def test_subframes(parent_idx_device, child_device):
parent_device, idx_device = parent_idx_device
g = dgl.graph((F.tensor([1, 2, 3],
dtype=F.int64), F.tensor([2, 3, 4],
dtype=F.int64)))
print(g.device)
g.ndata['x'] = F.randn((5, 4))
g.edata['a'] = F.randn((3, 6))
idx = F.tensor([1, 2], dtype=F.int64)
if parent_device == 'cuda':
g = g.to(F.cuda())
elif parent_device == 'uva':
g = g.to(F.cpu())
g.create_formats_()
g.pin_memory_()
elif parent_device == 'cpu':
g = g.to(F.cpu())
idx = F.copy_to(idx, idx_device)
sg = g.sample_neighbors(idx, 2).to(child_device)
assert sg.device == sg.ndata['x'].device
assert sg.device == sg.edata['a'].device
assert sg.device == child_device
if parent_device != 'uva':
sg = g.to(child_device).sample_neighbors(F.copy_to(idx, child_device),
2)
assert sg.device == sg.ndata['x'].device
assert sg.device == sg.edata['a'].device
assert sg.device == child_device
if parent_device == 'uva':
g.unpin_memory_()
def test_subframes(parent_idx_device, child_device):
parent_device, idx_device = parent_idx_device
g = dgl.graph((F.tensor([1,2,3], dtype=F.int64), F.tensor([2,3,4], dtype=F.int64)))
print(g.device)
g.ndata['x'] = F.randn((5, 4))
g.edata['a'] = F.randn((3, 6))
idx = F.tensor([1, 2], dtype=F.int64)
if parent_device == 'cuda':
g = g.to(F.cuda())
elif parent_device == 'uva':
if F.backend_name != 'pytorch':
pytest.skip("UVA only supported for PyTorch")
g = g.to(F.cpu())
g.create_formats_()
g.pin_memory_()
elif parent_device == 'cpu':
g = g.to(F.cpu())
idx = F.copy_to(idx, idx_device)
sg = g.sample_neighbors(idx, 2).to(child_device)
assert sg.device == F.context(sg.ndata['x'])
assert sg.device == F.context(sg.edata['a'])
assert sg.device == child_device
if parent_device != 'uva':
sg = g.to(child_device).sample_neighbors(F.copy_to(idx, child_device), 2)
assert sg.device == F.context(sg.ndata['x'])
assert sg.device == F.context(sg.edata['a'])
assert sg.device == child_device
if parent_device == 'uva':
g.unpin_memory_()
def test_segment_reduce(reducer):
ctx = F.ctx()
value = F.tensor(np.random.rand(10, 5))
v1 = F.attach_grad(F.clone(value))
v2 = F.attach_grad(F.clone(value))
seglen = F.tensor([2, 3, 0, 4, 1, 0, 0])
u = F.copy_to(F.arange(0, F.shape(value)[0], F.int32), ctx)
v = F.repeat(F.copy_to(F.arange(0, len(seglen), F.int32), ctx),
seglen,
dim=0)
num_nodes = {'_U': len(u), '_V': len(seglen)}
g = dgl.convert.heterograph({('_U', '_E', '_V'): (u, v)},
num_nodes_dict=num_nodes)
with F.record_grad():
rst1 = gspmm(g, 'copy_lhs', reducer, v1, None)
if reducer in ['max', 'min']:
rst1 = F.replace_inf_with_zero(rst1)
F.backward(F.reduce_sum(rst1))
grad1 = F.grad(v1)
with F.record_grad():
rst2 = segment_reduce(seglen, v2, reducer=reducer)
F.backward(F.reduce_sum(rst2))
assert F.allclose(rst1, rst2)
print('forward passed')
grad2 = F.grad(v2)
assert F.allclose(grad1, grad2)
print('backward passed')
def test_filter():
g = DGLGraph()
g.add_nodes(4)
g.add_edges([0, 1, 2, 3], [1, 2, 3, 0])
n_repr = np.zeros((4, 5))
e_repr = np.zeros((4, 5))
n_repr[[1, 3]] = 1
e_repr[[1, 3]] = 1
n_repr = F.copy_to(F.zerocopy_from_numpy(n_repr), F.ctx())
e_repr = F.copy_to(F.zerocopy_from_numpy(e_repr), F.ctx())
g.ndata['a'] = n_repr
g.edata['a'] = e_repr
def predicate(r):
return F.max(r.data['a'], 1) > 0
# full node filter
n_idx = g.filter_nodes(predicate)
assert set(F.zerocopy_to_numpy(n_idx)) == {1, 3}
# partial node filter
n_idx = g.filter_nodes(predicate, [0, 1])
assert set(F.zerocopy_to_numpy(n_idx)) == {1}
# full edge filter
e_idx = g.filter_edges(predicate)
assert set(F.zerocopy_to_numpy(e_idx)) == {1, 3}
# partial edge filter
e_idx = g.filter_edges(predicate, [0, 1])
assert set(F.zerocopy_to_numpy(e_idx)) == {1}
def _random_simple_graph(idtype, dtype, ctx, M, N, max_nnz, srctype, dsttype,
etype):
src = np.random.randint(0, M, (max_nnz, ))
dst = np.random.randint(0, N, (max_nnz, ))
val = np.random.randn(max_nnz)
a = ssp.csr_matrix((val, (src, dst)), shape=(M, N))
a.sum_duplicates()
a = a.tocoo()
# shuffle edges
perm = np.random.permutation(a.nnz)
row = a.row[perm]
col = a.col[perm]
val = a.data[perm]
a = ssp.csr_matrix((val, (row, col)), shape=(M, N))
A = dgl.heterograph(
{
(srctype, etype, dsttype):
(F.copy_to(F.tensor(row, dtype=idtype),
ctx), F.copy_to(F.tensor(col, dtype=idtype), ctx))
},
num_nodes_dict={
srctype: a.shape[0],
dsttype: a.shape[1]
})
A.edata['w'] = F.copy_to(F.tensor(val, dtype=dtype), ctx)
return a, A
def graph1():
g = dgl.graph(([0, 0, 0, 1, 1, 2, 2, 2, 3, 3, 3, 4, 6, 6, 7, 8, 9
], [4, 5, 1, 2, 4, 7, 9, 8, 6, 4, 1, 0, 1, 0, 2, 3, 5]),
device=F.cpu())
g.ndata['h'] = F.copy_to(F.randn((g.number_of_nodes(), 2)), F.cpu())
g.edata['w'] = F.copy_to(F.randn((g.number_of_edges(), 3)), F.cpu())
return g
def heterograph0():
g = dgl.heterograph({
('user', 'plays', 'game'): ([0, 1, 1, 2], [0, 0, 1, 1]),
('developer', 'develops', 'game'): ([0, 1], [0, 1])}, device=F.cpu())
g.nodes['user'].data['h'] = F.copy_to(F.randn((g.number_of_nodes('user'), 3)), F.cpu())
g.nodes['game'].data['h'] = F.copy_to(F.randn((g.number_of_nodes('game'), 2)), F.cpu())
g.nodes['developer'].data['h'] = F.copy_to(F.randn((g.number_of_nodes('developer'), 3)), F.cpu())
g.edges['plays'].data['h'] = F.copy_to(F.randn((g.number_of_edges('plays'), 1)), F.cpu())
g.edges['develops'].data['h'] = F.copy_to(F.randn((g.number_of_edges('develops'), 5)), F.cpu())
return g
def tensor_topo_traverse():
n = g.number_of_nodes()
mask = F.copy_to(F.ones((n, 1)), F.cpu())
degree = F.spmm(adjmat, mask)
while F.reduce_sum(mask) != 0.:
v = F.astype((degree == 0.), F.float32)
v = v * mask
mask = mask - v
frontier = F.copy_to(F.nonzero_1d(F.squeeze(v, 1)), F.cpu())
yield frontier
degree -= F.spmm(adjmat, v)
def test_to_device(index_dtype):
g1 = dgl.heterograph({('user', 'plays', 'game'): [(0, 0), (1, 1)]},
index_dtype=index_dtype)
g1.nodes['user'].data['h1'] = F.copy_to(F.tensor([[0.], [1.]]), F.cpu())
g1.nodes['user'].data['h2'] = F.copy_to(F.tensor([[3.], [4.]]), F.cpu())
g1.edges['plays'].data['h1'] = F.copy_to(F.tensor([[2.], [3.]]), F.cpu())
g2 = dgl.heterograph({('user', 'plays', 'game'): [(0, 0), (1, 0)]},
index_dtype=index_dtype)
g2.nodes['user'].data['h1'] = F.copy_to(F.tensor([[1.], [2.]]), F.cpu())
g2.nodes['user'].data['h2'] = F.copy_to(F.tensor([[4.], [5.]]), F.cpu())
g2.edges['plays'].data['h1'] = F.copy_to(F.tensor([[0.], [1.]]), F.cpu())
bg = dgl.batch_hetero([g1, g2])
if F.is_cuda_available():
bg1 = bg.to(F.cuda())
assert bg1 is not None
assert bg.batch_size == bg1.batch_size
assert bg.batch_num_nodes('user') == bg1.batch_num_nodes('user')
assert bg.batch_num_edges('plays') == bg1.batch_num_edges('plays')
# set feature
g1 = dgl.heterograph({('user', 'plays', 'game'): [(0, 0), (1, 1)]},
index_dtype=index_dtype)
g2 = dgl.heterograph({('user', 'plays', 'game'): [(0, 0), (1, 0)]},
index_dtype=index_dtype)
bg = dgl.batch_hetero([g1, g2])
if F.is_cuda_available():
bg1 = bg.to(F.cuda())
bg1.nodes['user'].data['test'] = F.copy_to(F.tensor([0, 1, 2, 3]),
F.cuda())
bg1.edata['test'] = F.copy_to(F.tensor([0, 1, 2, 3]), F.cuda())
def test_get_node_partition_from_book(idtype):
node_map = {"type_n": F.tensor([[0, 3], [4, 5], [6, 10]], dtype=idtype)}
edge_map = {"type_e": F.tensor([[0, 9], [10, 15], [16, 25]], dtype=idtype)}
book = gpb.RangePartitionBook(0, 3, node_map, edge_map, {"type_n": 0},
{"type_e": 0})
partition = gpb.get_node_partition_from_book(book, F.ctx())
assert partition.num_parts() == 3
assert partition.array_size() == 11
test_ids = F.copy_to(F.tensor([0, 2, 6, 7, 10], dtype=idtype), F.ctx())
act_ids = partition.map_to_local(test_ids)
exp_ids = F.copy_to(F.tensor([0, 2, 0, 1, 4], dtype=idtype), F.ctx())
assert F.array_equal(act_ids, exp_ids)
test_ids = F.copy_to(F.tensor([0, 2], dtype=idtype), F.ctx())
act_ids = partition.map_to_global(test_ids, 0)
exp_ids = F.copy_to(F.tensor([0, 2], dtype=idtype), F.ctx())
assert F.array_equal(act_ids, exp_ids)
test_ids = F.copy_to(F.tensor([0, 1], dtype=idtype), F.ctx())
act_ids = partition.map_to_global(test_ids, 1)
exp_ids = F.copy_to(F.tensor([4, 5], dtype=idtype), F.ctx())
assert F.array_equal(act_ids, exp_ids)
test_ids = F.copy_to(F.tensor([0, 1, 4], dtype=idtype), F.ctx())
act_ids = partition.map_to_global(test_ids, 2)
exp_ids = F.copy_to(F.tensor([6, 7, 10], dtype=idtype), F.ctx())
assert F.array_equal(act_ids, exp_ids)
def test_basic():
num_layers = 2
g = generate_rand_graph(100, connect_more=True)
nf = create_full_nodeflow(g, num_layers)
assert nf.number_of_nodes() == g.number_of_nodes() * (num_layers + 1)
assert nf.number_of_edges() == g.number_of_edges() * num_layers
assert nf.num_layers == num_layers + 1
assert nf.layer_size(0) == g.number_of_nodes()
assert nf.layer_size(1) == g.number_of_nodes()
check_basic(g, nf)
parent_nids = F.copy_to(F.arange(0, g.number_of_nodes()), F.cpu())
nids = nf.map_from_parent_nid(0, parent_nids, remap_local=True)
assert_array_equal(F.asnumpy(nids), F.asnumpy(parent_nids))
# should also work for negative layer ids
for l in range(-1, -num_layers, -1):
nids1 = nf.map_from_parent_nid(l, parent_nids, remap_local=True)
nids2 = nf.map_from_parent_nid(l + num_layers,
parent_nids,
remap_local=True)
assert_array_equal(F.asnumpy(nids1), F.asnumpy(nids2))
g = generate_rand_graph(100)
nf = create_mini_batch(g, num_layers)
assert nf.num_layers == num_layers + 1
check_basic(g, nf)
g = generate_rand_graph(100, add_self_loop=True)
nf = create_mini_batch(g, num_layers, add_self_loop=True)
assert nf.num_layers == num_layers + 1
check_basic(g, nf)
def test_subgraph_message_passing():
# Unit test for PR #2055
g = dgl.graph(([0, 1, 2], [2, 3, 4])).to(F.cpu())
g.ndata['x'] = F.copy_to(F.randn((5, 6)), F.cpu())
sg = g.subgraph([1, 2, 3]).to(F.ctx())
sg.update_all(lambda edges: {'x': edges.src['x']},
lambda nodes: {'y': F.sum(nodes.mailbox['x'], 1)})
def check_positive_edge_sampler():
g = generate_rand_graph(1000)
num_edges = g.number_of_edges()
edge_weight = F.copy_to(
F.tensor(np.full((num_edges, ), 1, dtype=np.float32)), F.cpu())
edge_weight[num_edges - 1] = num_edges**2
EdgeSampler = getattr(dgl.contrib.sampling, 'EdgeSampler')
# Correctness check
# Test the homogeneous graph.
batch_size = 128
edge_sampled = np.full((num_edges, ), 0, dtype=np.int32)
for pos_edges in EdgeSampler(g,
batch_size,
reset=False,
edge_weight=edge_weight):
_, _, pos_leid = pos_edges.all_edges(form='all', order='eid')
np.add.at(edge_sampled, F.asnumpy(pos_edges.parent_eid[pos_leid]), 1)
truth = np.full((num_edges, ), 1, dtype=np.int32)
edge_sampled = edge_sampled[:num_edges]
assert np.array_equal(truth, edge_sampled)
edge_sampled = np.full((num_edges, ), 0, dtype=np.int32)
for pos_edges in EdgeSampler(g,
batch_size,
reset=False,
shuffle=True,
edge_weight=edge_weight):
_, _, pos_leid = pos_edges.all_edges(form='all', order='eid')
np.add.at(edge_sampled, F.asnumpy(pos_edges.parent_eid[pos_leid]), 1)
truth = np.full((num_edges, ), 1, dtype=np.int32)
edge_sampled = edge_sampled[:num_edges]
assert np.array_equal(truth, edge_sampled)
def test_multi_send():
g = generate_graph()
def _fmsg(edges):
assert edges.src['h'].shape == (5, D)
return {'m': edges.src['h']}
g.register_message_func(_fmsg)
# many-many send
u = F.tensor([0, 0, 0, 0, 0])
v = F.tensor([1, 2, 3, 4, 5])
g.send((u, v))
# duplicate send
u = F.tensor([0])
v = F.tensor([1, 2, 3, 4, 5])
g.send((u, v))
# send more
u = F.tensor([1, 2, 3, 4, 5])
v = F.tensor([9])
g.send((u, v))
# check if message indicator is as expected
expected = F.copy_to(F.zeros((g.number_of_edges(), ), dtype=F.int64),
F.cpu())
eid = g.edge_ids([0, 0, 0, 0, 0, 1, 2, 3, 4, 5],
[1, 2, 3, 4, 5, 9, 9, 9, 9, 9])
expected = F.asnumpy(expected)
eid = F.asnumpy(eid)
expected[eid] = 1
assert np.array_equal(g._get_msg_index().tonumpy(), expected)
def test_node_dataloader(sampler_name, pin_graph):
g1 = dgl.graph(([0, 0, 0, 1, 1], [1, 2, 3, 3, 4]))
if F.ctx() != F.cpu() and pin_graph:
g1.create_formats_()
g1.pin_memory_()
g1.ndata['feat'] = F.copy_to(F.randn((5, 8)), F.cpu())
g1.ndata['label'] = F.copy_to(F.randn((g1.num_nodes(),)), F.cpu())
for num_workers in [0, 1, 2]:
sampler = {
'full': dgl.dataloading.MultiLayerFullNeighborSampler(2),
'neighbor': dgl.dataloading.MultiLayerNeighborSampler([3, 3]),
'neighbor2': dgl.dataloading.MultiLayerNeighborSampler([3, 3])}[sampler_name]
dataloader = dgl.dataloading.NodeDataLoader(
g1, g1.nodes(), sampler, device=F.ctx(),
batch_size=g1.num_nodes(),
num_workers=num_workers)
for input_nodes, output_nodes, blocks in dataloader:
_check_device(input_nodes)
_check_device(output_nodes)
_check_device(blocks)
g2 = dgl.heterograph({
('user', 'follow', 'user'): ([0, 0, 0, 1, 1, 1, 2], [1, 2, 3, 0, 2, 3, 0]),
('user', 'followed-by', 'user'): ([1, 2, 3, 0, 2, 3, 0], [0, 0, 0, 1, 1, 1, 2]),
('user', 'play', 'game'): ([0, 1, 1, 3, 5], [0, 1, 2, 0, 2]),
('game', 'played-by', 'user'): ([0, 1, 2, 0, 2], [0, 1, 1, 3, 5])
})
for ntype in g2.ntypes:
g2.nodes[ntype].data['feat'] = F.copy_to(F.randn((g2.num_nodes(ntype), 8)), F.cpu())
batch_size = max(g2.num_nodes(nty) for nty in g2.ntypes)
sampler = {
'full': dgl.dataloading.MultiLayerFullNeighborSampler(2),
'neighbor': dgl.dataloading.MultiLayerNeighborSampler([{etype: 3 for etype in g2.etypes}] * 2),
'neighbor2': dgl.dataloading.MultiLayerNeighborSampler([3, 3])}[sampler_name]
dataloader = dgl.dataloading.NodeDataLoader(
g2, {nty: g2.nodes(nty) for nty in g2.ntypes},
sampler, device=F.ctx(), batch_size=batch_size)
assert isinstance(iter(dataloader), Iterator)
for input_nodes, output_nodes, blocks in dataloader:
_check_device(input_nodes)
_check_device(output_nodes)
_check_device(blocks)
if g1.is_pinned():
g1.unpin_memory_()
def test_dynamic_addition():
N = 3
D = 1
g = DGLGraph()
def _init(shape, dtype, ctx, ids):
return F.copy_to(F.astype(F.randn(shape), dtype), ctx)
g.set_n_initializer(_init)
g.set_e_initializer(_init)
def _message(edges):
return {
'm':
edges.src['h1'] + edges.dst['h2'] + edges.data['h1'] +
edges.data['h2']
}
def _reduce(nodes):
return {'h': F.sum(nodes.mailbox['m'], 1)}
def _apply(nodes):
return {'h': nodes.data['h']}
g.register_message_func(_message)
g.register_reduce_func(_reduce)
g.register_apply_node_func(_apply)
g.set_n_initializer(dgl.init.zero_initializer)
g.set_e_initializer(dgl.init.zero_initializer)
# add nodes and edges
g.add_nodes(N)
g.ndata.update({'h1': F.randn((N, D)), 'h2': F.randn((N, D))})
g.add_nodes(3)
g.add_edge(0, 1)
g.add_edge(1, 0)
g.edata.update({'h1': F.randn((2, D)), 'h2': F.randn((2, D))})
g.send()
expected = F.copy_to(F.ones((g.number_of_edges(), ), dtype=F.int64),
F.cpu())
assert F.array_equal(g._get_msg_index().tousertensor(), expected)
# add more edges
g.add_edges([0, 2], [2, 0], {'h1': F.randn((2, D))})
g.send(([0, 2], [2, 0]))
g.recv(0)
g.add_edge(1, 2)
g.edges[4].data['h1'] = F.randn((1, D))
g.send((1, 2))
g.recv([1, 2])
h = g.ndata.pop('h')
# a complete round of send and recv
g.send()
g.recv()
assert F.allclose(h, g.ndata['h'])
def _test_callback_array_thread(dtype):
def cb(x):
return F.to_dgl_nd(F.from_dgl_nd(x) + 1)
arg = F.copy_to(F.tensor([1, 2, 3], dtype=dtype), F.ctx())
ret = F.from_dgl_nd(
dgl._api_internal._TestPythonCallbackThread(cb, F.to_dgl_nd(arg)))
assert np.allclose(F.asnumpy(ret), F.asnumpy(arg) + 1)
def _test_sampler(g, sampler, ntype):
seeds = F.copy_to(F.tensor([0, 2], dtype=F.int64), F.ctx())
neighbor_g = sampler(seeds)
assert neighbor_g.ntypes == [ntype]
u, v = neighbor_g.all_edges(form='uv', order='eid')
uv = list(zip(F.asnumpy(u).tolist(), F.asnumpy(v).tolist()))
assert (1, 0) in uv or (0, 0) in uv
assert (2, 2) in uv or (3, 2) in uv
def test_async_transferer_to_other():
cpu_ones = F.ones([100,75,25], dtype=F.int32, ctx=F.cpu())
tran = AsyncTransferer(F.ctx())
t = tran.async_copy(cpu_ones, F.ctx())
other_ones = t.wait()
assert F.context(other_ones) == F.ctx()
assert F.array_equal(F.copy_to(other_ones, ctx=F.cpu()), cpu_ones)
def test_column_subcolumn():
data = F.copy_to(
F.tensor([[1., 1., 1., 1.], [0., 2., 9., 0.], [3., 2., 1., 0.],
[1., 1., 1., 1.], [0., 2., 4., 0.]]), F.ctx())
original = Column(data)
# subcolumn from cpu context
i1 = F.tensor([0, 2, 1, 3], dtype=F.int64)
l1 = original.subcolumn(i1)
assert len(l1) == i1.shape[0]
assert F.array_equal(l1.data, F.gather_row(data, i1))
# next subcolumn from target context
i2 = F.copy_to(F.tensor([0, 2], dtype=F.int64), F.ctx())
l2 = l1.subcolumn(i2)
assert len(l2) == i2.shape[0]
i1i2 = F.copy_to(F.gather_row(i1, F.copy_to(i2, F.context(i1))), F.ctx())
assert F.array_equal(l2.data, F.gather_row(data, i1i2))
# next subcolumn also from target context
i3 = F.copy_to(F.tensor([1], dtype=F.int64), F.ctx())
l3 = l2.subcolumn(i3)
assert len(l3) == i3.shape[0]
i1i2i3 = F.copy_to(F.gather_row(i1i2, F.copy_to(i3, F.context(i1i2))),
F.ctx())
assert F.array_equal(l3.data, F.gather_row(data, i1i2i3))
def test_self_loop():
n = 100
num_hops = 2
g = generate_rand_graph(n, complete=True)
nf = create_mini_batch(g, num_hops, add_self_loop=True)
for i in range(1, nf.num_layers):
in_deg = nf.layer_in_degree(i)
deg = F.copy_to(F.ones(in_deg.shape, dtype=F.int64), F.cpu()) * n
assert_array_equal(F.asnumpy(in_deg), F.asnumpy(deg))
def test_edge_prediction_sampler(idtype):
g = create_test_graph(idtype)
sampler = NeighborSampler([10, 10])
sampler = as_edge_prediction_sampler(
sampler, negative_sampler=negative_sampler.Uniform(1))
seeds = F.copy_to(F.arange(0, 2, dtype=idtype), ctx=F.ctx())
# just a smoke test to make sure we don't fail internal assertions
result = sampler.sample(g, {'follows': seeds})
def test_uva_subgraph(idtype, device):
g = create_test_heterograph(idtype)
g = g.to(F.cpu())
g.create_formats_()
g.pin_memory_()
indices = {'user': F.copy_to(F.tensor([0], idtype), device)}
edge_indices = {'follows': F.copy_to(F.tensor([0], idtype), device)}
assert g.subgraph(indices).device == device
assert g.edge_subgraph(edge_indices).device == device
assert g.in_subgraph(indices).device == device
assert g.out_subgraph(indices).device == device
if dgl.backend.backend_name != 'tensorflow':
# (BarclayII) Most of Tensorflow functions somehow do not preserve device: a CPU tensor
# becomes a GPU tensor after operations such as concat(), unique() or even sin().
# Not sure what should be the best fix.
assert g.khop_in_subgraph(indices, 1)[0].device == device
assert g.khop_out_subgraph(indices, 1)[0].device == device
assert g.sample_neighbors(indices, 1).device == device
g.unpin_memory_()
def test_sage_conv2(idtype):
# TODO: add test for blocks
# Test the case for graphs without edges
g = dgl.heterograph({('_U', '_E', '_V'): ([], [])}, {'_U': 5, '_V': 3})
g = g.astype(idtype).to(F.ctx())
ctx = F.ctx()
sage = nn.SAGEConv((3, 3), 2, 'gcn')
feat = (F.randn((5, 3)), F.randn((3, 3)))
sage = sage.to(ctx)
h = sage(g, (F.copy_to(feat[0], F.ctx()), F.copy_to(feat[1], F.ctx())))
assert h.shape[-1] == 2
assert h.shape[0] == 3
for aggre_type in ['mean', 'pool', 'lstm']:
sage = nn.SAGEConv((3, 1), 2, aggre_type)
feat = (F.randn((5, 3)), F.randn((3, 1)))
sage = sage.to(ctx)
h = sage(g, feat)
assert h.shape[-1] == 2
assert h.shape[0] == 3
def test_sage_conv2(idtype):
# TODO: add test for blocks
# Test the case for graphs without edges
g = dgl.bipartite([], num_nodes=(5, 3))
g = g.astype(idtype).to(F.ctx())
ctx = F.ctx()
sage = nn.SAGEConv((3, 3), 2, 'gcn')
feat = (F.randn((5, 3)), F.randn((3, 3)))
sage = sage.to(ctx)
h = sage(g, (F.copy_to(feat[0], F.ctx()), F.copy_to(feat[1], F.ctx())))
assert h.shape[-1] == 2
assert h.shape[0] == 3
for aggre_type in ['mean', 'pool', 'lstm']:
sage = nn.SAGEConv((3, 1), 2, aggre_type)
feat = (F.randn((5, 3)), F.randn((3, 1)))
sage = sage.to(ctx)
h = sage(g, feat)
assert h.shape[-1] == 2
assert h.shape[0] == 3
def test_multi_recv():
# basic recv test
g = generate_graph()
h = g.ndata['h']
g.register_message_func(message_func)
g.register_reduce_func(reduce_func)
g.register_apply_node_func(apply_node_func)
expected = F.copy_to(F.zeros((g.number_of_edges(), ), dtype=F.int64),
F.cpu())
# two separate round of send and recv
u = [4, 5, 6]
v = [9]
g.send((u, v))
eid = g.edge_ids(u, v)
expected[eid] = 1
assert F.array_equal(g._get_msg_index().tousertensor(), expected)
g.recv(v)
expected[eid] = 0
assert F.array_equal(g._get_msg_index().tousertensor(), expected)
u = [0]
v = [1, 2, 3]
g.send((u, v))
eid = g.edge_ids(u, v)
expected[eid] = 1
assert F.array_equal(g._get_msg_index().tousertensor(), expected)
g.recv(v)
expected[eid] = 0
assert F.array_equal(g._get_msg_index().tousertensor(), expected)
h1 = g.ndata['h']
# one send, two recv
g.ndata['h'] = h
u = F.tensor([0, 0, 0, 4, 5, 6])
v = F.tensor([1, 2, 3, 9, 9, 9])
g.send((u, v))
eid = g.edge_ids(u, v)
expected[eid] = 1
assert F.array_equal(g._get_msg_index().tousertensor(), expected)
u = [4, 5, 6]
v = [9]
g.recv(v)
eid = g.edge_ids(u, v)
expected[eid] = 0
assert F.array_equal(g._get_msg_index().tousertensor(), expected)
u = [0]
v = [1, 2, 3]
g.recv(v)
eid = g.edge_ids(u, v)
expected[eid] = 0
assert F.array_equal(g._get_msg_index().tousertensor(), expected)
h2 = g.ndata['h']
assert F.allclose(h1, h2)
def _random_simple_graph(idtype, dtype, ctx, M, N, max_nnz, srctype, dsttype,
etype):
src = np.random.randint(0, M, (max_nnz, ))
dst = np.random.randint(0, N, (max_nnz, ))
val = np.random.randn(max_nnz)
a = ssp.csr_matrix((val, (src, dst)), shape=(M, N))
a.sum_duplicates()
a = a.tocoo()
A = dgl.heterograph(
{
('A', 'AB', 'B'):
(F.copy_to(F.tensor(a.row, dtype=idtype),
ctx), F.copy_to(F.tensor(a.col, dtype=idtype), ctx))
},
num_nodes_dict={
'A': a.shape[0],
'B': a.shape[1]
})
A.edata['w'] = F.copy_to(F.tensor(a.data, dtype=dtype), ctx)
return a, A
def test_node_dataloader():
sampler = dgl.dataloading.MultiLayerFullNeighborSampler(2)
g1 = dgl.graph(([0, 0, 0, 1, 1], [1, 2, 3, 3, 4]))
g1.ndata['feat'] = F.copy_to(F.randn((5, 8)), F.cpu())
dataloader = dgl.dataloading.NodeDataLoader(g1,
g1.nodes(),
sampler,
device=F.ctx(),
batch_size=g1.num_nodes())
for input_nodes, output_nodes, blocks in dataloader:
_check_device(input_nodes)
_check_device(output_nodes)
_check_device(blocks)
g2 = dgl.heterograph({
('user', 'follow', 'user'): ([0, 0, 0, 1, 1, 1,
2], [1, 2, 3, 0, 2, 3, 0]),
('user', 'followed-by', 'user'): ([1, 2, 3, 0, 2, 3,
0], [0, 0, 0, 1, 1, 1, 2]),
('user', 'play', 'game'): ([0, 1, 1, 3, 5], [0, 1, 2, 0, 2]),
('game', 'played-by', 'user'): ([0, 1, 2, 0, 2], [0, 1, 1, 3, 5])
})
for ntype in g2.ntypes:
g2.nodes[ntype].data['feat'] = F.copy_to(
F.randn((g2.num_nodes(ntype), 8)), F.cpu())
batch_size = max(g2.num_nodes(nty) for nty in g2.ntypes)
dataloader = dgl.dataloading.NodeDataLoader(
g2, {nty: g2.nodes(nty)
for nty in g2.ntypes},
sampler,
device=F.ctx(),
batch_size=batch_size)
assert isinstance(iter(dataloader), Iterator)
for input_nodes, output_nodes, blocks in dataloader:
_check_device(input_nodes)
_check_device(output_nodes)
_check_device(blocks)
def test_serialize_deserialize_dtype():
data = F.copy_to(F.tensor([[1., 1., 1., 1.],
[0., 2., 9., 0.],
[3., 2., 1., 0.],
[1., 1., 1., 1.],
[0., 2., 4., 0.]]), F.ctx())
original = Column(data)
original = original.astype(F.int64)
serial = pickle.dumps(original)
new = pickle.loads(serial)
assert new.dtype == F.int64
def test_issue_2484(idtype):
import dgl.function as fn
g = dgl.graph(([0, 1, 2], [1, 2, 3]), idtype=idtype, device=F.ctx())
x = F.copy_to(F.randn((4, )), F.ctx())
g.ndata['x'] = x
g.pull([2, 1], fn.u_add_v('x', 'x', 'm'), fn.sum('m', 'x'))
y1 = g.ndata['x']
g.ndata['x'] = x
g.pull([1, 2], fn.u_add_v('x', 'x', 'm'), fn.sum('m', 'x'))
y2 = g.ndata['x']
assert F.allclose(y1, y2)
