/**
 * 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.
 */

/*!
 * \file aclnn_nonzero.cpp
 * \brief
 */
#include "aclnn_nonzero.h"
#include "aclnn_kernels/cast.h"
#include "aclnn_kernels/contiguous.h"
#include "nonzero.h"
#include "aclnn_kernels/common/op_error_check.h"
#include "aclnn/aclnn_base.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"
#include "opdev/platform.h"
#include "op_api/aclnn_util.h"

using namespace op;
#ifdef __cplusplus
extern "C" {
#endif

constexpr size_t MAX_DIM_LEN = 8;

// 根据API定义，需要列出所能支持的所有dtype
static const std::initializer_list<op::DataType> ASCEND910_DTYPE_SUPPORT_LIST = {
    op::DataType::DT_FLOAT,   op::DataType::DT_INT32,  op::DataType::DT_INT64,
    op::DataType::DT_FLOAT16, op::DataType::DT_INT16,  op::DataType::DT_INT8,
    op::DataType::DT_UINT8,   op::DataType::DT_DOUBLE, op::DataType::DT_BOOL};

static const std::initializer_list<op::DataType> ASCEND910B_DTYPE_SUPPORT_LIST = {
    op::DataType::DT_FLOAT, op::DataType::DT_INT32, op::DataType::DT_INT64, op::DataType::DT_FLOAT16,
    op::DataType::DT_INT16, op::DataType::DT_INT8,  op::DataType::DT_UINT8, op::DataType::DT_DOUBLE,
    op::DataType::DT_BOOL,  op::DataType::DT_BF16};

static const std::initializer_list<op::DataType> REGBASE_DTYPE_SUPPORT_LIST = {
    op::DataType::DT_FLOAT, op::DataType::DT_INT32, op::DataType::DT_INT64,  op::DataType::DT_FLOAT16,
    op::DataType::DT_INT16, op::DataType::DT_INT8,  op::DataType::DT_UINT8,  op::DataType::DT_DOUBLE,
    op::DataType::DT_BOOL,  op::DataType::DT_BF16,  op::DataType::DT_UINT16, op::DataType::DT_UINT32,
    op::DataType::DT_UINT64};

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

static inline const std::initializer_list<op::DataType>& GetDtypeSupportList()
{
    auto curArch = GetCurrentPlatformInfo().GetCurNpuArch();
    if (Ops::NN::AclnnUtil::IsRegbase(curArch)) {
        return REGBASE_DTYPE_SUPPORT_LIST;
    }
    if (curArch == NpuArch::DAV_2201 || Ops::NN::AclnnUtil::IsRegbase(curArch)) {
        return ASCEND910B_DTYPE_SUPPORT_LIST;
    } else {
        return ASCEND910_DTYPE_SUPPORT_LIST;
    }
}

inline static bool CheckDtypeValid(const aclTensor* self, const aclTensor* out)
{
    // 检查self的数据类型是否在支持列表内
    auto supportList = GetDtypeSupportList();
    OP_CHECK_DTYPE_NOT_SUPPORT(self, supportList, return false);

    // 检查out的数据类型是否在支持列表内
    OP_CHECK_DTYPE_NOT_MATCH(out, op::DataType::DT_INT64, return false);
    return true;
}

inline static bool CheckShape(const aclTensor* self, const aclTensor* out)
{
    OP_CHECK_MAX_DIM(self, MAX_DIM_LEN, return false);
    OP_CHECK_MAX_DIM(out, MAX_DIM_LEN, return false);
    return true;
}

inline 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.  检查输出shape
    CHECK_RET(CheckShape(self, out), ACLNN_ERR_PARAM_INVALID);
    // 4. 检查输入shape
    CHECK_RET(CheckDims(self), ACLNN_ERR_PARAM_INVALID);
    return ACLNN_SUCCESS;
}

aclnnStatus aclnnNonzeroGetWorkspaceSize(
    const aclTensor* self, aclTensor* out, uint64_t* workspaceSize, aclOpExecutor** executor)
{
    L2_DFX_PHASE_1(aclnnNonzero, DFX_IN(self), DFX_OUT(out));
    // 固定写法，创建OpExecutor
    auto uniqueExecutor = CREATE_EXECUTOR();
    CHECK_RET(uniqueExecutor.get() != nullptr, ACLNN_ERR_INNER_CREATE_EXECUTOR);

    // 固定写法，参数检查
    auto ret = CheckParams(self, out);
    CHECK_RET(ret == ACLNN_SUCCESS, ret);

    if (self->IsEmpty() || out->IsEmpty()) {
        // 根据实际支持情况补充
        *workspaceSize = 0;
        uniqueExecutor.ReleaseTo(executor);
        return ACLNN_SUCCESS;
    }

    // 固定写法，将输入self转换成连续的tensor
    auto selfContiguous = l0op::Contiguous(self, uniqueExecutor.get());
    CHECK_RET(selfContiguous != nullptr, ACLNN_ERR_INNER_NULLPTR);

    // 还原out的StorageShape和OriginalShape为ViewShape
    auto outViewShape = out->GetViewShape();
    out->SetStorageShape(outViewShape);
    out->SetOriginalShape(outViewShape);

    // 调用主体计算函数
    out = l0op::Nonzero(selfContiguous, out, uniqueExecutor.get());
    CHECK_RET(out != 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();
    // 需要把 uniqueExecutor持有executor转移给executor
    uniqueExecutor.ReleaseTo(executor);
    return ACLNN_SUCCESS;
}

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

#ifdef __cplusplus
}
#endif