/**
* Copyright (c) 2025 Huawei Technologies Co., Ltd.
* This program is free software, you can redistribute it and/or modify it under the terms and conditions of
* CANN Open Software License Agreement Version 2.0 (the "License").
* Please refer to the License for details. You may not use this file except in compliance with the License.
* THIS SOFTWARE IS PROVIDED ON AN "AS IS" BASIS, WITHOUT WARRANTIES OF ANY KIND, EITHER EXPRESS OR IMPLIED,
* INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT, MERCHANTABILITY, OR FITNESS FOR A PARTICULAR PURPOSE.
* See LICENSE in the root of the software repository for the full text of the License.
*/
#include "infershape_case_executor.h"
#include <gtest/gtest.h>
#include "base/registry/op_impl_space_registry_v2.h"
#include "exe_graph/runtime/kernel_run_context.h"
#define DO_INFERSHAPE(infershapeContextPara) \
auto contextFaker = gert::InferShapeContextFaker(); \
/* 1. input/output information */ \
size_t inputNum = infershapeContextPara.inputTensorDesc_.size(); \
size_t outputNum = infershapeContextPara.outputTensorDesc_.size(); \
if (infershapeContextPara.inputInstanceNum_.size() != 0 || infershapeContextPara.outputInstanceNum_.size() != 0) { \
contextFaker.IrInstanceNum(infershapeContextPara.inputInstanceNum_, infershapeContextPara.outputInstanceNum_); \
} else { \
contextFaker.NodeIoNum(inputNum, outputNum); \
} \
std::vector<gert::Tensor*> inputTensors = {}; \
std::vector<std::unique_ptr<gert::Tensor>> inputTensorsKeepAlive = {}; \
for (size_t index = 0; index < inputNum; index++) { \
std::unique_ptr<gert::Tensor> curTensor = std::make_unique<gert::Tensor>( \
infershapeContextPara.inputTensorDesc_[index].shape_, \
gert::StorageFormat(infershapeContextPara.inputTensorDesc_[index].format_, \
infershapeContextPara.inputTensorDesc_[index].format_, gert::ExpandDimsType()), \
gert::TensorPlacement::kOnHost, infershapeContextPara.inputTensorDesc_[index].dtype_, \
infershapeContextPara.inputTensorDesc_[index].isConst_ ? \
infershapeContextPara.inputTensorDesc_[index].constValue_ : \
nullptr); \
inputTensors.push_back(curTensor.get()); \
inputTensorsKeepAlive.push_back(std::move(curTensor)); \
} \
for (size_t index = 0; index < outputNum; index++) { \
contextFaker.NodeOutputTd(index, infershapeContextPara.outputTensorDesc_[index].dtype_, \
infershapeContextPara.outputTensorDesc_[index].format_, \
infershapeContextPara.outputTensorDesc_[index].format_); \
} \
contextFaker.InputTensors(inputTensors); \
for (auto& attrInfo : infershapeContextPara.attrs_) { \
switch (attrInfo.attr_.type_) { \
case Ops::Math::AnyValue::ValueType::VT_BOOL: { \
contextFaker.Attr(attrInfo.attrName_, *reinterpret_cast<bool*>(attrInfo.attr_.valuePtr_.get())); \
break; \
} \
case Ops::Math::AnyValue::ValueType::VT_INT: { \
contextFaker.Attr(attrInfo.attrName_, *reinterpret_cast<int64_t*>(attrInfo.attr_.valuePtr_.get())); \
break; \
} \
case Ops::Math::AnyValue::ValueType::VT_FLOAT: { \
contextFaker.Attr(attrInfo.attrName_, *reinterpret_cast<float*>(attrInfo.attr_.valuePtr_.get())); \
break; \
} \
case Ops::Math::AnyValue::ValueType::VT_STRING: { \
contextFaker.Attr( \
attrInfo.attrName_, \
ge::AscendString(reinterpret_cast<std::string*>(attrInfo.attr_.valuePtr_.get())->c_str())); \
break; \
} \
case Ops::Math::AnyValue::ValueType::VT_LIST_BOOL: { \
contextFaker.Attr(attrInfo.attrName_, \
*reinterpret_cast<std::vector<bool>*>(attrInfo.attr_.valuePtr_.get())); \
break; \
} \
case Ops::Math::AnyValue::ValueType::VT_LIST_INT: { \
contextFaker.Attr(attrInfo.attrName_, \
*reinterpret_cast<std::vector<int64_t>*>(attrInfo.attr_.valuePtr_.get())); \
break; \
} \
case Ops::Math::AnyValue::ValueType::VT_LIST_LIST_INT: { \
contextFaker.Attr(attrInfo.attrName_, *reinterpret_cast<std::vector<std::vector<int64_t>>*>( \
attrInfo.attr_.valuePtr_.get())); \
break; \
} \
case Ops::Math::AnyValue::ValueType::VT_LIST_FLOAT: { \
contextFaker.Attr(attrInfo.attrName_, \
*reinterpret_cast<std::vector<float>*>(attrInfo.attr_.valuePtr_.get())); \
break; \
} \
default: \
std::cout << "[ERROR]" << __FILE__ << ":" << __LINE__ << "The ValueType " << attrInfo.attr_.type_ \
<< "is not supported!" << std::endl; \
} \
} \
auto contextHolder = contextFaker.SetOpType(infershapeContextPara.opName_.c_str()).Build(); \
/* Set null input tensor pointers to nullptr in the context */ \
for (auto nullIdx : infershapeContextPara.nullInputIndices_) { \
/* InferShapeContext has standard layout (POD), and its address equals KernelRunContext address */ \
auto* kernelCtx = reinterpret_cast<KernelRunContext*>(contextHolder.GetContext()); \
if (nullIdx < kernelCtx->input_size) { \
kernelCtx->values[nullIdx]->data.pointer = nullptr; \
} \
} \
/* 2. get infershape func */ \
auto spaceRegistry = gert::DefaultOpImplSpaceRegistryV2::GetInstance().GetSpaceRegistry(); \
auto infershapeFunc = spaceRegistry->GetOpImpl(infershapeContextPara.opName_.c_str())->infer_shape; \
/* 3. check infershape func */ \
auto infershapeRet = infershapeFunc(contextHolder.GetContext());
static std::vector<int64_t> ToVector(const gert::Shape& shape)
{
size_t shapeSize = shape.GetDimNum();
std::vector<int64_t> shapeVec(shapeSize, 0);
for (size_t i = 0; i < shapeSize; i++) {
shapeVec[i] = shape.GetDim(i);
}
return shapeVec;
}
void ExecuteTestCase(gert::InfershapeContextPara& infershapeContextPara, ge::graphStatus expectResult,
const std::vector<std::vector<int64_t>>& expectOutputShape)
{
DO_INFERSHAPE(infershapeContextPara);
// check infershape func
EXPECT_EQ(infershapeRet, expectResult);
if (expectResult == ge::GRAPH_FAILED) {
return;
}
// check output shape
for (int i = 0; i < expectOutputShape.size(); i++) {
EXPECT_EQ(ToVector(*contextHolder.GetContext()->GetOutputShape(i)), expectOutputShape[i]);
}
}