtype='B')
+    >>> g = dgl.heterograph({('A', '_E', 'B'): ([0, 1, 2], [1, 2, 3])})
 
     If you don't specify any node of type A on the output side, the node type ``A``
     in the block would have zero nodes on the output side.

python/dgl/traversal.py

@@ -36,7 +36,7 @@ def bfs_nodes_generator(graph, source, reverse=False):
              / \\
         0 - 1 - 3 - 5
 
-    >>> g = dgl.graph([(0, 1), (1, 2), (1, 3), (2, 3), (2, 4), (3, 5)])
+    >>> g = dgl.graph(([0, 1, 1, 2, 2, 3], [1, 2, 3, 3, 4, 5]))
     >>> list(dgl.bfs_nodes_generator(g, 0))
     [tensor([0]), tensor([1]), tensor([2, 3]), tensor([4, 5])]
     """
@@ -80,7 +80,7 @@ def bfs_edges_generator(graph, source, reverse=False):
              / \\
         0 - 1 - 3 - 5
 
-    >>> g = dgl.graph([(0, 1), (1, 2), (1, 3), (2, 3), (2, 4), (3, 5)])
+    >>> g = dgl.graph(([0, 1, 1, 2, 2, 3], [1, 2, 3, 3, 4, 5]))
     >>> list(dgl.bfs_edges_generator(g, 0))
     [tensor([0]), tensor([1, 2]), tensor([4, 5])]
     """
@@ -121,7 +121,7 @@ def topological_nodes_generator(graph, reverse=False):
              / \\
         0 - 1 - 3 - 5
 
-    >>> g = dgl.graph([(0, 1), (1, 2), (1, 3), (2, 3), (2, 4), (3, 5)])
+    >>> g = dgl.graph(([0, 1, 1, 2, 2, 3], [1, 2, 3, 3, 4, 5]))
     >>> list(dgl.topological_nodes_generator(g))
     [tensor([0]), tensor([1]), tensor([2]), tensor([3, 4]), tensor([5])]
     """
@@ -169,7 +169,7 @@ def dfs_edges_generator(graph, source, reverse=False):
 
     Edge addition order [(0, 1), (1, 2), (1, 3), (2, 3), (2, 4), (3, 5)]
 
-    >>> g = dgl.graph([(0, 1), (1, 2), (1, 3), (2, 3), (2, 4), (3, 5)])
+    >>> g = dgl.graph(([0, 1, 1, 2, 2, 3], [1, 2, 3, 3, 4, 5]))
     >>> list(dgl.dfs_edges_generator(g, 0))
     [tensor([0]), tensor([1]), tensor([3]), tensor([5]), tensor([4])]
     """
@@ -243,7 +243,7 @@ def dfs_labeled_edges_generator(
 
     Edge addition order [(0, 1), (1, 2), (1, 3), (2, 3), (2, 4), (3, 5)]
 
-    >>> g = dgl.graph([(0, 1), (1, 2), (1, 3), (2, 3), (2, 4), (3, 5)])
+    >>> g = dgl.graph(([0, 1, 1, 2, 2, 3], [1, 2, 3, 3, 4, 5]))
     >>> list(dgl.dfs_labeled_edges_generator(g, 0, has_nontree_edge=True))
     (tensor([0]), tensor([1]), tensor([3]), tensor([5]), tensor([4]), tensor([2])),
     (tensor([0]), tensor([0]), tensor([0]), tensor([0]), tensor([0]), tensor([2]))

python/dgl/utils/checks.py

@@ -27,7 +27,6 @@ def prepare_tensor(g, data, name):
     Tensor
         Data in tensor object.
     """
-    ret = None
     if F.is_tensor(data):
         if F.dtype(data) != g.idtype or F.context(data) != g.device:
             raise DGLError('Expect argument "{}" to have data type {} and device '
@@ -35,9 +34,15 @@ def prepare_tensor(g, data, name):
                                name, g.idtype, g.device, F.dtype(data), F.context(data)))
         ret = data
     else:
-        ret = F.copy_to(F.tensor(data, g.idtype), g.device)
-
-    if F.ndim(ret) != 1:
+        data = F.tensor(data)
+        if F.dtype(data) not in (F.int32, F.int64):
+            raise DGLError('Expect argument "{}" to have data type int32 or int64,'
+                           ' but got {}.'.format(name, F.dtype(data)))
+        ret = F.copy_to(F.astype(data, g.idtype), g.device)
+
+    if F.ndim(ret) == 0:
+        ret = F.unsqueeze(ret, 0)
+    if F.ndim(ret) > 1:
         raise DGLError('Expect a 1-D tensor for argument "{}". But got {}.'.format(
             name, ret))
     return ret
@@ -158,3 +163,15 @@ def check_all_same_schema(feat_dict_list, keys, name):
                                ' and feature size, but got\n\t{} {}\nand\n\t{} {}.'.format(
                                    name, k, F.dtype(t1), F.shape(t1)[1:],
                                    F.dtype(t2), F.shape(t2)[1:]))
+
+def check_valid_idtype(idtype):
+    """Check whether the value of the idtype argument is valid (int32/int64)
+
+    Parameters
+    ----------
+    idtype : data type
+        The framework object of a data type.
+    """
+    if idtype not in [None, F.int32, F.int64]:
+        raise DGLError('Expect idtype to be a framework object of int32/int64, '
+                       'got {}'.format(idtype))

python/dgl/utils/data.py

@@ -5,6 +5,7 @@ import networkx as nx
 
 from ..base import DGLError
 from .. import backend as F
+from . import checks
 
 def elist2tensor(elist, idtype):
     """Function to convert an edge list to edge tensors.
@@ -49,7 +50,7 @@ def scipy2tensor(spmat, idtype):
     col = F.tensor(spmat.col, idtype)
     return row, col
 
-def networkx2tensor(nx_graph, idtype, edge_id_attr_name='id'):
+def networkx2tensor(nx_graph, idtype, edge_id_attr_name=None):
     """Function to convert a networkx graph to edge tensors.
 
     Parameters
@@ -60,7 +61,7 @@ def networkx2tensor(nx_graph, idtype, edge_id_attr_name='id'):
         Integer ID type. Must be int32 or int64.
     edge_id_attr_name : str, optional
         Key name for edge ids in the NetworkX graph. If not found, we
-        will consider the graph not to have pre-specified edge ids. (Default: 'id')
+        will consider the graph not to have pre-specified edge ids. (Default: None)
 
     Returns
     -------
@@ -72,19 +73,17 @@ def networkx2tensor(nx_graph, idtype, edge_id_attr_name='id'):
 
     # Relabel nodes using consecutive integers
     nx_graph = nx.convert_node_labels_to_integers(nx_graph, ordering='sorted')
-
-    # nx_graph.edges(data=True) returns src, dst, attr_dict
-    if nx_graph.number_of_edges() > 0:
-        has_edge_id = edge_id_attr_name in next(iter(nx_graph.edges(data=True)))[-1]
-    else:
-        has_edge_id = False
+    has_edge_id = edge_id_attr_name is not None
 
     if has_edge_id:
         num_edges = nx_graph.number_of_edges()
         src = [0] * num_edges
         dst = [0] * num_edges
         for u, v, attr in nx_graph.edges(data=True):
-            eid = attr[edge_id_attr_name]
+            eid = int(attr[edge_id_attr_name])
+            if eid < 0 or eid >= nx_graph.number_of_edges():
+                raise DGLError('Expect edge IDs to be a non-negative integer smaller than {:d}, '
+                               'got {:d}'.format(num_edges, eid))
             src[eid] = u
             dst[eid] = v
     else:
@@ -97,7 +96,7 @@ def networkx2tensor(nx_graph, idtype, edge_id_attr_name='id'):
     dst = F.tensor(dst, idtype)
     return src, dst
 
-def graphdata2tensors(data, idtype=None, bipartite=False):
+def graphdata2tensors(data, idtype=None, bipartite=False, **kwargs):
     """Function to convert various types of data to edge tensors and infer
     the number of nodes.
 
@@ -111,6 +110,14 @@ def graphdata2tensors(data, idtype=None, bipartite=False):
     bipartite : bool, optional
         Whether infer number of nodes of a bipartite graph --
         num_src and num_dst can be different.
+    kwargs
+
+        - edge_id_attr_name : The name (str) of the edge attribute that stores the edge
+          IDs in the NetworkX graph.
+        - top_map : The dictionary mapping the original IDs of the source nodes to the
+          new ones.
+        - bottom_map : The dictionary mapping the original IDs of the destination nodes
+          to the new ones.
 
     Returns
     -------
@@ -127,19 +134,62 @@ def graphdata2tensors(data, idtype=None, bipartite=False):
         # preferred default idtype is int64
         # if data is tensor and idtype is None, infer the idtype from tensor
         idtype = F.int64
+    checks.check_valid_idtype(idtype)
+
+    if isinstance(data, tuple) and (not F.is_tensor(data[0]) or not F.is_tensor(data[1])):
+        # (Iterable, Iterable) type data, convert it to (Tensor, Tensor)
+        if len(data[0]) == 0:
+            # force idtype for empty list
+            data = F.tensor(data[0], idtype), F.tensor(data[1], idtype)
+        else:
+            # convert the iterable to tensor and keep its native data type so we can check
+            # its validity later
+            data = F.tensor(data[0]), F.tensor(data[1])
+
     if isinstance(data, tuple):
-        src, dst = F.tensor(data[0], idtype), F.tensor(data[1], idtype)
+        # (Tensor, Tensor) type data
+        src, dst = data
+        # sanity checks
+        # TODO(minjie): move these checks to C for faster graph construction.
+        if F.dtype(src) != F.dtype(dst):
+            raise DGLError('Expect the source and destination node IDs to have the same type,'
+                           ' but got {} and {}.'.format(F.dtype(src), F.dtype(dst)))
+        if F.context(src) != F.context(dst):
+            raise DGLError('Expect the source and destination node IDs to be on the same device,'
+                           ' but got {} and {}.'.format(F.context(src), F.context(dst)))
+        if F.dtype(src) not in (F.int32, F.int64):
+            raise DGLError('Expect the source ID tensor to have data type int32 or int64,'
+                           ' but got {}.'.format(F.dtype(src)))
+        if F.dtype(dst) not in (F.int32, F.int64):
+            raise DGLError('Expect the destination ID tensor to have data type int32 or int64,'
+                           ' but got {}.'.format(F.dtype(dst)))
+        if idtype is not None:
+            src, dst = F.astype(src, idtype), F.astype(dst, idtype)
     elif isinstance(data, list):
         src, dst = elist2tensor(data, idtype)
     elif isinstance(data, sp.sparse.spmatrix):
         src, dst = scipy2tensor(data, idtype)
     elif isinstance(data, nx.Graph):
+        edge_id_attr_name = kwargs.get('edge_id_attr_name', None)
         if bipartite:
-            src, dst = networkxbipartite2tensors(data, idtype)
+            top_map = kwargs.get('top_map')
+            bottom_map = kwargs.get('bottom_map')
+            src, dst = networkxbipartite2tensors(
+                data, idtype, top_map=top_map,
+                bottom_map=bottom_map, edge_id_attr_name=edge_id_attr_name)
         else:
-            src, dst = networkx2tensor(data, idtype)
+            src, dst = networkx2tensor(
+                data, idtype, edge_id_attr_name=edge_id_attr_name)
     else:
         raise DGLError('Unsupported graph data type:', type(data))
+
+    if len(src) != len(dst):
+        raise DGLError('Expect the source and destination ID tensors to have the same length,'
+                       ' but got {} and {}.'.format(len(src), len(dst)))
+    if len(src) > 0 and (F.as_scalar(F.min(src, 0)) < 0 or F.as_scalar(F.min(dst, 0)) < 0):
+        raise DGLError('All IDs must be non-negative integers.')
+
+    # infer number of nodes
     infer_from_raw = infer_num_nodes(data, bipartite=bipartite)
     if infer_from_raw is None:
         num_src, num_dst = infer_num_nodes((src, dst), bipartite=bipartite)
@@ -147,7 +197,7 @@ def graphdata2tensors(data, idtype=None, bipartite=False):
         num_src, num_dst = infer_from_raw
     return src, dst, num_src, num_dst
 
-def networkxbipartite2tensors(nx_graph, idtype, edge_id_attr_name='id'):
+def networkxbipartite2tensors(nx_graph, idtype, top_map, bottom_map, edge_id_attr_name=None):
     """Function to convert a networkx bipartite to edge tensors.
 
     Parameters
@@ -155,49 +205,54 @@ def networkxbipartite2tensors(nx_graph, idtype, edge_id_attr_name='id'):
     nx_graph : nx.Graph
         NetworkX graph. It must follow the bipartite graph convention of networkx.
         Each node has an attribute ``bipartite`` with values 0 and 1 indicating
-        which set it belongs to. Only edges from node set 0 to node set 1 are
-        added to the returned graph.
+        which set it belongs to.
+    top_map : dict
+        The dictionary mapping the original node labels to the node IDs for the source type.
+    bottom_map : dict
+        The dictionary mapping the original node labels to the node IDs for the destination type.
     idtype : int32, int64, optional
         Integer ID type. Must be int32 or int64.
     edge_id_attr_name : str, optional
         Key name for edge ids in the NetworkX graph. If not found, we
-        will consider the graph not to have pre-specified edge ids. (Default: 'id')
+        will consider the graph not to have pre-specified edge ids. (Default: None)
 
     Returns
     -------
     (Tensor, Tensor)
         Edge tensors.
     """
-    if not nx_graph.is_directed():
-        nx_graph = nx_graph.to_directed()
-
-    top_nodes = {n for n, d in nx_graph.nodes(data=True) if d['bipartite'] == 0}
-    bottom_nodes = set(nx_graph) - top_nodes
-    top_nodes = sorted(top_nodes)
-    bottom_nodes = sorted(bottom_nodes)
-    top_map = {n : i for i, n in enumerate(top_nodes)}
-    bottom_map = {n : i for i, n in enumerate(bottom_nodes)}
-
-    if nx_graph.number_of_edges() > 0:
-        has_edge_id = edge_id_attr_name in next(iter(nx_graph.edges(data=True)))[-1]
-    else:
-        has_edge_id = False
+    has_edge_id = edge_id_attr_name is not None
 
     if has_edge_id:
         num_edges = nx_graph.number_of_edges()
         src = [0] * num_edges
         dst = [0] * num_edges
         for u, v, attr in nx_graph.edges(data=True):
-            eid = attr[edge_id_attr_name]
+            if u not in top_map:
+                raise DGLError('Expect the node {} to have attribute bipartite=0 '
+                               'with edge {}'.format(u, (u, v)))
+            if v not in bottom_map:
+                raise DGLError('Expect the node {} to have attribute bipartite=1 '
+                               'with edge {}'.format(v, (u, v)))
+            eid = int(attr[edge_id_attr_name])
+            if eid < 0 or eid >= nx_graph.number_of_edges():
+                raise DGLError('Expect edge IDs to be a non-negative integer smaller than {:d}, '
+                               'got {:d}'.format(num_edges, eid))
             src[eid] = top_map[u]
             dst[eid] = bottom_map[v]
     else:
         src = []
         dst = []
         for e in nx_graph.edges:
-            if e[0] in top_map:
-                src.append(top_map[e[0]])
-                dst.append(bottom_map[e[1]])
+            u, v = e[0], e[1]
+            if u not in top_map:
+                raise DGLError('Expect the node {} to have attribute bipartite=0 '
+                               'with edge {}'.format(u, (u, v)))
+            if v not in bottom_map:
+                raise DGLError('Expect the node {} to have attribute bipartite=1 '
+                               'with edge {}'.format(v, (u, v)))
+            src.append(top_map[u])
+            dst.append(bottom_map[v])
     src = F.tensor(src, dtype=idtype)
     dst = F.tensor(dst, dtype=idtype)
     return src, dst

src/graph/heterograph.cc

@@ -420,7 +420,7 @@ FlattenedHeteroGraphPtr HeteroGraph::FlattenImpl(const std::vector<dgl_type_t>&
       dsttype_set.push_back(dsttype);
     }
   }
-  // Sort the node types so that we can compare the sets and decide whether a homograph
+  // Sort the node types so that we can compare the sets and decide whether a homogeneous graph
   // should be returned.
   std::sort(srctype_set.begin(), srctype_set.end());
   std::sort(dsttype_set.begin(), dsttype_set.end());

src/graph/transform/partition_hetero.cc

@@ -52,7 +52,7 @@ HaloHeteroSubgraph GetSubgraphWithHalo(std::shared_ptr<HeteroGraph> hg,
                                        IdArray nodes, int num_hops) {
   CHECK_EQ(hg->NumBits(), 64) << "halo subgraph only supports 64bits graph";
   CHECK_EQ(hg->relation_graphs().size(), 1)
-    << "halo subgraph only supports homograph";
+    << "halo subgraph only supports homogeneous graph";
   CHECK_EQ(nodes->dtype.bits, 64)
     << "halo subgraph only supports 64bits nodes tensor";
   const dgl_id_t *nid = static_cast<dgl_id_t *>(nodes->data);

tests/compute/test_basics.py

@@ -145,30 +145,19 @@ def test_batch_setter_getter(idtype):
     truth = [0.] * 17
     truth[0] = truth[4] = truth[3] = truth[9] = truth[16] = 1.
     assert _pfc(g.edata['l']) == truth
-    # set partial edges (many-one)
     u = F.tensor([3, 4, 6], g.idtype)
-    v = F.tensor([9], g.idtype)
+    v = F.tensor([9, 9, 9], g.idtype)
     g.edges[u, v].data['l'] = F.ones((3, D))
     truth[5] = truth[7] = truth[11] = 1.
     assert _pfc(g.edata['l']) == truth
-    # set partial edges (one-many)
-    u = F.tensor([0], g.idtype)
+    u = F.tensor([0, 0, 0], g.idtype)
     v = F.tensor([4, 5, 6], g.idtype)
     g.edges[u, v].data['l'] = F.ones((3, D))
     truth[6] = truth[8] = truth[10] = 1.
     assert _pfc(g.edata['l']) == truth
-    # get partial edges (many-many)
     u = F.tensor([0, 6, 0], g.idtype)
     v = F.tensor([6, 9, 7], g.idtype)
-    assert _pfc(g.edges[u, v].data['l']) == [1., 1., 0.]
-    # get partial edges (many-one)
-    u = F.tensor([5, 6, 7], g.idtype)
-    v = F.tensor([9], g.idtype)
-    assert _pfc(g.edges[u, v].data['l']) == [1., 1., 0.]
-    # get partial edges (one-many)
-    u = F.tensor([0], g.idtype)
-    v = F.tensor([3, 4, 5], g.idtype)
-    assert _pfc(g.edges[u, v].data['l']) == [1., 1., 1.]
+    assert _pfc(g.edges[u, v].data['l']) == [1.0, 1.0, 0.0]
 
 @parametrize_dtype
 def test_batch_setter_autograd(idtype):
@@ -221,9 +210,7 @@ def _test_nx_conversion():
     n3 = F.randn((5, 4))
     e1 = F.randn((4, 5))
     e2 = F.randn((4, 7))
-    g = DGLGraph()
-    g.add_nodes(5)
-    g.add_edges([0,1,3,4], [2,4,0,3])
+    g = dgl.graph(([0, 1, 3, 4], [2, 4, 0, 3]))
     g.ndata.update({'n1': n1, 'n2': n2, 'n3': n3})
     g.edata.update({'e1': e1, 'e2': e2})
 
@@ -369,7 +356,7 @@ def test_update_routines(idtype):
 @parametrize_dtype
 def test_update_all_0deg(idtype):
     # test#1
-    g = dgl.graph([(1,0), (2,0), (3,0), (4,0)], idtype=idtype, device=F.ctx())
+    g = dgl.graph(([1, 2, 3, 4], [0, 0, 0, 0]), idtype=idtype, device=F.ctx())
     def _message(edges):
         return {'m' : edges.src['h']}
     def _reduce(nodes):
@@ -390,7 +377,7 @@ def test_update_all_0deg(idtype):
     assert F.allclose(new_repr[0], 2 * F.sum(old_repr, 0))
 
     # test#2: graph with no edge
-    g = dgl.graph([], num_nodes=5, idtype=idtype, device=F.ctx())
+    g = dgl.graph(([], []), num_nodes=5, idtype=idtype, device=F.ctx())
     g.ndata['h'] = old_repr
     g.update_all(_message, _reduce, lambda nodes : {'h' : nodes.data['h'] * 2})
     new_repr = g.ndata['h']
@@ -399,7 +386,7 @@ def test_update_all_0deg(idtype):
 
 @parametrize_dtype
 def test_pull_0deg(idtype):
-    g = dgl.graph([(0,1)], idtype=idtype, device=F.ctx())
+    g = dgl.graph(([0], [1]), idtype=idtype, device=F.ctx())
     def _message(edges):
         return {'m' : edges.src['h']}
     def _reduce(nodes):
@@ -465,7 +452,7 @@ def test_dynamic_addition():
 
 @parametrize_dtype
 def test_repr(idtype):
-    g = dgl.graph([(0,1), (0,2), (1,2)], num_nodes=10, idtype=idtype, device=F.ctx())
+    g = dgl.graph(([0, 0, 1], [1, 2, 2]), num_nodes=10, idtype=idtype, device=F.ctx())
     repr_string = g.__repr__()
     print(repr_string)
     g.ndata['x'] = F.zeros((10, 5))
@@ -475,7 +462,7 @@ def test_repr(idtype):
 
 @parametrize_dtype
 def test_local_var(idtype):
-    g = dgl.graph([(0,1), (1,2), (2,3), (3,4)], idtype=idtype, device=F.ctx())
+    g = dgl.graph(([0, 1, 2, 3], [1, 2, 3, 4]), idtype=idtype, device=F.ctx())
     g.ndata['h'] = F.zeros((g.number_of_nodes(), 3))
     g.edata['w'] = F.zeros((g.number_of_edges(), 4))
     # test override
@@ -512,7 +499,7 @@ def test_local_var(idtype):
     assert 'ww' not in g.edata
 
     # test initializer1
-    g = dgl.graph([(0,1), (1,1)], idtype=idtype, device=F.ctx())
+    g = dgl.graph(([0, 1], [1, 1]), idtype=idtype, device=F.ctx())
     g.set_n_initializer(dgl.init.zero_initializer)
     def foo(g):
         g = g.local_var()
@@ -533,7 +520,7 @@ def test_local_var(idtype):
 
 @parametrize_dtype
 def test_local_scope(idtype):
-    g = dgl.graph([(0,1), (1,2), (2,3), (3,4)], idtype=idtype, device=F.ctx())
+    g = dgl.graph(([0, 1, 2, 3], [1, 2, 3, 4]), idtype=idtype, device=F.ctx())
     g.ndata['h'] = F.zeros((g.number_of_nodes(), 3))
     g.edata['w'] = F.zeros((g.number_of_edges(), 4))
     # test override
@@ -584,7 +571,7 @@ def test_local_scope(idtype):
     assert 'ww' not in g.edata
 
     # test initializer1
-    g = dgl.graph([(0,1), (1,1)], idtype=idtype, device=F.ctx())
+    g = dgl.graph(([0, 1], [1, 1]), idtype=idtype, device=F.ctx())
     g.set_n_initializer(dgl.init.zero_initializer)
     def foo(g):
         with g.local_scope():
@@ -605,29 +592,27 @@ def test_local_scope(idtype):
 
 @parametrize_dtype
 def test_isolated_nodes(idtype):
-    g = dgl.graph([(0, 1), (1, 2)], num_nodes=5, idtype=idtype, device=F.ctx())
-    assert g.number_of_nodes() == 5
-
-    # Test backward compatibility
-    g = dgl.graph([(0, 1), (1, 2)], card=5, idtype=idtype, device=F.ctx())
+    g = dgl.graph(([0, 1], [1, 2]), num_nodes=5, idtype=idtype, device=F.ctx())
     assert g.number_of_nodes() == 5
 
-    g = dgl.bipartite([(0, 2), (0, 3), (1, 2)], 'user', 'plays',
-                      'game', num_nodes=(5, 7), idtype=idtype, device=F.ctx())
+    g = dgl.heterograph({
+        ('user', 'plays', 'game'): ([0, 0, 1], [2, 3, 2])
+    }, {'user': 5, 'game': 7}, idtype=idtype, device=F.ctx())
     assert g.idtype == idtype
     assert g.number_of_nodes('user') == 5
     assert g.number_of_nodes('game') == 7
 
     # Test backward compatibility
-    g = dgl.bipartite([(0, 2), (0, 3), (1, 2)], 'user', 'plays',
-                      'game', card=(5, 7), idtype=idtype, device=F.ctx())
+    g = dgl.heterograph({
+        ('user', 'plays', 'game'): ([0, 0, 1], [2, 3, 2])
+    }, {'user': 5, 'game': 7}, idtype=idtype, device=F.ctx())
     assert g.idtype == idtype
     assert g.number_of_nodes('user') == 5
     assert g.number_of_nodes('game') == 7
 
 @parametrize_dtype
 def test_send_multigraph(idtype):
-    g = dgl.graph([(0,1), (0,1), (0,1), (2,1)], idtype=idtype, device=F.ctx())
+    g = dgl.graph(([0, 0, 0, 2], [1, 1, 1, 1]), idtype=idtype, device=F.ctx())
 
     def _message_a(edges):
         return {'a': edges.data['a']}

tests/compute/test_batched_graph.py

@@ -12,7 +12,7 @@ def tree1(idtype):
      3   4
     Edges are from leaves to root.
     """
-    g = dgl.DGLGraph().astype(idtype).to(F.ctx())
+    g = dgl.graph(([], [])).astype(idtype).to(F.ctx())
     g.add_nodes(5)
     g.add_edge(3, 1)
     g.add_edge(4, 1)
@@ -31,7 +31,7 @@ def tree2(idtype):
      2   0
     Edges are from leaves to root.
     """
-    g = dgl.DGLGraph().astype(idtype).to(F.ctx())
+    g = dgl.graph(([], [])).astype(idtype).to(F.ctx())
     g.add_nodes(5)
     g.add_edge(2, 4)
     g.add_edge(0, 4)
@@ -120,10 +120,10 @@ def test_batch_unbatch_frame(idtype):
 @parametrize_dtype
 def test_batch_unbatch2(idtype):
     # test setting/getting features after batch
-    a = dgl.DGLGraph().astype(idtype).to(F.ctx())
+    a = dgl.graph(([], [])).astype(idtype).to(F.ctx())
     a.add_nodes(4)
     a.add_edges(0, [1, 2, 3])
-    b = dgl.DGLGraph().astype(idtype).to(F.ctx())
+    b = dgl.graph(([], [])).astype(idtype).to(F.ctx())
     b.add_nodes(3)
     b.add_edges(0, [1, 2])
     c = dgl.batch([a, b])
@@ -179,12 +179,12 @@ def test_batch_propagate(idtype):
 
 @parametrize_dtype
 def test_batched_edge_ordering(idtype):
-    g1 = dgl.DGLGraph().astype(idtype).to(F.ctx())
+    g1 = dgl.graph(([], [])).astype(idtype).to(F.ctx())
     g1.add_nodes(6)
     g1.add_edges([4, 4, 2, 2, 0], [5, 3, 3, 1, 1])
     e1 = F.randn((5, 10))
     g1.edata['h'] = e1
-    g2 = dgl.DGLGraph().astype(idtype).to(F.ctx())
+    g2 = dgl.graph(([], [])).astype(idtype).to(F.ctx())
     g2.add_nodes(6)
     g2.add_edges([0, 1 ,2 ,5, 4 ,5], [1, 2, 3, 4, 3, 0])
     e2 = F.randn((6, 10))
@@ -196,13 +196,13 @@ def test_batched_edge_ordering(idtype):
 
 @parametrize_dtype
 def test_batch_no_edge(idtype):
-    g1 = dgl.DGLGraph().astype(idtype).to(F.ctx())
+    g1 = dgl.graph(([], [])).astype(idtype).to(F.ctx())
     g1.add_nodes(6)
     g1.add_edges([4, 4, 2, 2, 0], [5, 3, 3, 1, 1])
-    g2 = dgl.DGLGraph().astype(idtype).to(F.ctx())
+    g2 = dgl.graph(([], [])).astype(idtype).to(F.ctx())
     g2.add_nodes(6)
     g2.add_edges([0, 1, 2, 5, 4, 5], [1 ,2 ,3, 4, 3, 0])
-    g3 = dgl.DGLGraph().astype(idtype).to(F.ctx())
+    g3 = dgl.graph(([], [])).astype(idtype).to(F.ctx())
     g3.add_nodes(1)  # no edges
     g = dgl.batch([g1, g3, g2]) # should not throw an error
 

tests/compute/test_batched_heterograph.py

@@ -44,14 +44,14 @@ def check_equivalence_between_heterographs(g1, g2, node_attrs=None, edge_attrs=N
 def test_topology(idtype):
     """Test batching two DGLHeteroGraphs where some nodes are isolated in some relations"""
     g1 = dgl.heterograph({
-        ('user', 'follows', 'user'): [(0, 1), (1, 2)],
-        ('user', 'follows', 'developer'): [(0, 1), (1, 2)],
-        ('user', 'plays', 'game'): [(0, 0), (1, 0), (2, 1), (3, 1)]
+        ('user', 'follows', 'user'): ([0, 1], [1, 2]),
+        ('user', 'follows', 'developer'): ([0, 1], [1, 2]),
+        ('user', 'plays', 'game'): ([0, 1, 2, 3], [0, 0, 1, 1])
     }, idtype=idtype, device=F.ctx())
     g2 = dgl.heterograph({
-        ('user', 'follows', 'user'): [(0, 1), (1, 2)],
-        ('user', 'follows', 'developer'): [(0, 1), (1, 2)],
-        ('user', 'plays', 'game'): [(0, 0), (1, 0), (2, 1)]
+        ('user', 'follows', 'user'): ([0, 1], [1, 2]),
+        ('user', 'follows', 'developer'): ([0, 1], [1, 2]),
+        ('user', 'plays', 'game'): ([0, 1, 2], [0, 0, 1])
     }, idtype=idtype, device=F.ctx())
     bg = dgl.batch([g1, g2])
 
@@ -113,17 +113,17 @@ def test_topology(idtype):
 def test_batching_batched(idtype):
     """Test batching a DGLHeteroGraph and a BatchedDGLHeteroGraph."""
     g1 = dgl.heterograph({
-        ('user', 'follows', 'user'): [(0, 1), (1, 2)],
-        ('user', 'plays', 'game'): [(0, 0), (1, 0)]
+        ('user', 'follows', 'user'): ([0, 1], [1, 2]),
+        ('user', 'plays', 'game'): ([0, 1], [0, 0])
     }, idtype=idtype, device=F.ctx())
     g2 = dgl.heterograph({
-        ('user', 'follows', 'user'): [(0, 1), (1, 2)],
-        ('user', 'plays', 'game'): [(0, 0), (1, 0)]
+        ('user', 'follows', 'user'): ([0, 1], [1, 2]),
+        ('user', 'plays', 'game'): ([0, 1], [0, 0])
     }, idtype=idtype, device=F.ctx())
     bg1 = dgl.batch([g1, g2])
     g3 = dgl.heterograph({
-        ('user', 'follows', 'user'): [(0, 1)],
-        ('user', 'plays', 'game'): [(1, 0)]
+        ('user', 'follows', 'user'): ([0], [1]),
+        ('user', 'plays', 'game'): ([1], [0])
     }, idtype=idtype, device=F.ctx())
     bg2 = dgl.batch([bg1, g3])
     assert bg2.idtype == idtype
@@ -169,8 +169,8 @@ def test_batching_batched(idtype):
 def test_features(idtype):
     """Test the features of batched DGLHeteroGraphs"""
     g1 = dgl.heterograph({
-        ('user', 'follows', 'user'): [(0, 1), (1, 2)],
-        ('user', 'plays', 'game'): [(0, 0), (1, 0)]
+        ('user', 'follows', 'user'): ([0, 1], [1, 2]),
+        ('user', 'plays', 'game'): ([0, 1], [0, 0])
     }, idtype=idtype, device=F.ctx())
     g1.nodes['user'].data['h1'] = F.tensor([[0.], [1.], [2.]])
     g1.nodes['user'].data['h2'] = F.tensor([[3.], [4.], [5.]])
@@ -181,8 +181,8 @@ def test_features(idtype):
     g1.edges['plays'].data['h1'] = F.tensor([[0.], [1.]])
 
     g2 = dgl.heterograph({
-        ('user', 'follows', 'user'): [(0, 1), (1, 2)],
-        ('user', 'plays', 'game'): [(0, 0), (1, 0)]
+        ('user', 'follows', 'user'): ([0, 1], [1, 2]),
+        ('user', 'plays', 'game'): ([0, 1], [0, 0])
     }, idtype=idtype, device=F.ctx())
     g2.nodes['user'].data['h1'] = F.tensor([[0.], [1.], [2.]])
     g2.nodes['user'].data['h2'] = F.tensor([[3.], [4.], [5.]])
@@ -243,8 +243,8 @@ def test_features(idtype):
 def test_empty_relation(idtype):
     """Test the features of batched DGLHeteroGraphs"""
     g1 = dgl.heterograph({
-        ('user', 'follows', 'user'): [(0, 1), (1, 2)],
-        ('user', 'plays', 'game'): []
+        ('user', 'follows', 'user'): ([0, 1], [1, 2]),
+        ('user', 'plays', 'game'): ([], [])
     }, idtype=idtype, device=F.ctx())
     g1.nodes['user'].data['h1'] = F.tensor([[0.], [1.], [2.]])
     g1.nodes['user'].data['h2'] = F.tensor([[3.], [4.], [5.]])
@@ -252,8 +252,8 @@ def test_empty_relation(idtype):
     g1.edges['follows'].data['h2'] = F.tensor([[2.], [3.]])
 
     g2 = dgl.heterograph({
-        ('user', 'follows', 'user'): [(0, 1), (1, 2)],
-        ('user', 'plays', 'game'): [(0, 0), (1, 0)]
+        ('user', 'follows', 'user'): ([0, 1], [1, 2]),
+        ('user', 'plays', 'game'): ([0, 1], [0, 0])
     }, idtype=idtype, device=F.ctx())
     g2.nodes['user'].data['h1'] = F.tensor([[0.], [1.], [2.]])
     g2.nodes['user'].data['h2'] = F.tensor([[3.], [4.], [5.]])
@@ -298,8 +298,8 @@ def test_empty_relation(idtype):
         edge_attrs={('user', 'follows', 'user'): ['h1']})
 
     # Test graphs without edges
-    g1 = dgl.bipartite([], 'u', 'r', 'v', num_nodes=(0, 4))
-    g2 = dgl.bipartite([], 'u', 'r', 'v', num_nodes=(1, 5))
+    g1 = dgl.heterograph({('u', 'r', 'v'): ([], [])}, {'u': 0, 'v': 4})
+    g2 = dgl.heterograph({('u', 'r', 'v'): ([], [])}, {'u': 1, 'v': 5})
     dgl.batch([g1, g2])
 
 @parametrize_dtype

tests/compute/test_graph.py

@@ -223,8 +223,6 @@ def test_query():
     _test(gen_from_data(elist_input(), False, False))
     _test(gen_from_data(elist_input(), True, False))
     _test(gen_from_data(elist_input(), True, True))
-    _test(gen_from_data(nx_input(), False, False))
-    _test(gen_from_data(nx_input(), True, False))
     _test(gen_from_data(scipy_coo_input(), False, False))
     _test(gen_from_data(scipy_coo_input(), True, False))