datasets

mkdocs.yml

@@ -44,6 +44,7 @@ nav:
           - "Protein":
             - "Protein sequence": "reference/datasets/generation/protein/protein-sequence.md"
             - "Protein structure": "reference/datasets/generation/protein/protein-structure.md"
+          - "Reaction": "reference/datasets/generation/reaction.md"
           - "Retrosynthesis": "reference/datasets/generation/retrosynthesis.md"
         - "Prediction":
           - "Antibody affinity": "reference/datasets/prediction/antibody-affinity.md"

src/beignet/datasets/__init__.py

@@ -79,6 +79,7 @@
 from ._uniref50_dataset import UniRef50Dataset
 from ._uniref90_dataset import UniRef90Dataset
 from ._uniref100_dataset import UniRef100Dataset
+from ._uspto_dataset import USPTODataset
 from ._uspto_reaction_product_dataset import USPTOReactionProductDataset
 from ._veith_cytochrome_p450_1a2_inhibition_dataset import (
     VeithCytochromeP4501A2InhibitionDataset,
@@ -163,6 +164,7 @@
     "ToxCastDataset",
     "TrEMBLDataset",
     "TrajectoryDataset",
+    "USPTODataset",
     "USPTOReactionProductDataset",
     "UniProtDataset",
     "UniRef100Dataset",

src/beignet/datasets/_uspto_dataset.py

@@ -0,0 +1,44 @@
+from pathlib import Path
+from typing import Callable
+
+from beignet.transforms import Transform
+
+from ._tdc_dataset import TDCDataset
+
+
+class USPTODataset(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=0,
+            suffix="",
+            checksum="",
+            x_keys=[""],
+            y_keys=[""],
+            transform=transform,
+            target_transform=target_transform,
+        )