-
Notifications
You must be signed in to change notification settings - Fork 55
/
dataset.py
183 lines (134 loc) · 7.03 KB
/
dataset.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
"""Dataset for auto-constraint model."""
import numpy as np
import torch
from sketchgraphs.data import sequence as datalib
from sketchgraphs.pipeline.graph_model.target import NODE_TYPES, EDGE_TYPES, EDGE_TYPES_PREDICTED, NODE_IDX_MAP, EDGE_IDX_MAP
from sketchgraphs.pipeline import graph_model as graph_utils
def _reindex_sparse_batch(sparse_batch, pack_batch_offsets):
return graph_utils.SparseFeatureBatch(
pack_batch_offsets[sparse_batch.index],
sparse_batch.value)
def collate(batch):
# Sort batch for packing
node_lengths = [len(x['node_features']) for x in batch]
sorted_indices = np.argsort(node_lengths)[::-1].copy()
batch = [batch[i] for i in sorted_indices]
graph = graph_utils.GraphInfo.merge(*[x['graph'] for x in batch])
edge_label = torch.tensor(
[x['target_edge_label'] for x in batch if x['target_edge_label'] != -1], dtype=torch.int64)
node_features = torch.nn.utils.rnn.pack_sequence([x['node_features'] for x in batch])
batch_offsets = graph_utils.offsets_from_counts(node_features.batch_sizes)
node_features_graph_index = torch.cat([
i + batch_offsets[:graph.node_counts[i]] for i in range(len(batch))
], dim=0)
sparse_node_features = {}
for k in batch[0]['sparse_node_features']:
sparse_node_features[k] = graph_utils.SparseFeatureBatch.merge(
[_reindex_sparse_batch(x['sparse_node_features'][k], batch_offsets) for x in batch], range(len(batch)))
last_graph_node_index = batch_offsets[graph.node_counts - 1] + torch.arange(len(graph.node_counts), dtype=torch.int64)
partner_index_index = []
partner_index = []
stop_partner_index_index = []
for i, x in enumerate(batch):
if x['partner_index'] == -1:
stop_partner_index_index.append(i)
continue
partner_index_index.append(i)
partner_index.append(x['partner_index'] + graph.node_offsets[i])
partner_index = graph_utils.SparseFeatureBatch(
torch.tensor(partner_index_index, dtype=torch.int64),
torch.tensor(partner_index, dtype=torch.int64)
)
stop_partner_index_index = torch.tensor(stop_partner_index_index, dtype=torch.int64)
return {
'graph': graph,
'edge_label': edge_label,
'partner_index': partner_index,
'stop_partner_index_index': stop_partner_index_index,
'node_features': node_features,
'node_features_graph_index': node_features_graph_index,
'sparse_node_features': sparse_node_features,
'last_graph_node_index': last_graph_node_index,
'sorted_indices': torch.as_tensor(sorted_indices)
}
def process_node_and_edge_ops(node_ops, edge_ops_in_graph, num_nodes_in_graph, node_feature_mappings):
all_node_labels = torch.tensor([NODE_IDX_MAP[op.label] for op in node_ops], dtype=torch.int64)
edge_labels = torch.tensor([EDGE_IDX_MAP[op.label] for op in edge_ops_in_graph], dtype=torch.int64)
if len(edge_ops_in_graph) > 0:
incidence = torch.tensor([(op.references[0], op.references[-1]) for op in edge_ops_in_graph],
dtype=torch.int64).T.contiguous()
incidence = torch.cat((incidence, torch.flip(incidence, [0])), dim=1)
else:
incidence = torch.empty([2, 0], dtype=torch.int64)
edge_features = edge_labels.repeat(2)
if node_feature_mappings is not None:
sparse_node_features = node_feature_mappings.all_sparse_features(node_ops)
else:
sparse_node_features = None
graph = graph_utils.GraphInfo.from_single_graph(incidence, None, edge_features, num_nodes_in_graph)
return {
'graph': graph,
'node_features': all_node_labels,
'sparse_node_features': sparse_node_features
}
class AutoconstraintDataset(torch.utils.data.Dataset):
def __init__(self, sequences, node_feature_mappings, seed=10):
self.sequences = sequences
self.node_feature_mappings = node_feature_mappings
self._rng = np.random.Generator(np.random.Philox(seed))
def __getitem__(self, idx):
idx = idx % len(self.sequences)
seq = self.sequences[idx]
if not isinstance(seq[0], datalib.NodeOp):
raise ValueError('First operation in sequence is not a NodeOp')
if seq[-1].label != 'Stop':
seq.append(datalib.NodeOp('stop', {}))
node_ops = [seq[0]]
edge_ops = []
num_predicted_edge_ops_per_node = []
num_non_predicted_edge_ops_per_node = []
predicted_edge_ops_for_current_node = 0
non_predicted_edge_ops_for_current_node = 0
for op in seq[1:]:
if isinstance(op, datalib.NodeOp):
num_predicted_edge_ops_per_node.append(predicted_edge_ops_for_current_node)
num_non_predicted_edge_ops_per_node.append(non_predicted_edge_ops_for_current_node)
predicted_edge_ops_for_current_node = 0
non_predicted_edge_ops_for_current_node = 0
node_ops.append(op)
else:
if op.label in EDGE_TYPES_PREDICTED:
predicted_edge_ops_for_current_node += 1
else:
non_predicted_edge_ops_for_current_node += 1
edge_ops.append(op)
node_ops = node_ops[:-1]
num_predicted_edge_ops_per_node = np.array(num_predicted_edge_ops_per_node, dtype=np.int64)
num_non_predicted_edge_ops_per_node = np.array(num_non_predicted_edge_ops_per_node, dtype=np.int64)
predicted_edge_ops_offsets = num_predicted_edge_ops_per_node.cumsum()
non_predicted_edge_ops_offsets = num_non_predicted_edge_ops_per_node.cumsum()
num_predicted_edge_ops = predicted_edge_ops_offsets[-1]
stop_target = self._rng.uniform() < len(node_ops) / (len(node_ops) + num_predicted_edge_ops)
if stop_target:
target_node_idx = self._rng.integers(len(node_ops))
num_nodes_in_graph = target_node_idx + 1
edge_ops_in_graph = edge_ops[:predicted_edge_ops_offsets[target_node_idx] + non_predicted_edge_ops_offsets[target_node_idx]]
target_edge_label = -1
partner_index = -1
else:
target_predicted_edge_idx = self._rng.integers(num_predicted_edge_ops)
target_node_idx = np.searchsorted(predicted_edge_ops_offsets, target_predicted_edge_idx, side='right')
num_nodes_in_graph = target_node_idx + 1
target_edge_idx = target_predicted_edge_idx + non_predicted_edge_ops_offsets[target_node_idx]
target_edge = edge_ops[target_edge_idx]
edge_ops_in_graph = edge_ops[:target_edge_idx]
target_edge_label = EDGE_IDX_MAP[target_edge.label]
partner_index = target_edge.references[-1]
assert target_edge_label < len(EDGE_TYPES_PREDICTED)
input_features = process_node_and_edge_ops(
node_ops, edge_ops_in_graph, num_nodes_in_graph, self.node_feature_mappings)
return {
**input_features,
'target_edge_label': target_edge_label,
'partner_index': partner_index,
}