Hyperlink 2012 graph dataset. Splits: pld_train, pld_test
| 486 | |
| 487 | |
| 488 | class Hyperlink2012(Dataset): |
| 489 | """ |
| 490 | Hyperlink 2012 graph dataset. |
| 491 | |
| 492 | Splits: |
| 493 | pld_train, pld_test |
| 494 | """ |
| 495 | def __init__(self): |
| 496 | super(Hyperlink2012, self).__init__( |
| 497 | "hyperlink2012", |
| 498 | urls={ |
| 499 | "pld_train": "http://data.dws.informatik.uni-mannheim.de/hyperlinkgraph/2012-08/pld-arc.gz", |
| 500 | "pld_valid": "http://data.dws.informatik.uni-mannheim.de/hyperlinkgraph/2012-08/pld-arc.gz", |
| 501 | "pld_test": "http://data.dws.informatik.uni-mannheim.de/hyperlinkgraph/2012-08/pld-arc.gz" |
| 502 | } |
| 503 | ) |
| 504 | |
| 505 | def pld_train_preprocess(self, graph_file, train_file): |
| 506 | valid_file = train_file[:train_file.rfind("pld_train.txt")] + "pld_valid.txt" |
| 507 | test_file = train_file[:train_file.rfind("pld_train.txt")] + "pld_test.txt" |
| 508 | self.link_prediction_split(graph_file, [train_file, valid_file, test_file], portions=[10000, 1, 1]) |
| 509 | |
| 510 | def pld_valid_preprocess(self, graph_file, valid_file): |
| 511 | train_file = valid_file[:valid_file.rfind("pld_valid.txt")] + "pld_train.txt" |
| 512 | test_file = valid_file[:valid_file.rfind("pld_valid.txt")] + "pld_test.txt" |
| 513 | self.link_prediction_split(graph_file, [train_file, valid_file, test_file], portions=[10000, 1, 1]) |
| 514 | |
| 515 | def pld_test_preprocess(self, graph_file, test_file): |
| 516 | train_file = test_file[:test_file.rfind("pld_test.txt")] + "pld_train.txt" |
| 517 | valid_file = test_file[:test_file.rfind("pld_test.txt")] + "pld_valid.txt" |
| 518 | self.link_prediction_split(graph_file, [train_file, valid_file, test_file], portions=[10000, 1, 1]) |
| 519 | |
| 520 | |
| 521 | class Friendster(Dataset): |