Source code for torch_geometric.datasets.elliptic

from typing import Any, Callable, List, Optional, Tuple

import torch

from torch_geometric.data import Data, InMemoryDataset
from torch_geometric.io import fs


[docs]class EllipticBitcoinDataset(InMemoryDataset): r"""The Elliptic Bitcoin dataset of Bitcoin transactions from the `"Anti-Money Laundering in Bitcoin: Experimenting with Graph Convolutional Networks for Financial Forensics" <https://arxiv.org/abs/1908.02591>`_ paper. :class:`EllipticBitcoinDataset` maps Bitcoin transactions to real entities belonging to licit categories (exchanges, wallet providers, miners, licit services, etc.) versus illicit ones (scams, malware, terrorist organizations, ransomware, Ponzi schemes, etc.) There exists 203,769 node transactions and 234,355 directed edge payments flows, with two percent of nodes (4,545) labelled as illicit, and twenty-one percent of nodes (42,019) labelled as licit. The remaining transactions are unknown. Args: root (str): Root directory where the dataset should be saved. transform (callable, optional): A function/transform that takes in an :obj:`torch_geometric.data.Data` object and returns a transformed version. The data object will be transformed before every access. (default: :obj:`None`) pre_transform (callable, optional): A function/transform that takes in an :obj:`torch_geometric.data.Data` object and returns a transformed version. The data object will be transformed before being saved to disk. (default: :obj:`None`) force_reload (bool, optional): Whether to re-process the dataset. (default: :obj:`False`) **STATS:** .. list-table:: :widths: 10 10 10 10 :header-rows: 1 * - #nodes - #edges - #features - #classes * - 203,769 - 234,355 - 165 - 2 """ url = 'https://data.pyg.org/datasets/elliptic' def __init__( self, root: str, transform: Optional[Callable] = None, pre_transform: Optional[Callable] = None, force_reload: bool = False, ) -> None: super().__init__(root, transform, pre_transform, force_reload=force_reload) self.load(self.processed_paths[0]) @property def raw_file_names(self) -> List[str]: return [ 'elliptic_txs_features.csv', 'elliptic_txs_edgelist.csv', 'elliptic_txs_classes.csv', ] @property def processed_file_names(self) -> str: return 'data.pt' def download(self) -> None: for file_name in self.raw_file_names: fs.cp(f'{self.url}/{file_name}.zip', self.raw_dir, extract=True) def _process_df(self, feat_df: Any, edge_df: Any, class_df: Any) -> Tuple[Any, Any, Any]: return feat_df, edge_df, class_df def process(self) -> None: import pandas as pd feat_df = pd.read_csv(self.raw_paths[0], header=None) edge_df = pd.read_csv(self.raw_paths[1]) class_df = pd.read_csv(self.raw_paths[2]) columns = {0: 'txId', 1: 'time_step'} feat_df = feat_df.rename(columns=columns) feat_df, edge_df, class_df = self._process_df( feat_df, edge_df, class_df, ) x = torch.from_numpy(feat_df.loc[:, 2:].values).to(torch.float) # There exists 3 different classes in the dataset: # 0=licit, 1=illicit, 2=unknown mapping = {'unknown': 2, '1': 1, '2': 0} class_df['class'] = class_df['class'].map(mapping) y = torch.from_numpy(class_df['class'].values) mapping = {idx: i for i, idx in enumerate(feat_df['txId'].values)} edge_df['txId1'] = edge_df['txId1'].map(mapping) edge_df['txId2'] = edge_df['txId2'].map(mapping) edge_index = torch.from_numpy(edge_df.values).t().contiguous() # Timestamp based split: # train_mask: 1 - 34 time_step, test_mask: 35-49 time_step time_step = torch.from_numpy(feat_df['time_step'].values) train_mask = (time_step < 35) & (y != 2) test_mask = (time_step >= 35) & (y != 2) data = Data(x=x, edge_index=edge_index, y=y, train_mask=train_mask, test_mask=test_mask) if self.pre_transform is not None: data = self.pre_transform(data) self.save([data], self.processed_paths[0]) @property def num_classes(self) -> int: return 2