[Typing] 修复 dynamic_decode 以及示例中的类型标注错误 (#67295)

megemini · web-flow · commit 4eb9d6e999db · 2024-08-13T12:13:33.000+08:00
* [Fix] typing

* [Fix] typing

* [Fix] typing

* [Fix] DatasetBase from base and fleet
diff --git a/python/paddle/autograd/backward_mode.py b/python/paddle/autograd/backward_mode.py
@@ -31,8 +31,8 @@
 
 @framework.dygraph_only
 def backward(
-    tensors: list[Tensor],
-    grad_tensors: list[Tensor | None] | None = None,
+    tensors: Tensor | Sequence[Tensor],
+    grad_tensors: Tensor | Sequence[Tensor | None] | None = None,
     retain_graph: bool = False,
 ) -> None:
     """
diff --git a/python/paddle/base/executor.py b/python/paddle/base/executor.py
@@ -62,6 +62,9 @@
     from paddle._typing import PlaceLike
     from paddle._typing.device_like import _Place
     from paddle.base.dataset import DatasetBase
+    from paddle.distributed.fleet.dataset.dataset import (
+        DatasetBase as _FleetDatasetBase,
+    )
     from paddle.static import CompiledProgram
 
 __all__ = []
@@ -3156,7 +3159,7 @@ def _run_pipeline(
     def infer_from_dataset(
         self,
         program: Program | CompiledProgram | None = None,
-        dataset: DatasetBase | None = None,
+        dataset: DatasetBase | _FleetDatasetBase | None = None,
         scope: core.Scope | None = None,
         thread: int = 0,
         debug: bool = False,
@@ -3279,7 +3282,7 @@ def start_heter_trainer(
     def train_from_dataset(
         self,
         program: Program | CompiledProgram | None = None,
-        dataset: DatasetBase | None = None,
+        dataset: DatasetBase | _FleetDatasetBase | None = None,
         scope: core.Scope | None = None,
         thread: int = 0,
         debug: bool = False,
diff --git a/python/paddle/distributed/communication/group.py b/python/paddle/distributed/communication/group.py
@@ -309,7 +309,7 @@ def barrier(group=None):
             >>> import paddle
             >>> from paddle.distributed import init_parallel_env
 
-            >>> paddle.set_device('gpu:%d'%paddle.distributed.ParallelEnv().dev_id)
+            >>> paddle.set_device(f'gpu:{paddle.distributed.ParallelEnv().dev_id}')
             >>> init_parallel_env()
             >>> paddle.distributed.barrier()
     """
diff --git a/python/paddle/distributed/fleet/layers/mpu/mp_ops.py b/python/paddle/distributed/fleet/layers/mpu/mp_ops.py
@@ -822,7 +822,7 @@ def split(
             >>> import paddle.distributed.fleet as fleet
 
             >>> paddle.enable_static()
-            >>> paddle.set_device('gpu:%d'%paddle.distributed.ParallelEnv().dev_id)
+            >>> paddle.set_device(f'gpu:{paddle.distributed.ParallelEnv().dev_id}')
             >>> fleet.init(is_collective=True)
             >>> data = paddle.randint(0, 8, shape=[10,4])
             >>> emb_out = paddle.distributed.split(
diff --git a/python/paddle/geometric/sampling/neighbors.py b/python/paddle/geometric/sampling/neighbors.py
@@ -322,7 +322,9 @@ def weighted_sample_neighbors(
             >>> colptr = paddle.to_tensor(colptr, dtype="int64")
             >>> weight = paddle.to_tensor(weight, dtype="float32")
             >>> nodes = paddle.to_tensor(nodes, dtype="int64")
-            >>> out_neighbors, out_count = paddle.geometric.weighted_sample_neighbors(row, colptr, weight, nodes, sample_size=sample_size)
+            >>> out_neighbors, out_count = paddle.geometric.weighted_sample_neighbors(
+            ...     row, colptr, weight, nodes, sample_size=sample_size, return_eids=False
+            ... )
 
     """
 
diff --git a/python/paddle/io/dataloader/dataset.py b/python/paddle/io/dataloader/dataset.py
@@ -127,7 +127,7 @@ class IterableDataset(Dataset[_T]):
             >>> from paddle.io import IterableDataset
 
             >>> # define a random dataset
-            >>> class RandomDataset(IterableDataset):
+            >>> class RandomDataset(IterableDataset): # type: ignore[type-arg]
             ...     def __init__(self, num_samples):
             ...         self.num_samples = num_samples
             ...
@@ -159,7 +159,7 @@ class IterableDataset(Dataset[_T]):
             >>> import numpy as np
             >>> from paddle.io import IterableDataset, DataLoader, get_worker_info
 
-            >>> class SplitedIterableDataset(IterableDataset):
+            >>> class SplitedIterableDataset(IterableDataset): # type: ignore[type-arg]
             ...     def __init__(self, start, end):
             ...         self.start = start
             ...         self.end = end
@@ -214,7 +214,7 @@ class IterableDataset(Dataset[_T]):
             >>> import numpy as np
             >>> from paddle.io import IterableDataset, DataLoader, get_worker_info
 
-            >>> class RangeIterableDataset(IterableDataset):
+            >>> class RangeIterableDataset(IterableDataset): # type: ignore[type-arg]
             ...     def __init__(self, start, end):
             ...         self.start = start
             ...         self.end = end
@@ -228,7 +228,7 @@ class IterableDataset(Dataset[_T]):
             >>> def worker_init_fn(worker_id):
             ...     worker_info = get_worker_info()
             ...
-            ...     dataset = worker_info.dataset
+            ...     dataset: RangeIterableDataset = worker_info.dataset # type: ignore[assignment]
             ...     start = dataset.start
             ...     end = dataset.end
             ...     num_per_worker = int(
@@ -550,14 +550,14 @@ def random_split(
             2
 
             >>> # output of the first subset
-            >>> for idx, v in enumerate(a_list[0]):  # type: ignore[arg-type, var-annotated]
+            >>> for idx, v in enumerate(a_list[0]):
             ...     print(idx, v) # doctest: +SKIP("The output depends on the environment.")
             0 7
             1 6
             2 5
 
             >>> # output of the second subset
-            >>> for idx, v in enumerate(a_list[1]):  # type: ignore[arg-type, var-annotated]
+            >>> for idx, v in enumerate(a_list[1]):
             ...     print(idx, v) # doctest: +SKIP("The output depends on the environment.")
             0 1
             1 9
diff --git a/python/paddle/nn/decode.py b/python/paddle/nn/decode.py
@@ -29,7 +29,6 @@
     from collections.abc import Callable
 
     from paddle import Tensor
-    from paddle._typing import NestedStructure
     from paddle.nn import Embedding, Layer, RNNCellBase
 
 
@@ -1028,7 +1027,7 @@ def dynamic_decode(
     is_test: bool = ...,
     return_length: Literal[False] = ...,
     **kwargs: Any,
-) -> tuple[NestedStructure[Tensor], NestedStructure[Tensor]]: ...
+) -> tuple[Tensor, BeamSearchDecoder.StateWrapper]: ...
 
 
 @overload
@@ -1041,7 +1040,7 @@ def dynamic_decode(
     is_test: bool = ...,
     return_length: Literal[True] = ...,
     **kwargs: Any,
-) -> tuple[NestedStructure[Tensor], NestedStructure[Tensor], Tensor]: ...
+) -> tuple[Tensor, BeamSearchDecoder.StateWrapper, Tensor]: ...
 
 
 @overload
@@ -1055,8 +1054,8 @@ def dynamic_decode(
     return_length: bool = ...,
     **kwargs: Any,
 ) -> (
-    tuple[NestedStructure[Tensor], NestedStructure[Tensor]]
-    | tuple[NestedStructure[Tensor], NestedStructure[Tensor], Tensor]
+    tuple[Tensor, BeamSearchDecoder.StateWrapper]
+    | tuple[Tensor, BeamSearchDecoder.StateWrapper, Tensor]
 ): ...
 
 
diff --git a/python/paddle/vision/image.py b/python/paddle/vision/image.py
@@ -75,7 +75,7 @@ def set_image_backend(backend: _ImageBackend) -> None:
 
             >>> pil_data_folder = DatasetFolder(temp_dir)
 
-            >>> for items in pil_data_folder:  # type: ignore
+            >>> for items in pil_data_folder:
             ...     break
 
             >>> print(type(items[0]))
@@ -86,7 +86,7 @@ def set_image_backend(backend: _ImageBackend) -> None:
 
             >>> cv2_data_folder = DatasetFolder(temp_dir)
 
-            >>> for items in cv2_data_folder:  # type: ignore
+            >>> for items in cv2_data_folder:
             ...     break
 
             >>> print(type(items[0]))
diff --git a/python/paddle/vision/transforms/transforms.py b/python/paddle/vision/transforms/transforms.py
@@ -457,11 +457,11 @@ class Resize(BaseTransform[_InputT, _RetT]):
 
             >>> fake_img = Image.fromarray((np.random.rand(256, 300, 3) * 255.).astype(np.uint8))
             >>> transform = Resize(size=224)
-            >>> converted_img = transform(fake_img)
+            >>> converted_img = transform(fake_img) # type: ignore[call-overload]
             >>> print(converted_img.size)
             (262, 224)
             >>> transform = Resize(size=(200,150))
-            >>> converted_img = transform(fake_img)
+            >>> converted_img = transform(fake_img) # type: ignore[call-overload]
             >>> print(converted_img.size)
             (150, 200)
     """
@@ -1039,10 +1039,10 @@ class BrightnessTransform(BaseTransform[_InputT, _RetT]):
 
             >>> transform = BrightnessTransform(0.4)
             >>> fake_img = Image.fromarray((np.random.rand(224, 224, 3) * 255.).astype(np.uint8))
-            >>> print(fake_img.load()[1,1])
+            >>> print(fake_img.load()[1,1]) # type: ignore[index]
             (60, 169, 34)
             >>> # doctest: +SKIP('random sample in Brightness function')
-            >>> fake_img = transform(fake_img)
+            >>> fake_img = transform(fake_img) # type: ignore[call-overload]
             >>> print(fake_img.load()[1,1])
             (68, 192, 38)
 
