/**
 * 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 "aclnn_logdet.h"

#include "../../../add/op_api/add.h"
#include "../../../log/op_api/log.h"
#include "../../../slogdet/op_host/op_api/log_matrix_determinant.h"
#include "aclnn_kernels/cast.h"
#include "aclnn_kernels/contiguous.h"
#include "aclnn_kernels/reshape.h"

#include "op_api/op_api_def.h"
#include "aclnn_kernels/common/op_error_check.h"
#include "opdev/common_types.h"
#include "opdev/data_type_utils.h"
#include "opdev/format_utils.h"
#include "opdev/op_dfx.h"
#include "opdev/op_executor.h"
#include "opdev/op_log.h"
#include "opdev/shape_utils.h"
#include "opdev/tensor_view_utils.h"

using namespace op;
// 根据API定义，需要列出所能支持的所有dtype
static const std::initializer_list<op::DataType> DTYPE_SUPPORT_LIST = {
    op::DataType::DT_FLOAT, op::DataType::DT_DOUBLE, op::DataType::DT_COMPLEX64, op::DataType::DT_COMPLEX128};


static inline bool CheckNotNull(const aclTensor *self, const aclTensor *out) {
  OP_CHECK_NULL(self, return false);
  OP_CHECK_NULL(out, return false);
  return true;
}


static bool CheckDtypeValid(const aclTensor *self, const aclTensor *out) {
  OP_CHECK_DTYPE_NOT_SUPPORT(self, DTYPE_SUPPORT_LIST, return false);
  OP_CHECK_DTYPE_NOT_SUPPORT(out, DTYPE_SUPPORT_LIST, return false);

  // 当输入数据类型是Complex时, 输出数据类型也必须是Complex
  if (IsComplexType(self->GetDataType()) && !IsComplexType(out->GetDataType())) {
    OP_LOGE(ACLNN_ERR_PARAM_INVALID,
            "The self's dtype is %s, out's dtype should also be complex type but got signOut with dtype %s.",
            op::ToString(self->GetDataType()).GetString(), op::ToString(out->GetDataType()).GetString());
    return false;
  }

  return true;
}

static bool CheckShape(const aclTensor *self, const aclTensor *out) {
  // self >= 2维
  auto dim = self->GetViewShape().GetDimNum();
  OP_CHECK_MIN_DIM(self, 2, return false);
  // self是一组方阵，最后2维相等
  auto mDim = self->GetViewShape().GetDim(dim - 2);
  auto nDim = self->GetViewShape().GetDim(dim - 1);
  OP_CHECK(mDim == nDim,
           OP_LOGE(ACLNN_ERR_PARAM_INVALID,
                   "The last two dimensions of self must be equal, but they are %ld by %ld matrices.", mDim, nDim),
           return false);

  // out是self去掉最后2维
  auto selfBatchShapeVec = ToShapeVector(self->GetViewShape());
  selfBatchShapeVec.pop_back();
  selfBatchShapeVec.pop_back();
  op::Shape selfBatchShape;
  ToShape(selfBatchShapeVec, selfBatchShape);
  OP_CHECK(out->GetViewShape() == selfBatchShape,
           OP_LOGE(ACLNN_ERR_PARAM_INVALID, "expect shape of out is %s, but got %s.",
                   op::ToString(selfBatchShape).GetString(), op::ToString(out->GetViewShape()).GetString()),
           return false);

  return true;
}

static aclnnStatus CheckParams(const aclTensor *self, const aclTensor *out) {
  // 1. 检查参数是否为空指针
  CHECK_RET(CheckNotNull(self, out), ACLNN_ERR_PARAM_NULLPTR);

  // 2. 检查输入的数据类型是否在API支持的数据类型范围之内，需要根据api定义校验
  CHECK_RET(CheckDtypeValid(self, out), ACLNN_ERR_PARAM_INVALID);

  // 3. 检查shape是否满足条件
  CHECK_RET(CheckShape(self, out), ACLNN_ERR_PARAM_INVALID);

  return ACLNN_SUCCESS;
}

static aclnnStatus ReshapeDim(const aclTensor *self, op::Shape &selfBatchShape, const aclTensor *&selfReshapeOut,
                              aclOpExecutor *executor) {
  auto selfOriginalShape = self->GetViewShape();
  auto dim = self->GetViewShape().GetDimNum();
  auto lastDim = self->GetViewShape().GetDim(dim - 1);
  auto newDim = self->Size() / (lastDim * lastDim);
  op::Shape selfNewShape = {newDim, lastDim, lastDim};
  selfReshapeOut = l0op::Reshape(self, selfNewShape, executor);
  CHECK_RET(selfReshapeOut != nullptr, ACLNN_ERR_INNER_NULLPTR);

  auto shapeVec = ToShapeVector(selfOriginalShape);
  shapeVec.pop_back();
  shapeVec.pop_back();
  ToShape(shapeVec, selfBatchShape);
  return ACLNN_SUCCESS;
}

static void CheckFormat(const aclTensor* self) {
    ge::Format selfStorageFormat = self->GetStorageFormat();
    if (selfStorageFormat == ge::Format::FORMAT_FRACTAL_NZ) {
        OP_LOGW("aclnnLogdet doesn't support format NZ.");
    }
}

#ifdef __cplusplus
extern "C" {
#endif

aclnnStatus aclnnLogdetGetWorkspaceSize(const aclTensor *self, aclTensor *out, uint64_t *workspaceSize,
                                         aclOpExecutor **executor) {
  L2_DFX_PHASE_1(aclnnLogdet, DFX_IN(self), DFX_OUT(out));
  // 参数检查
  auto ret = CheckParams(self, out);
  CHECK_RET(ret == ACLNN_SUCCESS, ret);

  // 创建OpExecutor
  auto uniqueExecutor = CREATE_EXECUTOR();
  CHECK_RET(uniqueExecutor.get() != nullptr, ACLNN_ERR_INNER_CREATE_EXECUTOR);

  if (self->IsEmpty()) {
    *workspaceSize = 0;
    uniqueExecutor.ReleaseTo(executor);
    return ACLNN_SUCCESS;
  }
  
  CheckFormat(self);

  auto selfContiguous = l0op::Contiguous(self, uniqueExecutor.get());
  CHECK_RET(selfContiguous != nullptr, ACLNN_ERR_INNER_NULLPTR);

  const aclTensor *selfReshapeOut = nullptr;
  auto selfOriginalShape = self->GetViewShape();
  auto dim = selfOriginalShape.GetDimNum();
  op::Shape selfBatchShape;
  // 8维以上需要reshape到3维
  if (dim > MAX_SUPPORT_DIMS_NUMS) {
    ret = ReshapeDim(selfContiguous, selfBatchShape, selfReshapeOut, uniqueExecutor.get());
    CHECK_RET(ret == ACLNN_SUCCESS, ret);
  } else {
    selfReshapeOut = selfContiguous;
  }

  auto logMatrixDeterminantOut = l0op::LogMatrixDeterminant(selfReshapeOut, uniqueExecutor.get());
  auto signValue = std::get<0>(logMatrixDeterminantOut);
  const float LOG_BASE = -1.0f;
  const float LOG_SCALE = 1.0f;
  const float LOG_SHIFT = 0.0f;
  auto offset = l0op::Log(signValue, LOG_BASE, LOG_SCALE, LOG_SHIFT, uniqueExecutor.get());
  CHECK_RET(offset != nullptr, ACLNN_ERR_INNER_NULLPTR);

  auto logValue = std::get<1>(logMatrixDeterminantOut);
  CHECK_RET(logValue != nullptr, ACLNN_ERR_INNER_NULLPTR);
  logValue = l0op::Add(logValue, offset, uniqueExecutor.get());
  CHECK_RET(logValue != nullptr, ACLNN_ERR_INNER_NULLPTR);

  // 8维以上需要reshape
  const aclTensor *logReshapeOut = nullptr;
  if (dim > MAX_SUPPORT_DIMS_NUMS) {
    logReshapeOut = l0op::Reshape(logValue, selfBatchShape, uniqueExecutor.get());
    CHECK_RET(logReshapeOut != nullptr, ACLNN_ERR_INNER_NULLPTR);
  } else {
    logReshapeOut = logValue;
  }
  auto logCastOut = l0op::Cast(logReshapeOut, out->GetDataType(), uniqueExecutor.get());
  CHECK_RET(logCastOut != nullptr, ACLNN_ERR_INNER_NULLPTR);

  // 固定写法，将计算结果拷贝到输出out上，out可能是非连续的tensor
  auto logCopyResult = l0op::ViewCopy(logCastOut, out, uniqueExecutor.get());
  CHECK_RET(logCopyResult != nullptr, ACLNN_ERR_INNER_NULLPTR);
  // 固定写法，获取计算过程中需要使用的workspace大小
  *workspaceSize = uniqueExecutor->GetWorkspaceSize();
  uniqueExecutor.ReleaseTo(executor);  // 需要把 uniqueExecutor持有executor转移给executor

  return ACLNN_SUCCESS;
}

aclnnStatus aclnnLogdet(void *workspace, uint64_t workspaceSize, aclOpExecutor *executor, aclrtStream stream) {
  L2_DFX_PHASE_2(aclnnLogdet);
  // 固定写法，调用框架能力，完成计算
  return CommonOpExecutorRun(workspace, workspaceSize, executor, stream);
}

#ifdef __cplusplus
}
#endif