glitchlings 0.4.1__cp312-cp312-macosx_11_0_universal2.whl → 0.4.3__cp312-cp312-macosx_11_0_universal2.whl

glitchlings/dlc/__init__.py

@@ -1,5 +1,7 @@
 """Optional DLC integrations for Glitchlings."""
 
 from .huggingface import install as install_huggingface
+from .pytorch import install as install_pytorch
+from .pytorch_lightning import install as install_pytorch_lightning
 
-__all__ = ["install_huggingface"]
+__all__ = ["install_huggingface", "install_pytorch", "install_pytorch_lightning"]

glitchlings/dlc/_shared.py

@@ -0,0 +1,68 @@
+"""Shared utilities for DLC integrations."""
+
+from __future__ import annotations
+
+from collections.abc import Callable, Sequence
+from typing import Any
+
+
+def resolve_environment(
+    env: Any,
+    *,
+    loader: Callable[[str], Any],
+    environment_type: type[Any],
+) -> Any:
+    """Return a fully-instantiated verifier environment."""
+    if isinstance(env, str):
+        env = loader(env)
+
+    if not isinstance(env, environment_type):
+        raise TypeError("Invalid environment type")
+
+    return env
+
+
+def resolve_columns(dataset: Any, columns: Sequence[str] | None) -> list[str]:
+    """Identify which dataset columns should be corrupted."""
+    available = set(getattr(dataset, "column_names", ()))
+
+    if columns is not None:
+        missing = sorted(set(columns) - available)
+        if missing:
+            missing_str = ", ".join(missing)
+            raise ValueError(f"Columns not found in dataset: {missing_str}")
+        return list(columns)
+
+    for candidate in ("prompt", "question"):
+        if candidate in available:
+            return [candidate]
+
+    try:
+        dataset_length = len(dataset)
+    except TypeError:
+        preview_rows: list[dict[str, Any]]
+        take_fn = getattr(dataset, "take", None)
+        if callable(take_fn):
+            preview_rows = list(take_fn(1))
+        else:
+            iterator = iter(dataset)
+            try:
+                first_row = next(iterator)
+            except StopIteration:
+                preview_rows = []
+            else:
+                preview_rows = [first_row]
+        sample = dict(preview_rows[0]) if preview_rows else {}
+    else:
+        sample = dataset[0] if dataset_length else {}
+    inferred = [
+        name for name in getattr(dataset, "column_names", ()) if isinstance(sample.get(name), str)
+    ]
+
+    if inferred:
+        return inferred
+
+    raise ValueError("Unable to determine which dataset columns to corrupt.")
+
+
+__all__ = ["resolve_columns", "resolve_environment"]

glitchlings/dlc/huggingface.py

@@ -3,21 +3,15 @@
 from __future__ import annotations
 
 from collections.abc import Iterable, Sequence
-from typing import Any
+from typing import Any, cast
 
-try:  # pragma: no cover - optional dependency is required at runtime
-    from datasets import Dataset as _DatasetsDataset
-except ModuleNotFoundError as _datasets_error:  # pragma: no cover - optional dependency
-    _DatasetsDataset = None  # type: ignore[assignment]
-else:
-    _datasets_error = None
-
-from ..zoo import Gaggle, Glitchling, summon
+from ..compat import datasets, get_datasets_dataset, require_datasets
+from ..util.adapters import coerce_gaggle
+from ..zoo import Gaggle, Glitchling
 
 
 def _normalise_columns(column: str | Sequence[str]) -> list[str]:
     """Normalise a column specification to a list."""
-
     if isinstance(column, str):
         return [column]
 
@@ -27,20 +21,6 @@ def _normalise_columns(column: str | Sequence[str]) -> list[str]:
     return normalised
 
 
-def _as_gaggle(glitchlings: Glitchling | Gaggle | str | Iterable[str | Glitchling], seed: int) -> Gaggle:
-    """Coerce any supported glitchling specification into a :class:`Gaggle`."""
-
-    if isinstance(glitchlings, Gaggle):
-        return glitchlings
-
-    if isinstance(glitchlings, (Glitchling, str)):
-        resolved: Iterable[str | Glitchling] = [glitchlings]
-    else:
-        resolved = glitchlings
-
-    return summon(list(resolved), seed=seed)
-
-
 def _glitch_dataset(
     dataset: Any,
     glitchlings: Glitchling | Gaggle | str | Iterable[str | Glitchling],
@@ -48,23 +28,28 @@ def _glitch_dataset(
     *,
     seed: int = 151,
 ) -> Any:
-    """Internal helper implementing :meth:`Dataset.glitch`."""
-
+    """Apply glitchlings to the provided dataset columns."""
     columns = _normalise_columns(column)
-    gaggle = _as_gaggle(glitchlings, seed=seed)
+    gaggle = coerce_gaggle(glitchlings, seed=seed)
     return gaggle.corrupt_dataset(dataset, columns)
 
 
 def _ensure_dataset_class() -> Any:
     """Return the Hugging Face :class:`~datasets.Dataset` patched with ``.glitch``."""
-
-    if _DatasetsDataset is None:  # pragma: no cover - datasets is an install-time dependency
-        message = "datasets is not installed"
-        raise ModuleNotFoundError(message) from _datasets_error
-
-    if getattr(_DatasetsDataset, "glitch", None) is None:
-
-        def glitch(  # type: ignore[override]
+    dataset_cls = get_datasets_dataset()
+    if dataset_cls is None:  # pragma: no cover - datasets is an install-time dependency
+        require_datasets("datasets is not installed")
+        dataset_cls = get_datasets_dataset()
+        if dataset_cls is None:
+            message = "datasets is not installed"
+            error = datasets.error
+            if error is not None:
+                raise ModuleNotFoundError(message) from error
+            raise ModuleNotFoundError(message)
+
+    if getattr(dataset_cls, "glitch", None) is None:
+
+        def glitch(
             self: Any,
             glitchlings: Glitchling | Gaggle | str | Iterable[str | Glitchling],
             *,
@@ -73,24 +58,24 @@ def _ensure_dataset_class() -> Any:
             **_: Any,
         ) -> Any:
             """Return a lazily corrupted copy of the dataset."""
-
             return _glitch_dataset(self, glitchlings, column, seed=seed)
 
-        setattr(_DatasetsDataset, "glitch", glitch)
+        setattr(dataset_cls, "glitch", glitch)
 
-    return _DatasetsDataset
+    return cast(type[Any], dataset_cls)
 
 
 def install() -> None:
     """Monkeypatch the Hugging Face :class:`~datasets.Dataset` with ``.glitch``."""
-
     _ensure_dataset_class()
 
 
-if _DatasetsDataset is not None:
+Dataset: type[Any] | None
+_DatasetAlias = get_datasets_dataset()
+if _DatasetAlias is not None:
     Dataset = _ensure_dataset_class()
 else:  # pragma: no cover - datasets is an install-time dependency
-    Dataset = None  # type: ignore[assignment]
+    Dataset = None
 
 
 __all__ = ["Dataset", "install"]

glitchlings/dlc/prime.py

@@ -4,79 +4,60 @@ from __future__ import annotations
 
 from collections.abc import Iterable, Sequence
 from enum import Enum
-from typing import Any, Callable
+from typing import Any, Callable, Protocol, cast
 
-import verifiers as vf
+from ..compat import require_datasets, require_jellyfish, require_verifiers
+from ..util.adapters import coerce_gaggle
+from ..zoo import Gaggle, Glitchling, Mim1c, Typogre
+from ._shared import resolve_columns as _resolve_columns_shared
+from ._shared import resolve_environment as _resolve_environment_shared
 
-from jellyfish import damerau_levenshtein_distance
 
-try:
-    from .huggingface import Dataset
-except ModuleNotFoundError:  # pragma: no cover - optional dependency
-    Dataset = object  # type: ignore[assignment]
-else:
-    if Dataset is None:  # pragma: no cover - optional dependency
-        Dataset = object  # type: ignore[assignment]
+class VerifierEnvironment(Protocol):
+    """Minimal interface for verifiers environments."""
 
-from ..zoo import Gaggle, Glitchling, Mim1c, Typogre, summon
+    dataset: Any
 
 
-def _resolve_environment(env: str | vf.Environment) -> vf.Environment:
-    """Return a fully-instantiated verifier environment."""
-
-    if isinstance(env, str):
-        env = vf.load_environment(env)
+class VerifierSingleTurnEnv(Protocol):
+    """Minimal interface for single-turn verifier environments."""
 
-    if not isinstance(env, vf.Environment):
-        raise TypeError("Invalid environment type")
+    dataset: Any
+    rubric: Any
 
-    return env
 
+vf = require_verifiers("verifiers is not installed; install glitchlings[prime]")
+_jellyfish = require_jellyfish("jellyfish is not installed; install glitchlings[prime]")
+damerau_levenshtein_distance = _jellyfish.damerau_levenshtein_distance
 
-def _resolve_columns(dataset: Dataset, columns: Sequence[str] | None) -> list[str]:
-    """Identify which dataset columns should be corrupted."""
-
-    available = set(dataset.column_names)
+try:
+    from .huggingface import Dataset as _HuggingFaceDataset
+except ModuleNotFoundError:  # pragma: no cover - optional dependency
+    _HuggingFaceDataset = None
+else:
+    if _HuggingFaceDataset is None:  # pragma: no cover - optional dependency
+        _HuggingFaceDataset = None
 
-    if columns is not None:
-        missing = sorted(set(columns) - available)
-        if missing:
-            missing_str = ", ".join(missing)
-            raise ValueError(f"Columns not found in dataset: {missing_str}")
-        return list(columns)
+Dataset: type[Any]
+if _HuggingFaceDataset is None:
+    Dataset = object
+else:
+    Dataset = _HuggingFaceDataset
 
-    for candidate in ("prompt", "question"):
-        if candidate in available:
-            return [candidate]
 
-    try:
-        dataset_length = len(dataset)  # type: ignore[arg-type]
-    except TypeError:
-        preview_rows: list[dict[str, Any]]
-        take_fn = getattr(dataset, "take", None)
-        if callable(take_fn):
-            preview_rows = list(take_fn(1))
-        else:
-            iterator = iter(dataset)
-            try:
-                first_row = next(iterator)
-            except StopIteration:
-                preview_rows = []
-            else:
-                preview_rows = [first_row]
-        sample = dict(preview_rows[0]) if preview_rows else {}
-    else:
-        sample = dataset[0] if dataset_length else {}
-    inferred = [
-        name
-        for name in dataset.column_names
-        if isinstance(sample.get(name), str)
-    ]
+def _resolve_environment(env: str | VerifierEnvironment) -> VerifierEnvironment:
+    """Return a fully-instantiated verifier environment."""
+    resolved = _resolve_environment_shared(
+        env,
+        loader=vf.load_environment,
+        environment_type=cast(type[Any], vf.Environment),
+    )
+    return cast(VerifierEnvironment, resolved)
 
-    if inferred:
-        return inferred
 
-    raise ValueError("Unable to determine which dataset columns to corrupt.")
+def _resolve_columns(dataset: Any, columns: Sequence[str] | None) -> list[str]:
+    """Identify which dataset columns should be corrupted."""
+    return _resolve_columns_shared(dataset, columns)
 
 
 class Difficulty(Enum):
@@ -90,12 +71,11 @@ class Difficulty(Enum):
 
 
 def tutorial_level(
-    env: vf.Environment | str,
+    env: VerifierEnvironment | str,
     seed: int = 151,
     difficulty: Difficulty = Difficulty.Normal,
-) -> vf.Environment:
+) -> VerifierEnvironment:
     """Create a low-corruption environment using tuned defaults."""
-
     tuned_mim1c = Mim1c(rate=0.01 * difficulty.value)
     tuned_typogre = Typogre(rate=0.025 * difficulty.value)
 
@@ -107,28 +87,19 @@ def tutorial_level(
 
 
 def load_environment(
-    env: str | vf.Environment,
+    env: str | VerifierEnvironment,
     glitchlings: Iterable[str | Glitchling] | Glitchling | str | Gaggle | None = None,
     *,
     seed: int = 151,
     columns: Sequence[str] | None = None,
-) -> vf.Environment:
+) -> VerifierEnvironment:
     """Load an environment and optionally corrupt it with glitchlings."""
-
     environment = _resolve_environment(env)
 
     if glitchlings is None:
         return environment
 
-    if isinstance(glitchlings, Gaggle):
-        gaggle = glitchlings
-    else:
-        if isinstance(glitchlings, (Glitchling, str)):
-            resolved = [glitchlings]
-        else:
-            resolved = list(glitchlings)
-
-        gaggle = summon(resolved, seed=seed)
+    gaggle = coerce_gaggle(glitchlings, seed=seed)
 
     dataset = environment.dataset
     corrupt_columns = _resolve_columns(dataset, columns)
@@ -142,21 +113,11 @@ def _as_gaggle(
     seed: int,
 ) -> Gaggle:
     """Coerce any supported glitchling specification into a :class:`Gaggle`."""
-
-    if isinstance(glitchlings, Gaggle):
-        return glitchlings
-
-    if isinstance(glitchlings, (Glitchling, str)):
-        resolved: Iterable[str | Glitchling] = [glitchlings]
-    else:
-        resolved = glitchlings
-
-    return summon(list(resolved), seed=seed)
+    return coerce_gaggle(glitchlings, seed=seed)
 
 
 def _extract_completion_text(completion: Any) -> str:
     """Normalise a completion payload into a plain string."""
-
     if isinstance(completion, str):
         return completion
 
@@ -175,11 +136,10 @@ def symmetric_damerau_levenshtein_similarity(
     answer: str,
 ) -> float:
     """Return ``1 - (distance / max_len)`` using Damerau-Levenshtein distance."""
-
     completion_text = _extract_completion_text(completion)
     target = answer or ""
     denominator = max(len(completion_text), len(target), 1)
-    distance = damerau_levenshtein_distance(completion_text, target)
+    distance = cast(int, damerau_levenshtein_distance(completion_text, target))
     score = 1.0 - (distance / denominator)
     return max(0.0, min(1.0, score))
 
@@ -199,32 +159,34 @@ def echo_chamber(
     reward_function: Callable[..., float] | None = None,
     split: str | None = None,
     **load_dataset_kwargs: Any,
-) -> vf.Environment:
+) -> VerifierSingleTurnEnv:
     """Create an Echo Chamber Prime environment from a Hugging Face dataset column.
 
     Args:
         dataset_id: Identifier of the Hugging Face dataset to load.
         column: Name of the column whose text should be glitched.
         glitchlings: Glitchling specifiers that will corrupt the prompts.
-        seed: RNG seed forwarded to :func:`summon`.
+        seed: RNG seed forwarded to :func:`glitchlings.util.adapters.coerce_gaggle`.
         instructions: System instructions supplied to the environment prompts.
         reward_function: Optional callable used to score completions. Defaults to
             :func:`symmetric_damerau_levenshtein_similarity` when omitted.
         split: Optional dataset split to load.
         **load_dataset_kwargs: Extra keyword arguments forwarded to
             :func:`datasets.load_dataset`.
-    """
 
-    try:
-        from datasets import Dataset as HFDataset, DatasetDict, load_dataset
-    except ModuleNotFoundError as exc:  # pragma: no cover - optional dependency
+    """
+    datasets_module = require_datasets("datasets is required to build an echo chamber")
+    load_dataset = getattr(datasets_module, "load_dataset", None)
+    if load_dataset is None:  # pragma: no cover - defensive
         message = "datasets is required to build an echo chamber"
-        raise ModuleNotFoundError(message) from exc
+        raise ModuleNotFoundError(message)
 
-    hf_dataset: HFDataset | DatasetDict
+    dataset_dict_cls = getattr(datasets_module, "DatasetDict", dict)
+
+    hf_dataset: Any
     if split is None:
         hf_dataset = load_dataset(dataset_id, **load_dataset_kwargs)
-        if isinstance(hf_dataset, DatasetDict):
+        if isinstance(hf_dataset, dataset_dict_cls):
             try:
                 hf_dataset = next(iter(hf_dataset.values()))
             except StopIteration as exc:  # pragma: no cover - defensive
@@ -232,10 +194,8 @@ def echo_chamber(
     else:
         hf_dataset = load_dataset(dataset_id, split=split, **load_dataset_kwargs)
 
-    if isinstance(hf_dataset, DatasetDict):
-        raise ValueError(
-            "Specify which split to use when the dataset loads as a DatasetDict."
-        )
+    if isinstance(hf_dataset, dataset_dict_cls):
+        raise ValueError("Specify which split to use when the dataset loads as a DatasetDict.")
 
     filtered_dataset = hf_dataset.filter(
         lambda row: row.get(column) is not None,
@@ -259,7 +219,7 @@ def echo_chamber(
     )
 
     try:
-        dataset_length = len(base_dataset)  # type: ignore[arg-type]
+        dataset_length = len(base_dataset)
     except TypeError:
         preview_rows: list[dict[str, Any]]
         take_fn = getattr(base_dataset, "take", None)
@@ -288,4 +248,7 @@ def echo_chamber(
 
     rubric_func = reward_function or symmetric_damerau_levenshtein_similarity
     rubric = vf.Rubric(funcs=[rubric_func], weights=[1.0])
-    return vf.SingleTurnEnv(dataset=glitched_dataset, rubric=rubric)
+    return cast(
+        VerifierSingleTurnEnv,
+        vf.SingleTurnEnv(dataset=glitched_dataset, rubric=rubric),
+    )

glitchlings/dlc/pytorch.py

@@ -0,0 +1,216 @@
+"""Integration helpers for PyTorch data loaders."""
+
+from __future__ import annotations
+
+from collections.abc import Iterable, Iterator, Mapping, MutableMapping, Sequence
+from typing import Any, cast
+
+from ..compat import get_torch_dataloader, require_torch
+from ..compat import torch as _torch_dependency
+from ..util.adapters import coerce_gaggle
+from ..zoo import Gaggle, Glitchling
+from ..zoo.core import _is_transcript
+
+
+def _normalise_columns(columns: str | int | Sequence[str | int] | None) -> list[str | int] | None:
+    """Normalise a column specification into a list of keys or indices."""
+    if columns is None:
+        return None
+
+    if isinstance(columns, (str, int)):
+        return [columns]
+
+    normalised = list(columns)
+    if not normalised:
+        raise ValueError("At least one column must be specified")
+    return normalised
+
+
+def _is_textual_candidate(value: Any) -> bool:
+    """Return ``True`` when ``value`` looks like text that glitchlings can corrupt."""
+    if isinstance(value, str):
+        return True
+
+    if _is_transcript(value, allow_empty=False, require_all_content=True):
+        return True
+
+    if isinstance(value, Sequence) and not isinstance(value, (bytes, bytearray, str)):
+        if not value:
+            return False
+        if all(isinstance(item, str) for item in value):
+            return True
+        if _is_transcript(list(value), allow_empty=False, require_all_content=True):
+            return True
+
+    return False
+
+
+def _corrupt_text(value: Any, gaggle: Gaggle) -> Any:
+    """Return ``value`` with glitchlings applied when possible."""
+    if isinstance(value, str):
+        return gaggle.corrupt(value)
+
+    if _is_transcript(value, allow_empty=True):
+        return gaggle.corrupt(value)
+
+    if isinstance(value, list) and value and all(isinstance(item, str) for item in value):
+        return [gaggle.corrupt(item) for item in value]
+
+    if isinstance(value, tuple) and value and all(isinstance(item, str) for item in value):
+        return tuple(gaggle.corrupt(item) for item in value)
+
+    return value
+
+
+def _apply_to_batch(batch: Any, targets: list[str | int] | None, gaggle: Gaggle) -> Any:
+    """Return ``batch`` with glitchlings applied to the specified ``targets``."""
+    if targets is None:
+        return _corrupt_text(batch, gaggle)
+
+    if isinstance(batch, Mapping):
+        mutated = cast(MutableMapping[str, Any], dict(batch))
+        for key in targets:
+            if not isinstance(key, str):
+                raise TypeError("Mapping batches require string column names")
+            if key not in mutated:
+                raise ValueError(f"Column '{key}' not found in DataLoader batch")
+            mutated[key] = _corrupt_text(mutated[key], gaggle)
+        return mutated
+
+    if isinstance(batch, Sequence) and not isinstance(batch, (bytes, bytearray, str)):
+        mutated_sequence = list(batch)
+        for index in targets:
+            if not isinstance(index, int):
+                raise TypeError("Sequence batches require integer column indices")
+            try:
+                mutated_sequence[index] = _corrupt_text(mutated_sequence[index], gaggle)
+            except IndexError as exc:  # pragma: no cover - defensive
+                raise IndexError("Column index out of range for DataLoader batch") from exc
+        if isinstance(batch, tuple):
+            return tuple(mutated_sequence)
+        return mutated_sequence
+
+    raise TypeError("Unsupported DataLoader batch type for glitching")
+
+
+def _infer_targets(batch: Any) -> list[str | int] | None:
+    """Infer which fields should be glitched from a representative ``batch``."""
+    if isinstance(batch, Mapping):
+        inferred = [key for key, value in batch.items() if _is_textual_candidate(value)]
+        if inferred:
+            return inferred
+        raise ValueError("Unable to infer which mapping columns contain text")
+
+    if isinstance(batch, Sequence) and not isinstance(batch, (bytes, bytearray, str)):
+        inferred_indices: list[str | int] = [
+            idx for idx, value in enumerate(batch) if _is_textual_candidate(value)
+        ]
+        if inferred_indices:
+            return inferred_indices
+        raise ValueError("Unable to infer which sequence indices contain text")
+
+    if _is_textual_candidate(batch):
+        return None
+
+    raise TypeError("Unsupported DataLoader batch type for glitching")
+
+
+class _GlitchedDataLoader(Iterable[Any]):
+    """Wrapper that applies glitchlings lazily to each batch from a data loader."""
+
+    def __init__(
+        self,
+        dataloader: Any,
+        gaggle: Gaggle,
+        *,
+        columns: list[str | int] | None,
+    ) -> None:
+        self._dataloader = dataloader
+        self._gaggle = gaggle
+        self._explicit_columns = columns
+        self._inferred_columns: list[str | int] | None | _Sentinel = _UNINITIALISED
+
+    def __iter__(self) -> Iterator[Any]:
+        # Reset all glitchling RNGs before each fresh pass for determinism.
+        self._gaggle.sort_glitchlings()
+        for batch in self._dataloader:
+            targets = self._resolve_columns(batch)
+            yield _apply_to_batch(batch, targets, self._gaggle)
+
+    def __len__(self) -> int:
+        return len(self._dataloader)
+
+    def __getattr__(self, attribute: str) -> Any:
+        return getattr(self._dataloader, attribute)
+
+    def _resolve_columns(self, batch: Any) -> list[str | int] | None:
+        if self._explicit_columns is not None:
+            return self._explicit_columns
+
+        if self._inferred_columns is _UNINITIALISED:
+            self._inferred_columns = _infer_targets(batch)
+
+        return cast(list[str | int] | None, self._inferred_columns)
+
+
+class _Sentinel:
+    """Sentinel type for deferred column inference."""
+
+
+_UNINITIALISED = _Sentinel()
+
+
+def _ensure_dataloader_class() -> type[Any]:
+    """Return :class:`torch.utils.data.DataLoader` patched with ``.glitch``."""
+    dataloader_cls = get_torch_dataloader()
+    if dataloader_cls is None:
+        require_torch("torch is not installed; install glitchlings[torch]")
+        dataloader_cls = get_torch_dataloader()
+        if dataloader_cls is None:  # pragma: no cover - defensive
+            message = "torch.utils.data.DataLoader is not available"
+            error = _torch_dependency.error
+            if error is not None:
+                raise ModuleNotFoundError(message) from error
+            raise ModuleNotFoundError(message)
+
+    if getattr(dataloader_cls, "glitch", None) is None:
+
+        def glitch(
+            self: Any,
+            glitchlings: Iterable[str | Glitchling] | Glitchling | str | Gaggle,
+            *,
+            columns: str | int | Sequence[str | int] | None = None,
+            seed: int = 151,
+        ) -> _GlitchedDataLoader:
+            """Return a lazily glitched view of the loader's batches."""
+            gaggle = coerce_gaggle(glitchlings, seed=seed)
+            normalised = _normalise_columns(columns)
+            return _GlitchedDataLoader(self, gaggle, columns=normalised)
+
+        setattr(dataloader_cls, "glitch", glitch)
+
+    return cast(type[Any], dataloader_cls)
+
+
+def _optional_dataloader_class() -> type[Any] | None:
+    """Return the PyTorch :class:`~torch.utils.data.DataLoader` when importable."""
+    dataloader_cls = get_torch_dataloader()
+    if dataloader_cls is None:
+        return None
+    return cast(type[Any], dataloader_cls)
+
+
+def install() -> None:
+    """Monkeypatch PyTorch's :class:`~torch.utils.data.DataLoader` with ``.glitch``."""
+    _ensure_dataloader_class()
+
+
+DataLoader: type[Any] | None
+_DataLoaderAlias = _optional_dataloader_class()
+if _DataLoaderAlias is not None:
+    DataLoader = _ensure_dataloader_class()
+else:  # pragma: no cover - torch is an optional dependency
+    DataLoader = None
+
+
+__all__ = ["DataLoader", "install"]