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 GGI(Dataset):
"""A pair of networks synthesized from the human gene-gene interaction (PPI) network from
`IsoBase <https://cb.csail.mit.edu/mna/isobase/>`_. Nodes represent human genes and edges represent gene-gene
interactions. The two networks are noisy permutations of the original network generated by randomly inserting 10%
edges (GGI1) and deleting 15% edges (GGI2) from the original network, respectively.
There are in total 10,403 common nodes across two networks.
.. list-table::
:widths: 10 10 10 10 10
:header-rows: 1
* - Graph
- #nodes
- #edges
- #node attrs
- #edge attrs
* - GGI1
- 10,403
- 130,868
- 0
- 0
* - GGI2
- 10,403
- 100,603
- 0
- 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='1BbnuLpo1gyiH1I7m1RNrRzjCwHv0C6aq',
save_filename='GGI.pt',
root=root)
if not self._check_integrity(root):
raise RuntimeError('PPI dataset not found or corrupted. You can use download=True to download it')
super(GGI, self).__init__(root=root, name='GGI', train_ratio=train_ratio, dtype=dtype, seed=seed)
def _check_integrity(self, root):
return os.path.exists(os.path.join(root, 'GGI.pt'))