/**
 * 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_kernels/cast.h"
#include "isclose.h"
#include "opdev/common_types.h"
#include "opdev/data_type_utils.h"
#include "opdev/format_utils.h"
#include "opdev/make_op_executor.h"
#include "opdev/op_log.h"
#include "opdev/op_dfx.h"
#include "conversion/broadcast_to/op_api/broadcast_to.h"
#include "aclnn_kernels/contiguous.h"
#include "aclnn_kernels/common/op_error_check.h"
#include "aclnn_isclose.h"

using namespace op;

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> DTYPE_SUPPORT_LIST_910 = {
    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> DTYPE_SUPPORT_LIST_910B = {
    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 bool CheckNotNull(const aclTensor* self, const aclTensor* other, const aclTensor* out)
{
    OP_CHECK_NULL(self, return false);
    OP_CHECK_NULL(other, return false);
    OP_CHECK_NULL(out, return false);

    return true;
}

static bool CheckDtypeValid(const aclTensor* self, const aclTensor* other)
{
    bool isBf16Support =
        (GetCurrentPlatformInfo().GetSocVersion() >= SocVersion::ASCEND910B &&
         GetCurrentPlatformInfo().GetSocVersion() <= SocVersion::ASCEND910E);
    const std::initializer_list<op::DataType> dtypeSupportList =
        isBf16Support ? DTYPE_SUPPORT_LIST_910B : DTYPE_SUPPORT_LIST_910;

    // 检查self的数据类型是否在支持列表内
    OP_CHECK_DTYPE_NOT_SUPPORT(self, dtypeSupportList, return false);

    // 检查other的数据类型是否在支持列表内
    OP_CHECK_DTYPE_NOT_SUPPORT(other, dtypeSupportList, return false);

    // self和other数据类型必须一样
    OP_CHECK_DTYPE_NOT_MATCH(self, other->GetDataType(), return false);

    return true;
}

// self、other的shape需要满足broadcast规则，out的shape为broadcast后的shape
static bool CheckShape(const aclTensor* self, const aclTensor* other, const aclTensor* out)
{
    op::Shape broadcastShape;
    // self的数据维度不能超过8
    OP_CHECK_MAX_DIM(self, MAX_DIM_LEN, return false);

    // other的数据维度不能超过8
    OP_CHECK_MAX_DIM(other, MAX_DIM_LEN, return false);

    OP_CHECK_BROADCAST_AND_INFER_SHAPE(self, other, broadcastShape, return false);
    OP_CHECK_SHAPE_NOT_EQUAL_WITH_EXPECTED_SIZE(out, broadcastShape, return false);

    return true;
}

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

    // 3. 检查self和other以及out的维度匹配关系
    CHECK_RET(CheckShape(self, other, out), ACLNN_ERR_PARAM_INVALID);

    return ACLNN_SUCCESS;
}

aclnnStatus aclnnIsCloseGetWorkspaceSize(
    const aclTensor* self, const aclTensor* other, double rtol, double atol, bool equal_nan, aclTensor* out,
    uint64_t* workspaceSize, aclOpExecutor** executor)
{
    OP_CHECK_COMM_INPUT(workspaceSize, executor);

    L2_DFX_PHASE_1(aclnnIsClose, DFX_IN(self, other, rtol, atol, equal_nan), DFX_OUT(out));

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

    // 将rtol转为float32
    auto rtol_fp32 = static_cast<float>(rtol);

    // 将atol转为float32
    auto atol_fp32 = static_cast<float>(atol);

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

    // ISCLOSE算子的空tensor在kernel中支持
    if (self->IsEmpty() || other->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selfContiguous转换成连续的tensor
    auto selfContiguous = l0op::Contiguous(self, uniqueExecutor.get());
    CHECK_RET(selfContiguous != nullptr, ACLNN_ERR_INNER_NULLPTR);

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

    // 调用IsClose算子kernel
    auto isCloseOpOut =
        l0op::IsClose(selfContiguous, otherContiguous, rtol_fp32, atol_fp32, equal_nan, out, uniqueExecutor.get());
    CHECK_RET(isCloseOpOut != 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 viewCopyResult = l0op::ViewCopy(isCloseOpOut, out, uniqueExecutor.get());
    CHECK_RET(viewCopyResult != 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 aclnnIsClose(void* workspace, uint64_t workspaceSize, aclOpExecutor* executor, aclrtStream stream)
{
    L2_DFX_PHASE_2(aclnnIsClose);
    // 固定写法，调用框架能力，完成计算
    return CommonOpExecutorRun(workspace, workspaceSize, executor, stream);
}