Add BreastInvasiveCarcinoma dataset (#7905)

This is a dataset that was generated by integrating the breast cancer
(BRCA TCGA) dataset from the cBioPortal (cbioportal.org) and a
biological network for node connections from Pathway Commons
(www.pathwaycommons.org). The dataset contains the gene features of each
patient and the overall survival time (in months) of each patient, which
are the labels.

---------

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Co-authored-by: Akihiro Nitta <nitta@akihironitta.com>
Co-authored-by: Matthias Fey <matthias.fey@tu-dortmund.de>

CHANGELOG.md

@@ -7,6 +7,7 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/).
 
 ### Added
 
+- Added the `BrcaTcga` dataset ([#7905](https://github.com/pyg-team/pytorch_geometric/pull/7905))
 - Added the `MyketDataset` ([#7959](https://github.com/pyg-team/pytorch_geometric/pull/7959))
 - Added a multi-GPU `ogbn-papers100M` example ([#7921](https://github.com/pyg-team/pytorch_geometric/pull/7921))
 - Added `group_argsort` implementation ([#7948](https://github.com/pyg-team/pytorch_geometric/pull/7948))

torch_geometric/datasets/__init__.py

@@ -73,6 +73,7 @@
 from .jodie import JODIEDataset
 from .wikidata import Wikidata5M
 from .myket import MyketDataset
+from .brca_tgca import BrcaTcga
 
 from .dbp15k import DBP15K
 from .aminer import AMiner
@@ -178,6 +179,7 @@
     'JODIEDataset',
     'Wikidata5M',
     'MyketDataset',
+    'BrcaTcga',
 ]
 
 hetero_datasets = [

torch_geometric/datasets/brca_tgca.py

@@ -0,0 +1,104 @@
+import os
+import os.path as osp
+import shutil
+from typing import Callable, List, Optional
+
+import numpy as np
+import torch
+
+from torch_geometric.data import (
+    Data,
+    InMemoryDataset,
+    download_url,
+    extract_zip,
+)
+
+
+class BrcaTcga(InMemoryDataset):
+    r"""The breast cancer (BRCA TCGA) dataset from the `cBioPortal
+    <https://www.cbioportal.org>`_ and the biological network for node
+    connections from `Pathway Commons <https://www.pathwaycommons.org>`_.
+    The dataset contains the gene features of 1,082 patients, and the overall
+    survival time (in months) of each patient as label.
+
+    Pre-processing and example model codes on how to use this dataset can be
+    found `here <https://github.com/cannin/pyg_pathway_commons_cbioportal>`_.
+
+    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`)
+        pre_filter (callable, optional): A function that takes in an
+            :obj:`torch_geometric.data.Data` object and returns a boolean
+            value, indicating whether the data object should be included in the
+            final dataset. (default: :obj:`None`)
+
+    **STATS:**
+
+    .. list-table::
+        :widths: 10 10 10 10
+        :header-rows: 1
+
+        * - #graphs
+          - #nodes
+          - #edges
+          - #features
+        * - 1,082
+          - 9,288
+          - 271,771
+          - 1,082
+    """
+    url = 'https://zenodo.org/record/8251328/files/brca_tcga.zip?download=1'
+
+    def __init__(
+        self,
+        root: str,
+        transform: Optional[Callable] = None,
+        pre_transform: Optional[Callable] = None,
+        pre_filter: Optional[Callable] = None,
+    ):
+        super().__init__(root, transform, pre_transform, pre_filter)
+        self.load(self.processed_paths[0])
+
+    @property
+    def raw_file_names(self) -> List[str]:
+        return ['graph_idx.csv', 'graph_labels.csv', 'edge_index.pt']
+
+    @property
+    def processed_file_names(self) -> str:
+        return 'data.pt'
+
+    def download(self):
+        path = download_url(self.url, self.root)
+        extract_zip(path, self.root)
+        os.unlink(path)
+        shutil.rmtree(self.raw_dir)
+        os.rename(osp.join(self.root, 'brca_tcga'), self.raw_dir)
+
+    def process(self):
+        import pandas as pd
+
+        graph_feat = pd.read_csv(self.raw_paths[0], index_col=0).values
+        graph_feat = torch.from_numpy(graph_feat).to(torch.float)
+        graph_label = np.loadtxt(self.raw_paths[1], delimiter=',')
+        graph_label = torch.from_numpy(graph_label).to(torch.float)
+        edge_index = torch.load(self.raw_paths[2])
+
+        data_list = []
+        for x, y in zip(graph_feat, graph_label):
+            data = Data(x=x, edge_index=edge_index, y=y)
+
+            if self.pre_filter is not None and not self.pre_filter(data):
+                continue
+            if self.pre_transform is not None:
+                data = self.pre_transform(data)
+
+            data_list.append(data)
+
+        self.save(data_list, self.processed_paths[0])