mindspore 2.4.10__cp310-cp310-win_amd64.whl → 2.6.0rc1__cp310-cp310-win_amd64.whl

mindspore/ops_generate/api/functional_map_cpp_generator.py

@@ -0,0 +1,504 @@
+# Copyright 2024 Huawei Technologies Co., Ltd
+#
+# 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.
+# ============================================================================
+"""
+Generates C++ functional map header files for graph mode.
+"""
+
+import os
+
+import common.gen_constants as K
+import common.template as template
+from common.gen_utils import save_file, OrderedSet
+from common.base_generator import BaseGenerator
+from pyboost import pyboost_utils
+
+
+class FunctionalMapCppGenerator(BaseGenerator):
+    """
+    Generates C++ functional map header files for graph mode.
+    """
+
+    def __init__(self):
+        """
+        Initializes the generator with templates for the functional map.
+        """
+        self.FUNCTIONAL_MAP_CC_TEMPLATE = template.FUNCTIONAL_MAP_CC_TEMPLATE
+        self.FUNCTIONAL_MAP_H_TEMPLATE = template.FUNCTIONAL_MAP_H_TEMPLATE
+        self.class_to_method_template = template.Template("{\"${class_name}\", \"${method_name}\"}")
+        self.functional_map_template = template.Template("{\"${func_api_name}\", {${class_to_method_str}}},")
+        self.k_prim_op_template = template.Template("prim::kPrim${camel_op_name}")
+        self.tensor_method_kwonlyargs_map_template = template.Template(
+            "{\"${camel_op_name}\", {${kw_only_args_list}}},")
+        self.tensor_method_varargs_map_template = \
+            template.Template("{\"${op_name}\", ${vararg_index}},")
+        self.deprecated_method_decl_template = template.Template(
+            "auto ${dep_op_name} = std::make_shared<prim::DeprecatedTensorMethod>(\"${dep_op_name}\", \"${op_name}\");")
+        self.functional_method_map_template = template.Template("{\"${op_name}\", {${sort_func_method_list_str}}},")
+
+        self.arg_handler_map = {"to_2d_paddings": ["tuple[int]", "list[int]", "int"],
+                                "dtype_to_type_id": ["int", "type"],
+                                "to_kernel_size": ["tuple[int]", "list[int]", "int"],
+                                "to_strides": ["tuple[int]", "list[int]", "int"],
+                                "str_to_enum": ["str"],
+                                "to_pair": ["tuple[int]", "list[int]", "int", "float"],
+                                "to_dilations": ["tuple[int]", "list[int]", "int"],
+                                "to_output_padding": ["tuple[int]", "list[int]", "int"],
+                                "to_rates": ["tuple[int]", "list[int]", "int"]}
+
+        self.prompt_type_map = {"any": "any",
+                                "int": "int",
+                                "float": "float",
+                                "str": "str",
+                                "bool": "bool",
+                                "number": "number",
+                                "tensor": "Tensor",
+                                "type": "mstype",
+                                "None": "None"}
+
+    def generate(self, work_path, tensor_method_protos_data, mint_func_protos_data, alias_func_mapping):
+        """
+        Generates the functional map header file.
+
+        Args:
+            work_path (str): The directory path to save the generated file.
+            tensor_method_protos_data (dict): A dictionary mapping function API names to their prototype data.
+            mint_func_protos_data (dict): A dictionary mapping mint API names to their prototype data.
+            alias_func_mapping (dict): A dictionary mapping function name to its alias function names.
+
+        Returns:
+            None
+        """
+        ops_inc_head_set = set()
+        dep_method_decl_list = self._get_dep_method_decl_list(tensor_method_protos_data)
+        tensor_method_overload_list, op_inc_1 = self._get_functional_method_map(tensor_method_protos_data,
+                                                                                alias_func_mapping)
+        mint_overload_list, op_inc_2 = self._get_functional_mint_map(mint_func_protos_data, alias_func_mapping)
+        tensor_method_kw_only_args_list = self._get_tensor_method_kwonlyargs_map(tensor_method_protos_data)
+        mint_kw_only_args_list = self._get_mint_kwonlyargs_map(mint_func_protos_data, alias_func_mapping)
+        tensor_varargs_map_list = self._get_tensor_varargs_map_list(tensor_method_protos_data)
+        mint_varargs_map_list = self._get_mint_varargs_map_list(mint_func_protos_data, alias_func_mapping)
+        funcs_sig_map_list = (
+            self._get_func_sigs_list(tensor_method_protos_data, alias_func_mapping, is_tensor_method=True))
+        funcs_mint_sigs_map = (
+            self._get_func_sigs_list(mint_func_protos_data, alias_func_mapping, is_tensor_method=False))
+        merge_op_inc = op_inc_1 | op_inc_2
+        for op_inc in merge_op_inc:
+            ops_inc_head_set.add(template.OP_DEF_INC_HEAD_TEMPLATE.replace(prefix_char=op_inc[0].lower()))
+        functional_map_cc_code = (
+            self.FUNCTIONAL_MAP_CC_TEMPLATE.replace(ops_inc=list(sorted(ops_inc_head_set)),
+                                                    deprecated_method_decl=dep_method_decl_list,
+                                                    tensor_method_map=tensor_method_overload_list,
+                                                    mint_map=mint_overload_list,
+                                                    tensor_method_kwonlyargs_map=tensor_method_kw_only_args_list,
+                                                    mint_kwonlyargs_map=mint_kw_only_args_list,
+                                                    tensor_varargs_map=tensor_varargs_map_list,
+                                                    mint_varargs_map=mint_varargs_map_list,
+                                                    tensor_method_sigs_map=funcs_sig_map_list,
+                                                    mint_sigs_map=funcs_mint_sigs_map))
+        save_path = os.path.join(work_path, K.FUNCTIONAL_OVERLOAD_GEN_PATH)
+        save_file(save_path, "functional_map.cc", functional_map_cc_code)
+        save_file(save_path, "functional_map.h", self.FUNCTIONAL_MAP_H_TEMPLATE.replace())
+
+    def _get_func_sigs_list(self, tensor_method_protos_data, alias_func_mapping, is_tensor_method):
+        """
+        Generates a list of function signatures for each function API name based on the provided prototype data.
+
+        Args: tensor_method_protos_data (dict): A dictionary mapping function API names to their corresponding prototype
+                                    data. Each prototype contains information necessary to generate function signatures.
+              alias_func_mapping (dict): A dictionary mapping function name to its alias function names.
+              is_tensor_method (bool): Whether the prototype data is a tensor method or a mint function.
+
+        Returns: list: A list of function signature strings for each function API, which are generated based on the
+        prototype data.
+        """
+        funcs_list = []
+        for func_api_name, func_protos in tensor_method_protos_data.items():
+            func_signatures = self._generate_func_signatures_str(func_api_name, func_protos, is_tensor_method)
+            funcs_list.append(func_signatures)
+            if func_api_name in alias_func_mapping:
+                for alias_api_name in alias_func_mapping[func_api_name]:
+                    func_signatures = self._generate_func_signatures_str(alias_api_name, func_protos, is_tensor_method)
+                    funcs_list.append(func_signatures)
+
+        return funcs_list
+
+    def _generate_func_signatures_str(self, func_api_name, func_protos, is_tensor_method) -> str:
+        """
+        Generates function signatures as a string from the given prototypes.
+
+        Args:
+            func_api_name (str): The name of the API to generate signatures for.
+            func_protos (list): List of TensorFuncProto objects representing the function prototypes.
+
+        Returns:
+            str: Generated function signatures string.
+        """
+        sig_set = OrderedSet()
+        for tensor_proto in func_protos:
+            sig_set.add(self._generate_single_signature_str(func_api_name, tensor_proto, is_tensor_method))
+        sig_list = list(sig_set)
+        sig_str = '{' + f'\"{func_api_name}\",\n ' + '{'
+        first_sig = True
+        for sig in sig_list:
+            if not first_sig:
+                sig_str += ',\n'
+            first_sig = False
+            sig_str += sig
+        sig_str += '}\n},'
+        return sig_str
+
+    def _is_input_arg(self, arg_name, op_name):
+        res = False
+        if op_name in K.INPUT_NAME_MAP and arg_name == K.INPUT_NAME_MAP[op_name]:
+            res = True
+        elif op_name not in K.INPUT_NAME_MAP and arg_name in K.INPUT_ARGS_NAME:
+            res = True
+        return res
+
+    def _generate_single_signature_str(self, func_api_name, tensor_proto, is_tensor_method) -> str:
+        """
+        Generates a single function signature string for the given operation prototype.
+
+        Args:
+            func_api_name (str): The name of the API to generate signatures for.
+            tensor_proto (OpProto): TensorFuncProto objects representing the function prototypes.
+
+        Returns:
+            str: Generated function signature string.
+        """
+        op_proto = tensor_proto.op_proto
+        op_name = tensor_proto.op_proto.op_class.name
+        args_str = f'"Tensor.{func_api_name}(' if is_tensor_method else f'"{func_api_name}('
+        first_arg = True
+        is_kw_args_init = False
+        arg_valid_types = []
+        for _, arg in enumerate(op_proto.op_args):
+            arg_name = arg.arg_name
+            if is_tensor_method and self._is_input_arg(arg_name, op_name):
+                continue
+
+            arg_valid_types = self._handle_arg_valid_types(arg, arg_name, arg_valid_types, func_api_name)
+            single_arg = f'{arg_name}=<' + ','.join(arg_valid_types) + '>'
+            prefix, is_kw_args_init, first_arg = self._build_prefix(arg_name, first_arg, is_kw_args_init, tensor_proto)
+            args_str += prefix + single_arg
+            arg_valid_types = []
+        return args_str + ')"'
+
+    def _build_prefix(self, arg_name, first_arg, is_kw_args_init, tensor_proto):
+        """
+        Build and return the prefix for the current argument, handling insertion of kw-only args if needed.
+
+        Args:
+            arg_name (str): Name of the current argument.
+            first_arg (bool): Indicates whether this is the first argument.
+            is_kw_args_init (bool): Indicates whether kw-only args prefix has been inserted.
+            tensor_proto: Contains information about kw_only_args.
+
+        Returns:
+            tuple:
+                prefix (str): Generated prefix (possibly including '*, ').
+                is_kw_args_init (bool): Updated kw-only args insertion status.
+                first_arg (bool): Updated first-argument status.
+        """
+        prefix = "" if first_arg else ", "
+        if tensor_proto.kw_only_args and not is_kw_args_init and arg_name == tensor_proto.kw_only_args[0]:
+            prefix += "*, "
+            is_kw_args_init = True
+        if first_arg:
+            first_arg = False
+        return prefix, is_kw_args_init, first_arg
+
+    def _handle_arg_valid_types(self, arg, arg_name, arg_valid_types, func_api_name):
+        """
+        Collect and return valid argument types based on the arg handler, defaults, and casts.
+
+        Args:
+            arg: Argument object containing handler, dtype, and cast info.
+            arg_name (str): Name of the current argument.
+            arg_valid_types (list): Existing valid types to be extended.
+            func_api_name (str): Name of the current API function.
+
+        Returns:
+            list: Sorted list of valid argument types (descending order).
+        """
+        arg_handler = arg.arg_handler
+        if arg_handler != '':
+            if arg_handler in self.arg_handler_map:
+                arg_valid_types.extend(self.arg_handler_map[arg_handler])
+            else:
+                raise ValueError("Generate failed. Check if {} is registered in TensorFuncRegCppGenerator."
+                                 .format(arg_handler))
+        else:
+            arg_valid_types.append(arg.arg_dtype)
+            for cast_type in arg.type_cast:
+                arg_valid_types.append(cast_type)
+
+        if arg.as_init_arg and str(arg.default) == 'None':
+            arg_valid_types.append('None')
+
+        arg_valid_types = self._parse_arg_type_list(func_api_name, arg_name, arg_valid_types)
+
+        return sorted(arg_valid_types, reverse=True)
+
+    def _parse_arg_type_list(self, func_api_name, arg_name, arg_valid_types):
+        """
+        Parses a list of argument types and maps them to generalized types.
+
+        Args:
+            func_api_name (str): The name of the function API for which the argument types are being parsed.
+            arg_name (str): The name of the argument whose valid types are being generalized.
+            arg_valid_types (list): A list of valid argument types that need to be generalized.
+
+        Returns:
+            set: A set of generalized argument types (e.g., 'List', 'Tuple') based on the input types.
+
+        Raises:
+            ValueError: If an unrecognized or invalid type is encountered in the argument types list.
+        """
+        generalized_type_list = set()
+        for arg_type in arg_valid_types:
+            if arg_type in self.prompt_type_map:
+                generalized_type_list.add(self.prompt_type_map[arg_type])
+            elif "list" in arg_type:
+                generalized_type_list.add('List')
+            elif "tuple" in arg_type:
+                generalized_type_list.add('Tuple')
+            else:
+                raise ValueError(f"Invalid type {arg_type} in api: {func_api_name} {arg_name}.")
+        return generalized_type_list
+
+    def _get_dep_method_decl_list(self, func_protos_data):
+        """
+        Extracts and generates declarations for deprecated methods from the provided function prototypes.
+
+        Args:
+            func_protos_data (dict): A dictionary where keys are function API names and values are lists
+                of function prototypes. Each prototype contains an operation name.
+
+        Returns:
+            list: A list of strings, each representing a declaration for a deprecated method.
+        """
+        deprecated_method_decl_list = []
+        for func_api_name, func_protos in func_protos_data.items():
+            for func_proto in func_protos:
+                op_name = func_proto.op_proto.op_name
+                if not op_name.startswith("deprecated"):
+                    continue
+
+                deprecated_method_name = ''.join(word.capitalize() for word in op_name.split('_'))
+                if func_proto.op_proto.op_name[-1] == '_':
+                    deprecated_method_name += '_'
+                deprecated_method_decl_list.append(
+                    self.deprecated_method_decl_template.replace(dep_op_name=deprecated_method_name,
+                                                                 op_name=func_api_name))
+
+        return deprecated_method_decl_list
+
+    def _get_functional_method_map(self, tensor_method_protos_data, alias_func_mapping):
+        """
+            Generates a list of functional method maps from the provided function prototypes and alias mappings.
+
+            Args:
+                tensor_method_protos_data (dict): A dictionary where keys are function API names and values are lists
+                    of function prototypes.
+                alias_func_mapping (dict): A dictionary mapping function API names to their aliases.
+
+            Returns:
+                list: A list of strings, each representing a functional method map.
+        """
+
+        op_inc_set = set()
+
+        def get_sort_func_method_list(func_protos):
+            """
+            Retrieves a sorted list of operator primitives, prioritizing deprecated operators.
+            """
+            func_method_list = []
+            for func_proto in func_protos:
+                k_op_name = pyboost_utils.get_op_name(func_proto.op_proto.op_name, func_proto.op_proto.op_class.name)
+                if k_op_name.startswith("Deprecated"):
+                    func_method_list.append(k_op_name)
+                else:
+                    func_method_list.append(self.k_prim_op_template.replace(camel_op_name=k_op_name))
+                    op_inc_set.add(k_op_name)
+
+            func_method_list.sort(key=lambda x: x.startswith("Deprecated"), reverse=True)
+            return func_method_list
+
+        deprecated_method_decl_list = []
+        for func_api_name, func_protos in tensor_method_protos_data.items():
+            sort_func_method_list = get_sort_func_method_list(func_protos)
+            deprecated_method_decl_list.append(
+                self.functional_method_map_template.replace(op_name=func_api_name,
+                                                            sort_func_method_list_str=sort_func_method_list))
+
+            if func_api_name in alias_func_mapping:
+                for alias in alias_func_mapping[func_api_name]:
+                    deprecated_method_decl_list.append(
+                        self.functional_method_map_template.replace(op_name=alias,
+                                                                    sort_func_method_list_str=sort_func_method_list))
+
+        return deprecated_method_decl_list, op_inc_set
+
+    def _get_functional_mint_map(self, mint_func_protos_data, alias_func_mapping):
+        """
+        mint_func_protos_data (dict): A dictionary mapping mint API names to their prototype data.
+        """
+
+        op_inc_set = set()
+
+        def get_mint_func_list(func_protos):
+            """
+            Retrieves a sorted list of operator primitives, prioritizing deprecated operators.
+            """
+            func_method_list = []
+            for func_proto in func_protos:
+                k_op_name = pyboost_utils.get_op_name(func_proto.op_proto.op_name, func_proto.op_proto.op_class.name)
+                func_method_list.append(self.k_prim_op_template.replace(camel_op_name=k_op_name))
+                op_inc_set.add(k_op_name)
+
+            return func_method_list
+
+        mint_func_decl_list = []
+        for func_api_name, func_protos in mint_func_protos_data.items():
+            mint_func_list = get_mint_func_list(func_protos)
+            mint_func_decl_list.append(
+                self.functional_method_map_template.replace(op_name=func_api_name,
+                                                            sort_func_method_list_str=mint_func_list))
+            if func_api_name in alias_func_mapping:
+                for alias in alias_func_mapping[func_api_name]:
+                    mint_func_decl_list.append(
+                        self.functional_method_map_template.replace(op_name=alias,
+                                                                    sort_func_method_list_str=mint_func_list))
+        return mint_func_decl_list, op_inc_set
+
+    def _get_and_append_single_op_kw_only_args_list(self, func_protos, single_op_kw_only_args_list):
+        """
+        Extracts keyword-only arguments from a list of function prototypes and appends them to a list.
+
+        Args:
+            func_protos (list): A list of function prototypes.
+            single_op_kw_only_args_list (list): The list to append the keyword-only arguments to.
+
+        Returns:
+            None
+        """
+        for func_proto in func_protos:
+            camel_op_name = pyboost_utils.get_op_name(func_proto.op_proto.op_name, func_proto.op_proto.op_class.name)
+            kw_only_args = func_proto.kw_only_args
+            if kw_only_args:
+                kw_only_args_list = ", ".join(f"\"{kw_arg}\"" for kw_arg in kw_only_args)
+                single_op_kw_only_args_list.append(
+                    self.tensor_method_kwonlyargs_map_template.replace(camel_op_name=camel_op_name,
+                                                                       kw_only_args_list=kw_only_args_list)
+                )
+
+    def _get_tensor_method_kwonlyargs_map(self, tensor_method_protos_data):
+        """
+        Generates a list of keyword-only arguments for tensor methods.
+
+        Args:
+            tensor_method_protos_data (dict): A dictionary of tensor method prototype data.
+
+        Returns:
+            list: A list of formatted strings representing the keyword-only arguments.
+        """
+        tensor_method_kw_only_args_list = []
+        for _, func_protos in tensor_method_protos_data.items():
+            self._get_and_append_single_op_kw_only_args_list(func_protos,
+                                                             tensor_method_kw_only_args_list)
+        return tensor_method_kw_only_args_list
+
+    def _get_mint_kwonlyargs_map(self, mint_func_protos_data, alias_func_mapping):
+        """
+        Generates a list of keyword-only arguments for mint functions.
+
+        Args:
+            mint_func_protos_data (dict): A dictionary of mint function prototype data.
+            alias_func_mapping (dict): A dictionary mapping original function names to alias function names.
+
+        Returns:
+            list: A list of formatted strings representing the keyword-only arguments.
+        """
+        mint_kw_only_args_list = []
+        for _, func_protos in mint_func_protos_data.items():
+            self._get_and_append_single_op_kw_only_args_list(func_protos,
+                                                             mint_kw_only_args_list)
+        return mint_kw_only_args_list
+
+    def _get_and_append_single_op_varargs_list(self, func_protos, single_op_varargs_list):
+        """
+        Extracts variable arguments from a list of function prototypes and appends them to a list.
+
+        Args:
+            func_protos (list): A list of function prototypes.
+            single_op_varargs_list (list): The list to append the variable arguments to.
+
+        Returns:
+            None
+        """
+        for func_proto in func_protos:
+            varargs = func_proto.varargs
+            args = func_proto.op_proto.op_args
+            op_name = func_proto.op_proto.op_class.name
+            if varargs:
+                if len(varargs) != 1:
+                    raise ValueError(
+                        f'There must be only one variable argument. But got {len(vararg_index)} in {op_name}')
+                vararg_index = [i for i in range(len(args)) if args[i].arg_name == varargs[0]]
+                if len(vararg_index) != 1:
+                    raise ValueError(
+                        f'The variable arguments list of {op_name} is wrong, please check.')
+                single_op_varargs_list.append(
+                    self.tensor_method_varargs_map_template.replace(op_name=op_name,
+                                                                    vararg_index=vararg_index[0])
+                )
+
+    def _get_tensor_varargs_map_list(self, tensor_method_protos_data):
+        """
+        Generates a list of variable arguments for tensor methods.
+
+        Args:
+            tensor_method_protos_data (dict): A dictionary of tensor method prototype data.
+
+        Returns:
+            list: A list of formatted strings representing the Variable arguments.
+        """
+        tensor_method_varargs_list = []
+        for _, func_protos in tensor_method_protos_data.items():
+            self._get_and_append_single_op_varargs_list(func_protos,
+                                                        tensor_method_varargs_list)
+        return tensor_method_varargs_list
+
+    def _get_mint_varargs_map_list(self, mint_func_protos_data, alias_func_mapping):
+        """
+        Generates a list of variable arguments for mint functions.
+
+        Args:
+            mint_func_protos_data (dict): A dictionary of mint function prototype data.
+            alias_func_mapping (dict): A dictionary mapping original function names to alias function names.
+
+        Returns:
+            list: A list of formatted strings representing the variable arguments.
+        """
+        mint_varargs_list = []
+        for func_api_name, func_protos in mint_func_protos_data.items():
+            self._get_and_append_single_op_varargs_list(func_protos,
+                                                        mint_varargs_list)
+
+            if mint_varargs_list and func_api_name in alias_func_mapping:
+                self._get_and_append_single_op_varargs_list(func_protos,
+                                                            mint_varargs_list)
+        return mint_varargs_list

mindspore/ops_generate/api/functional_overload_py_generator.py

@@ -0,0 +1,112 @@
+# Copyright 2024 Huawei Technologies Co., Ltd
+#
+# 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.
+# ============================================================================
+"""
+Module for generating C++ header files with operator name definitions.
+
+This module defines the `OpsNameHGenerator` class, which produces C++ code to define
+constants for operator names based on given prototypes.
+"""
+
+import os
+
+import common.gen_constants as K
+import common.gen_utils as gen_utils
+import common.template as template
+from common.template import Template
+
+from common.base_generator import BaseGenerator
+
+
+class FunctionalOverloadPyGenerator(BaseGenerator):
+    """
+    Class for generating C++ header files containing operator name constants.
+    """
+
+    def __init__(self):
+        """
+        Initializes the OpsNameHGenerator instance.
+        """
+        self.FUNCTIONAL_OVERLOAD_PY_TEMPLATE = template.FUNCTIONAL_OVERLOAD_PY_TEMPLATE
+
+        self.mint_func_doc_yaml_dir_path = os.path.join(K.WORK_DIR, K.MS_MINT_FUNC_DOC_YAML_PATH)
+        self.import_mint_template = Template("from mindspore._c_expression import _${cpp_func_name}_instance\n")
+        self.mint_def_template = Template(
+            'def ${mint_func_name}(*args, **kwargs):\n'
+            '    r"""\n${docstr}\n    """\n'
+            '    return _${cpp_func_name}_instance(*args, **kwargs)\n\n\n'
+        )
+
+    def generate(self, work_path, mint_func_protos_data, alias_api_mapping):
+        """
+        Generates python code for operator names and saves it to a header file.
+
+        Args:
+            mint_func_protos_data (dict): A dictionary mapping mint API names to their prototype data.
+            function_doc_data (dict): A dictionary mapping function names to their docstring data.
+            alias_api_mapping (dict): A dictionary mapping aliases to their prototype data.
+        """
+        function_doc_data = gen_utils.safe_load_yaml_from_dir(self.mint_func_doc_yaml_dir_path)
+        validate_func_docs(mint_func_protos_data, function_doc_data, alias_api_mapping)
+        import_mint_list, mint_init_list, mint_def_list, add_to_all_list = [], [], [], []
+        for mint_api_name, _ in mint_func_protos_data.items():
+            func_docstr = _format_docstring(function_doc_data[mint_api_name]["description"])
+            import_mint_list.append(self.import_mint_template.replace(cpp_func_name=mint_api_name))
+            mint_def_list.append(self.mint_def_template.replace(mint_func_name=mint_api_name,
+                                                                docstr=func_docstr,
+                                                                cpp_func_name=mint_api_name))
+            add_to_all_list.append(f'"{mint_api_name}",\n')
+            if mint_api_name in alias_api_mapping:
+                for alias_api_name in alias_api_mapping[mint_api_name]:
+                    func_docstr = _format_docstring(function_doc_data[alias_api_name]["description"])
+                    mint_def_list.append(self.mint_def_template.replace(mint_func_name=alias_api_name,
+                                                                        docstr=func_docstr,
+                                                                        cpp_func_name=mint_api_name))
+                    add_to_all_list.append(f'"{alias_api_name}",\n')
+
+        func_overload_py_file = self.FUNCTIONAL_OVERLOAD_PY_TEMPLATE.replace(import_mint_list=import_mint_list,
+                                                                             mint_init_list=mint_init_list,
+                                                                             mint_def_list=mint_def_list,
+                                                                             add_to_all_list=add_to_all_list)
+        save_path = os.path.join(work_path, K.MS_MINT_FUNC_OVERLOAD_PATH)
+        file_name = "functional_overload.py"
+        gen_utils.save_file(save_path, file_name, func_overload_py_file)
+
+
+def _format_docstring(docstring, indent_size=4):
+    if docstring is None:
+        return None
+
+    lines = docstring.split('\n')
+    # Add 4 spaces to each line except first line
+    formatted_lines = ([' ' * indent_size + lines[0]] +
+                       [' ' * indent_size + line if line.strip() else line for line in lines[1:]])
+    return '\n'.join(formatted_lines)
+
+
+def validate_func_docs(mint_func_protos_data, function_doc_data, alias_api_mapping):
+    """
+    Ensure that the generated API includes corresponding docstrings; otherwise, raise an error to prompt the developer.
+    """
+    mint_api_names = set(mint_func_protos_data.keys())
+    mint_doc_names = set(function_doc_data.keys())
+    all_api_names = set()
+    for mint_api_name in mint_api_names:
+        if mint_api_name in alias_api_mapping:
+            all_api_names = all_api_names.union(set(alias_api_mapping[mint_api_name]))
+    all_api_names = all_api_names.union(mint_api_names)
+    missing_docs = mint_doc_names - all_api_names
+    if missing_docs:
+        raise KeyError(f"Missing valid API references for the following doc names: {missing_docs}, "
+                       f"please check if their doc.yaml files are defined in mindspore/ops/api_def/function_doc.")