feat(menta): support all annotation

src/mentabotix/modules/menta.py

@@ -15,6 +15,7 @@
     Dict,
     Any,
     Set,
+    Type,
 )
 
 from .exceptions import BadSignatureError, RequirementError
@@ -171,7 +172,7 @@ def construct_inlined_function(
         usages: List[SamplerUsage],
         judging_source: List[SourceCode] | SourceCode,
         extra_context: Dict[str, Any] = None,
-        return_type_varname: Optional[str] = None,
+        return_type: Optional[Type] = None,
         return_raw: bool = False,
         function_name: str = "_func",
     ) -> Callable[[], Any] | Tuple[str, Dict[str, Any]]:
@@ -194,14 +195,6 @@ def construct_inlined_function(
             RequirementError: 如果返回类型变量名在extra_context中未定义，或者判断源中缺少必要的占位符。
             RuntimeError: 如果遇到了不支持的采样器类型。
         """
-        if (
-            return_type_varname
-            and return_type_varname not in extra_context
-            and return_type_varname not in BUILTIN_TYPES
-        ):
-            raise RequirementError(
-                f'The return_type_varname "{return_type_varname}" is not defined in extra_context: {extra_context} and is not in builtin types: {BUILTIN_TYPES}'
-            )
 
         # 将judging_source统一处理为字符串格式
         judging_source: str = judging_source if isinstance(judging_source, str) else "\n ".join(judging_source)
@@ -265,17 +258,22 @@ def construct_inlined_function(
         )
         _logger.debug(f"Created temp_var_source: {temp_var_source}")
 
-        # 构建完整的函数源码并编译执行
-        function_head: str = (
-            f"def {function_name}() -> {return_type_varname}:" if return_type_varname else f"def {function_name}():"
-        )
+        if return_type:
+
+            return_type_varname = f"{function_name}_return_type"
+            compile_context.update({return_type_varname: return_type})
+            # 构建完整的函数源码并编译执行
+            function_head: str = f"def {function_name}() -> {return_type_varname}:"
+        else:
+            function_head: str = f"def {function_name}():"
         func_source = f"{function_head}\n" f" {temp_var_source}\n" f" {judging_source}\n" f" return {RET_IDENTIFIER}"
         _logger.debug(f"Created func_source: {func_source}")
         _logger.debug("Compiling func_source")
 
         # 更新执行环境中的采样器和额外上下文信息
 
         compile_context.update(extra_context) if extra_context else None
+
         if return_raw:
             return func_source, compile_context
         exec(func_source, compile_context)  # exec the source with the context

tests/test_menta.py

@@ -268,6 +268,24 @@ def test_function_name(self):
         self.assertEqual(func.__name__, function_name)
         print(func())
 
+    def test_func_ret_type(self):
+        from inspect import signature
+
+        sages = [
+            SamplerUsage(used_sampler_index=0, required_data_indexes=[0, 2]),
+            SamplerUsage(used_sampler_index=1, required_data_indexes=[5]),
+            SamplerUsage(used_sampler_index=2, required_data_indexes=[0, 1, 2]),
+        ]
+        function_name = "func_a"
+        func = self.menta.construct_inlined_function(
+            usages=sages,
+            judging_source="ret=s0,s1,s2,s3,s4+s5",
+            return_raw=False,
+            return_type=int,
+            function_name=function_name,
+        )
+        self.assertEqual(signature(func).return_annotation, int)
+
     def tearDown(self):
         # 清理可能的副作用
         pass