from typing import Union, Optional
from pathlib import Path
import os
import torch
from PlanetAlign.data import Dataset
from .utils import download_file_from_google_drive
[docs]
class Cora(Dataset):
"""A pair of networks synthesized from the citation network Cora. Each network
Nodes represent publications and an edge exists between two publications if they have a citation relationship.
The two networks are noisy permutations of the original network generated by randomly inserting 10% edges (Cora1)
and deleting 15% edges (Cora2) from the original network, respectively. There are in total 2,708 common nodes across
two networks.
.. list-table::
:widths: 10 10 10 10 10
:header-rows: 1
* - Graph
- #nodes
- #edges
- #node attrs
- #edge attrs
* - Cora1
- 2,708
- 6,334
- 1,433
- 0
* - Cora2
- 2,708
- 4,542
- 1,433
- 0
"""
def __init__(self,
root: Union[str, Path],
download: Optional[bool] = False,
train_ratio: Optional[float] = 0.2,
dtype: torch.dtype = torch.float32,
seed: Optional[int] = 0):
if download:
download_file_from_google_drive(
remote_file_id='1cG9DoOxUcpCzWgWsxyTlp4XvHGNkY-oF',
save_filename='cora.pt',
root=root)
if not self._check_integrity(root):
raise RuntimeError('Cora dataset not found or corrupted. You can use download=True to download it')
super(Cora, self).__init__(root=root, name='cora', train_ratio=train_ratio, dtype=dtype, seed=seed)
def _check_integrity(self, root):
return os.path.exists(os.path.join(root, 'cora.pt'))