| 49 | |
| 50 | |
| 51 | class kHopSubgraphDataset(Dataset): |
| 52 | def __init__(self, data, num_hops=1, max_nodes=100, dataset_name="Cora", transform=None, pre_transform=None): |
| 53 | super(kHopSubgraphDataset, self).__init__(None, transform, pre_transform) |
| 54 | self.data = data |
| 55 | self.num_hops = num_hops |
| 56 | self.unique_classes = data.y.unique() |
| 57 | self.k_over_2 = math.ceil(len(self.unique_classes)/2) |
| 58 | if dataset_name == "Citeseer": |
| 59 | self.k_over_2 = 2 |
| 60 | elif dataset_name == "Arxiv": |
| 61 | self.k_over_2 = 5 |
| 62 | # elif dataset_name == "tech": |
| 63 | # self.k_over_2 = 2 |
| 64 | # elif dataset_name == "home": |
| 65 | # self.k_over_2 = 3 |
| 66 | |
| 67 | |
| 68 | self.max_nodes = max_nodes |
| 69 | self.dataset_name = dataset_name |
| 70 | self.subgraphs = self._create_subgraphs() |
| 71 | |
| 72 | def _create_subgraphs(self): |
| 73 | subgraphs = [] |
| 74 | for idx in tqdm(range(self.data.num_nodes)): |
| 75 | |
| 76 | subgraph_node_idx, subgraph_edge_index, mapping, edge_mask = k_hop_subgraph( |
| 77 | node_idx=idx, |
| 78 | num_hops=self.num_hops, |
| 79 | edge_index=self.data.edge_index, |
| 80 | relabel_nodes=True, |
| 81 | num_nodes=self.data.num_nodes |
| 82 | ) |
| 83 | |
| 84 | unique_classes_in_subgraph = np.unique(self.data.y[subgraph_node_idx].cpu().numpy()) |
| 85 | if len(unique_classes_in_subgraph) >= self.k_over_2 and len(subgraph_node_idx) <= self.max_nodes: |
| 86 | sub_data = Data(edge_index=subgraph_edge_index) |
| 87 | sub_data.y = self.data.y[subgraph_node_idx] |
| 88 | sub_data.raw_text = [self.data.raw_texts[i] for i in subgraph_node_idx.tolist()] |
| 89 | sub_data.label_text = self.data.label_text |
| 90 | sub_data.adjacency_matrix = to_dense_adj(subgraph_edge_index, max_num_nodes=mapping.size(0))[0] |
| 91 | sub_data.dataset_name = self.dataset_name |
| 92 | subgraphs.append(sub_data) |
| 93 | return subgraphs |
| 94 | |
| 95 | def len(self): |
| 96 | return len(self.subgraphs) |
| 97 | |
| 98 | def get(self, idx): |
| 99 | return self.subgraphs[idx] |
| 100 | |
| 101 | def __getitem__(self, idx): |
| 102 | return self.get(idx) |
| 103 | |
| 104 | # class kHopSubgraphDataset(Dataset): |
| 105 | # def __init__(self, data, num_hops=1, max_nodes=100, dataset_name="Cora", min_classes=None, transform=None, pre_transform=None): |
no outgoing calls
no test coverage detected