datasets

src/beignet/datasets/__init__.py

@@ -73,6 +73,7 @@
 from ._skempi_dataset import SKEMPIDataset
 from ._swissprot_dataset import SwissProtDataset
 from ._tdc_dataset import TDCDataset
+from ._therapeutic_antibody_profiler_dataset import TherapeuticAntibodyProfilerDataset
 from ._tox21_dataset import Tox21Dataset
 from ._toxcast_dataset import ToxCastDataset
 from ._trajectory_dataset import TrajectoryDataset
@@ -164,6 +165,7 @@
     "SizedSequenceDataset",
     "SwissProtDataset",
     "TDCDataset",
+    "TherapeuticAntibodyProfilerDataset",
     "Tox21Dataset",
     "ToxCastDataset",
     "TrEMBLDataset",

src/beignet/datasets/_therapeutic_antibody_profiler_dataset.py

@@ -0,0 +1,50 @@
+from pathlib import Path
+from typing import Callable
+
+from beignet.transforms import Transform
+
+from ._tdc_dataset import TDCDataset
+
+
+class TherapeuticAntibodyProfilerDataset(TDCDataset):
+    def __init__(
+        self,
+        root: str | Path,
+        *,
+        download: bool = False,
+        transform: Callable | Transform | None = None,
+        target_transform: Callable | Transform | None = None,
+    ):
+        r"""
+        Parameters
+        ----------
+        root : str | Path
+            Root directory of dataset.
+
+        download: bool
+            If `True`, downloads the dataset to the root directory. If dataset
+            already exists, it is not redownloaded. Default, `False`.
+
+        transform : Callable | Transform | None
+            Transforms the input.
+
+        target_transform : Callable | Transform | None
+            Transforms the target.
+        """
+        super().__init__(
+            root=root,
+            download=download,
+            identifier=4167113,
+            suffix="tsv",
+            checksum="md5:567e3550e984bc73e0391b3b8ee5b231",
+            x_keys=["X"],
+            y_keys=[
+                "CDR_Length",
+                "PNC",
+                "PPC",
+                "PSH",
+                "SFvCSP",
+            ],
+            transform=transform,
+            target_transform=target_transform,
+        )