gammagl.datasets.HGBDataset

class HGBDataset(root: str | None = None, name: str = 'acm', transform: Callable | None = None, pre_transform: Callable | None = None, force_reload: bool = False)[source]

A variety of heterogeneous graph benchmark datasets from the “Are We Really Making Much Progress? Revisiting, Benchmarking, and Refining Heterogeneous Graph Neural Networks” paper.

Note

Test labels are randomly given to prevent data leakage issues. If you want to obtain final test performance, you will need to submit your model predictions to the HGB leaderboard.

Parameters:

  • root (str, optional) – Root directory where the dataset should be saved.

  • name (str, optional) – The name of the dataset (one of "ACM", "DBLP", "Freebase", "IMDB")

  • transform (callable, optional) – A function/transform that takes in an gammmgl.transform object and returns a transformed version. The data object will be transformed before every access. (default: None)

  • pre_transform (callable, optional) – A function/transform that takes in an gammmgl.transform object and returns a transformed version. The data object will be transformed before being saved to disk. (default: None)

  • (bool (force_reload) – (default: False)

  • optional) (Whether to re-process the dataset.) – (default: False)

urls = {'acm_hgb': 'https://drive.google.com/uc?export=download&id=1xbJ4QE9pcDJOcALv7dYhHDCPITX2Iddz', 'dblp_hgb': 'https://drive.google.com/uc?export=download&id=1fLLoy559V7jJaQ_9mQEsC06VKd6Qd3SC', 'freebase_hgb': 'https://drive.google.com/uc?export=download&id=1vw-uqbroJZfFsWpriC1CWbtHCJMGdWJ7', 'imdb_hgb': 'https://drive.google.com/uc?export=download&id=18qXmmwKJBrEJxVQaYwKTL3Ny3fPqJeJ2'}
names = {'acm_hgb': 'ACM', 'dblp_hgb': 'DBLP', 'freebase_hgb': 'Freebase', 'imdb_hgb': 'IMDB'}
property raw_dir: str
property processed_dir: str
property raw_file_names: List[str]

The name of the files in the self.raw_dir folder that must be present in order to skip downloading.

property processed_file_names: str

The name of the files in the self.processed_dir folder that must be present in order to skip processing.

download()[source]

Downloads the dataset to the self.raw_dir folder.

process()[source]

Processes the dataset to the self.processed_dir folder.