【Hackathon 5th No.42】 为Paddle代码转换工具新增API转换规则 (第1组 编号1-20)

paconvert/api_mapping.json

@@ -1473,7 +1473,10 @@
   },
   "torch.Tensor.is_inference": {},
   "torch.Tensor.is_meta": {},
-  "torch.Tensor.is_pinned": {},
+  "torch.Tensor.is_pinned": {
+    "Matcher": "Is_PinnedMatcher",
+    "min_input_args": 0
+  },
   "torch.Tensor.is_quantized": {},
   "torch.Tensor.is_set_to": {},
   "torch.Tensor.is_shared": {},
@@ -2321,7 +2324,8 @@
   },
   "torch.Tensor.pin_memory": {
     "Matcher": "GenericMatcher",
-    "paddle_api": "paddle.Tensor.pin_memory"
+    "paddle_api": "paddle.Tensor.pin_memory",
+    "min_input_args": 0
   },
   "torch.Tensor.pinverse": {
     "Matcher": "TensorFunc2PaddleFunc",
@@ -2395,7 +2399,25 @@
     ]
   },
   "torch.Tensor.qscheme": {},
-  "torch.Tensor.quantile": {},
+  "torch.Tensor.quantile": {
+    "Matcher": "GenericMatcher",
+    "paddle_api": "paddle.Tensor.quantile",
+    "min_input_args": 1,
+    "args_list": [
+      "q",
+      "dim",
+      "keepdim",
+      "*",
+      "interpolation",
+      "out"
+    ],
+    "kwargs_change": {
+      "dim": "axis"
+    },
+    "unsupport_args": [
+      "interpolation "
+    ]
+  },
   "torch.Tensor.rad2deg": {
     "Matcher": "UnchangeMatcher"
   },
@@ -2937,7 +2959,10 @@
     "paddle_api": "paddle.Tensor.tanh_"
   },
   "torch.Tensor.tensor_split": {},
-  "torch.Tensor.tile": {},
+  "torch.Tensor.tile": {
+    "Matcher": "TensorTileMatcher",
+    "paddle_api": "paddle.Tensor.tile"
+  },
   "torch.Tensor.to": {
     "Matcher": "TensorToMatcher"
   },
@@ -2946,7 +2971,14 @@
     "paddle_api": "paddle.Tensor.to_dense"
   },
   "torch.Tensor.to_mkldnn": {},
-  "torch.Tensor.to_sparse": {},
+  "torch.Tensor.to_sparse": {
+    "Matcher": "GenericMatcher",
+    "min_input_args": 0,
+    "paddle_api": "paddle.Tensor.to_sparse_coo",
+    "args_list": [
+      "sparse_dim"
+    ]
+  },
   "torch.Tensor.tolist": {
     "Matcher": "UnchangeMatcher"
   },
@@ -7422,6 +7454,26 @@
       "dim"
     ]
   },
+  "torch.nanquantile": {
+    "Matcher": "GenericMatcher",
+    "paddle_api": "paddle.nanquantile",
+    "args_list": [
+      "input",
+      "q",
+      "dim",
+      "keepdim",
+      "*",
+      "interpolation",
+      "out"
+    ],
+    "kwargs_change": {
+      "input": "x",
+      "dim": "axis"
+    },
+    "unsupport_args": [
+      "interpolation "
+    ]
+  },
   "torch.nansum": {
     "Matcher": "GenericMatcher",
     "paddle_api": "paddle.nansum",

paconvert/api_matcher.py

@@ -1121,6 +1121,29 @@ def generate_code(self, kwargs):
         return code
 
 
+class TensorTileMatcher(BaseMatcher):
+    def get_paddle_class_nodes(self, func, args, kwargs):
+        self.parse_func(func)
+        kwargs = self.parse_kwargs(kwargs)
+        if kwargs is None:
+            return None
+
+        if "dims" in kwargs:
+            kwargs = {"repeat_times": kwargs.pop("dims")}
+        else:
+            if len(args) > 1 or (len(args) == 1 and isinstance(args[0], ast.Constant)):
+                perm = self.parse_args(args)
+            elif isinstance(args[0], ast.Starred):
+                perm = astor.to_source(args[0].value).strip("\n")
+            else:
+                perm = self.parse_args(args)[0]
+
+            kwargs = {"repeat_times": str(perm).replace("'", "")}
+
+        code = "{}.tile({})".format(self.paddleClass, self.kwargs_to_str(kwargs))
+        return ast.parse(code).body
+
+
 class TensorNew_Matcher(BaseMatcher):
     def get_paddle_class_nodes(self, func, args, kwargs):
         self.parse_func(func)
@@ -4151,3 +4174,11 @@ def generate_code(self, kwargs):
         return ast.parse(
             "paddle.utils.cpp_extension.setup({})".format(self.kwargs_to_str(kwargs))
         )
+
+
+class Is_PinnedMatcher(BaseMatcher):
+    def generate_code(self, kwargs):
+
+        code = f"'pinned' in str({self.paddleClass}.place)"
+
+        return code

tests/test_Tensor_is_pinned.py

@@ -0,0 +1,34 @@
+# Copyright (c) 2023 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+#
+
+import textwrap
+
+from apibase import APIBase
+
+obj = APIBase("torch.Tensor.is_pinned")
+
+
+def test_case_1():
+    pytorch_code = textwrap.dedent(
+        """
+        import torch
+        if torch.cuda.is_available():
+            x = torch.randn(4,4).pin_memory()
+        else:
+            x = torch.randn(4,4)
+        result = x.is_pinned()
+        """
+    )
+    obj.run(pytorch_code, ["result"])

tests/test_Tensor_quantile.py

@@ -0,0 +1,132 @@
+# Copyright (c) 2023 PaddlePaddle Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+#
+
+import textwrap
+
+from apibase import APIBase
+
+obj = APIBase("torch.Tensor.quantile")
+
+
+def test_case_1():
+    pytorch_code = textwrap.dedent(
+        """
+        import torch
+        x = torch.tensor([0., 1., 2., 3.],dtype=torch.float64)
+        result = x.quantile(0.6)
+        """
+    )
+    obj.run(pytorch_code, ["result"])
+
+
+def test_case_2():
+    pytorch_code = textwrap.dedent(
+        """
+        import torch
+        x = torch.tensor([0., 1., 2., 3.],dtype=torch.float64)
+        result = x.quantile(q=0.6)
+        """
+    )
+    obj.run(pytorch_code, ["result"])
+
+
+def test_case_3():
+    pytorch_code = textwrap.dedent(
+        """
+        import torch
+        x = torch.tensor([0., 1., 2., 3.],dtype=torch.float64)
+        k = 0.6
+        result = x.quantile(k)
+        """
+    )
+    obj.run(pytorch_code, ["result"])
+
+
+def test_case_4():
+    pytorch_code = textwrap.dedent(
+        """
+        import torch
+        x = torch.tensor([0., 1., 2., 3.],dtype=torch.float64)
+        k = 0.6
+        result = x.quantile(q=k)
+        """
+    )
+    obj.run(pytorch_code, ["result"])
+
+
+def test_case_5():
+    pytorch_code = textwrap.dedent(
+        """
+        import torch
+        x = torch.tensor([0., 1., 2., 3.],dtype=torch.float64)
+        result = x.quantile(0.6, dim=None)
+        """
+    )
+    obj.run(pytorch_code, ["result"])
+
+
+def test_case_6():
+    pytorch_code = textwrap.dedent(
+        """
+        import torch
+        x = torch.tensor([0., 1., 2., 3.],dtype=torch.float64)
+        result = x.quantile(0.6, dim=None, keepdim=False)
+        """
+    )
+    obj.run(pytorch_code, ["result"])
+
+
+def test_case_7():
+    pytorch_code = textwrap.dedent(
+        """
+        import torch
+        x = torch.tensor([0., 1., 2., 3.],dtype=torch.float64)
+        result = x.quantile(q=0.6, dim=None, keepdim=False)
+        """
+    )
+    obj.run(pytorch_code, ["result"])
+
+
+def test_case_8():
+    pytorch_code = textwrap.dedent(
+        """
+        import torch
+        x = torch.tensor([0., 1., 2., 3.],dtype=torch.float64)
+        result = x.quantile(0.6, None, False)
+        """
+    )
+    obj.run(pytorch_code, ["result"])
+
+
+def test_case_9():
+    pytorch_code = textwrap.dedent(
+        """
+        import torch
+        x = torch.tensor([0., 1., 2., 3.],dtype=torch.float64)
+        result = x.quantile(q=0.6, keepdim=False, dim=None)
+        """
+    )
+    obj.run(pytorch_code, ["result"])
+
+
+def test_case_10():
+    pytorch_code = textwrap.dedent(
+        """
+        import torch
+        x = torch.tensor([[ 0.0795, -1.2117,  0.9765], [ 1.1707,  0.6706,  0.4884]],dtype=torch.float64)
+        result = x.quantile(0.6, dim=1, keepdim=True)
+        """
+    )
+    obj.run(pytorch_code, ["result"])