UniProt datasets

src/beignet/datasets/__init__.py

@@ -1,7 +1,9 @@
-from .__uniref_dataset import _UniRefDataset
+from .__uniprot_dataset import _UniRefDataset
 from ._fasta_dataset import FASTADataset
 from ._sequence_dataset import SequenceDataset
 from ._sized_sequence_dataset import SizedSequenceDataset
+from ._swissprot_dataset import SwissProtDataset
+from ._trembl_dataset import TrEMBLDataset
 from ._uniref50_dataset import UniRef50Dataset
 from ._uniref90_dataset import UniRef90Dataset
 from ._uniref100_dataset import UniRef100Dataset
@@ -10,6 +12,8 @@
     "FASTADataset",
     "SequenceDataset",
     "SizedSequenceDataset",
+    "SwissProtDataset",
+    "TrEMBLDataset",
     "UniRef100Dataset",
     "UniRef50Dataset",
     "UniRef90Dataset",

src/beignet/datasets/_swissprot_dataset.py

@@ -0,0 +1,39 @@
+from os import PathLike
+from typing import Callable
+
+from beignet.transforms import Transform
+
+from .__uniprot_dataset import _UniRefDataset
+
+
+class SwissProtDataset(_UniRefDataset):
+    def __init__(
+        self,
+        root: str | PathLike | None = None,
+        *,
+        transform: Callable | Transform | None = None,
+        target_transform: Callable | Transform | None = None,
+    ) -> None:
+        """
+        Parameters
+        ----------
+        root : str | PathLike, optional
+            Root directory where the dataset subdirectory exists or, if
+            `download` is `True`, the directory where the dataset subdirectory
+            will be created and the dataset downloaded.
+
+        transform : Callable, optional
+            A `Callable` or `Transform` that that maps a sequence to a
+            transformed sequence (default: `None`).
+
+        target_transform : Callable, optional
+            A `Callable` or `Transform` that maps a target to a transformed
+            target (default: `None`).
+        """
+        super().__init__(
+            "https://ftp.uniprot.org/pub/databases/uniprot/current_release/knowledgebase/complete/uniprot_sprot.fasta.gz",
+            root,
+            "md5:0766df3e5785fc5f1cfc496aa89e86ad",
+            transform=transform,
+            target_transform=target_transform,
+        )

src/beignet/datasets/_trembl_dataset.py

@@ -0,0 +1,39 @@
+from os import PathLike
+from typing import Callable
+
+from beignet.transforms import Transform
+
+from .__uniprot_dataset import _UniRefDataset
+
+
+class TrEMBLDataset(_UniRefDataset):
+    def __init__(
+        self,
+        root: str | PathLike | None = None,
+        *,
+        transform: Callable | Transform | None = None,
+        target_transform: Callable | Transform | None = None,
+    ) -> None:
+        """
+        Parameters
+        ----------
+        root : str | PathLike, optional
+            Root directory where the dataset subdirectory exists or, if
+            `download` is `True`, the directory where the dataset subdirectory
+            will be created and the dataset downloaded.
+
+        transform : Callable, optional
+            A `Callable` or `Transform` that that maps a sequence to a
+            transformed sequence (default: `None`).
+
+        target_transform : Callable, optional
+            A `Callable` or `Transform` that maps a target (a cluster
+            identifier) to a transformed target (default: `None`).
+        """
+        super().__init__(
+            "https://ftp.uniprot.org/pub/databases/uniprot/current_release/knowledgebase/complete/uniprot_trembl.fasta.gz",
+            root,
+            "md5:56f0f20479a88d28fb51db7ef4df90ed",
+            transform=transform,
+            target_transform=target_transform,
+        )

src/beignet/datasets/_uniref100_dataset.py

@@ -2,7 +2,7 @@
 from typing import Callable
 
 from ..transforms import Transform
-from .__uniref_dataset import _UniRefDataset
+from .__uniprot_dataset import _UniRefDataset
 
 
 class UniRef100Dataset(_UniRefDataset):

src/beignet/datasets/_uniref50_dataset.py

@@ -3,7 +3,7 @@
 
 from beignet.transforms import Transform
 
-from .__uniref_dataset import _UniRefDataset
+from .__uniprot_dataset import _UniRefDataset
 
 
 class UniRef50Dataset(_UniRefDataset):

src/beignet/datasets/_uniref90_dataset.py

@@ -2,7 +2,7 @@
 from typing import Callable
 
 from ..transforms import Transform
-from .__uniref_dataset import _UniRefDataset
+from .__uniprot_dataset import _UniRefDataset
 
 
 class UniRef90Dataset(_UniRefDataset):