/**
 * 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 <bitset>
#include "aclnn_reduce_sum.h"
#include "reduce_sum_op.h"
#include "aclnn_kernels/cast.h"
#include "aclnn_kernels/contiguous.h"
#include "aclnn_kernels/common/op_error_check.h"
#include "opdev/common_types.h"
#include "opdev/shape_utils.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/tensor_view_utils.h"
#include "opdev/platform.h"
#include "math/reduce_any/op_api/reduce_any.h"
#include "conversion/fill/op_api/fill.h"
#include "op_api/op_api_def.h"
#include "op_api/aclnn_check.h"

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

constexpr size_t MAX_MASK_LEN = 64;

// 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_FLOAT16, op::DataType::DT_FLOAT,     op::DataType::DT_INT8,      op::DataType::DT_INT16,
    op::DataType::DT_INT32,   op::DataType::DT_INT64,     op::DataType::DT_UINT8,     op::DataType::DT_BOOL,
    op::DataType::DT_DOUBLE,  op::DataType::DT_COMPLEX64, op::DataType::DT_COMPLEX128};

static const std::initializer_list<op::DataType> ASCEND910B_DTYPE_SUPPORT_LIST = {
    op::DataType::DT_FLOAT16, op::DataType::DT_FLOAT,     op::DataType::DT_INT8,       op::DataType::DT_INT16,
    op::DataType::DT_INT32,   op::DataType::DT_INT64,     op::DataType::DT_UINT8,      op::DataType::DT_BOOL,
    op::DataType::DT_DOUBLE,  op::DataType::DT_COMPLEX64, op::DataType::DT_COMPLEX128, op::DataType::DT_BF16};

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

static const std::initializer_list<op::DataType> NON_CONTIOUS_SUPPORT_DTYPE_SUPPORT_LIST = {op::DataType::DT_FLOAT};

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

static bool CheckDtypeValid(const aclTensor* self, const aclDataType dtype, const aclTensor* out)
{
    // 检查self和out的数据类型是否在支持列表内
    bool isAscend910BSocVersion =
        (GetCurrentPlatformInfo().GetSocVersion() == SocVersion::ASCEND910B ||
         GetCurrentPlatformInfo().GetSocVersion() == SocVersion::ASCEND910_93 || IsRegBase());
    const std::initializer_list<op::DataType> CURRENT_DTYPE_SUPPORT_LIST =
        isAscend910BSocVersion ? ASCEND910B_DTYPE_SUPPORT_LIST : ASCEND910_DTYPE_SUPPORT_LIST;

    OP_CHECK_DTYPE_NOT_SUPPORT(self, CURRENT_DTYPE_SUPPORT_LIST, return false);
    OP_CHECK_DTYPE_NOT_SUPPORT(out, CURRENT_DTYPE_SUPPORT_LIST, return false);
    // 检查dtype指定的数据类型是否支持
    if (!CheckType(op::ToOpDataType(dtype), CURRENT_DTYPE_SUPPORT_LIST)) {
        OP_LOGE(
            ACLNN_ERR_PARAM_INVALID, "type %s should be in dtype support list [%s].",
            op::ToString(op::ToOpDataType(dtype)).GetString(), op::ToString(CURRENT_DTYPE_SUPPORT_LIST).GetString());
        return false;
    }
    // 检查dtype数据类型与out是否一致
    OP_CHECK_DTYPE_NOT_MATCH(out, op::ToOpDataType(dtype), return false);
    return true;
}

static bool CheckMaxDimension(const aclTensor* self)
{
    OP_CHECK_MAX_DIM(self, MAX_SUPPORT_DIMS_NUMS, return false);
    return true;
}

static inline uint64_t GetPosDim(int64_t dim, int64_t dimNum)
{
    if (dimNum <= 0) {
        dimNum = 1;
    }
    return dim >= 0 ? dim : dim + dimNum;
}

static bool CheckDimValid(const aclTensor* self, const aclIntArray* dims)
{
    auto selfViewShape = self->GetViewShape();
    auto selfDimNum = static_cast<int64_t>(selfViewShape.GetDimNum());
    // self为标量时，dim range [-1, 0]
    if (selfDimNum <= 0) {
        selfDimNum = 1;
    }
    // dim为负时需要转正校验
    std::bitset<MAX_MASK_LEN> dimMask = std::bitset<MAX_MASK_LEN>();

    for (size_t i = 0; i < dims->Size(); i++) {
        int64_t curDim = (*dims)[i];
        if (curDim >= selfDimNum || curDim < (-selfDimNum)) {
            OP_LOGE(
                ACLNN_ERR_PARAM_INVALID, "Provided dim %ld must be in the range of [%ld, %ld].", curDim, -selfDimNum,
                selfDimNum - 1);
            return false;
        }
        uint64_t index = GetPosDim(curDim, selfDimNum);
        // dim重复
        if (dimMask[index]) {
            OP_LOGE(ACLNN_ERR_PARAM_INVALID, "Dim %lu appears multiple times in the list of dims.", index);
            return false;
        }

        dimMask.set(index);
    }

    return true;
}

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

    // 2. 检查self、out的数据类型是否合法
    CHECK_RET(CheckDtypeValid(self, dtype, out), ACLNN_ERR_PARAM_INVALID);

    // 3. 检查最大维度是否超过8
    CHECK_RET(CheckMaxDimension(self), ACLNN_ERR_PARAM_INVALID);

    // 4. 检查reduce的轴是否超出self维度范围
    CHECK_RET(CheckDimValid(self, dims), ACLNN_ERR_PARAM_INVALID);

    return ACLNN_SUCCESS;
}

static aclnnStatus FillScalar(aclTensor* out, float val, aclOpExecutor* executor)
{
    OP_CHECK_DTYPE_NOT_SUPPORT(out, EMPTY_INPUT_DTYPE_SUPPORT_LIST, return ACLNN_ERR_PARAM_INVALID);
    FVector<int64_t> shape;
    size_t dimNum = out->GetViewShape().GetDimNum();

    if (out->IsEmpty()) {
        return ACLNN_SUCCESS;
    }

    for (size_t idx = 0; idx < dimNum; idx++) {
        int64_t tmpVal = out->GetViewShape().GetDim(idx);
        shape.push_back(tmpVal);
    }

    auto dims = executor->ConvertToTensor(shape.data(), shape.size(), DataType::DT_INT64);
    auto shapeArray = executor->AllocIntArray(shape.data(), shape.size());

    FVector<float> valVector = {val};
    auto valTensor = executor->ConvertToTensor(valVector.data(), valVector.size(), out->GetDataType());
    auto fillOut = l0op::Fill(dims, valTensor, shapeArray, executor);
    CHECK_RET(fillOut != nullptr, ACLNN_ERR_INNER_NULLPTR);
    auto viewCopyResult = l0op::ViewCopy(fillOut, out, executor);
    CHECK_RET(viewCopyResult != nullptr, ACLNN_ERR_INNER_NULLPTR);
    return ACLNN_SUCCESS;
}

static bool IsNonContiguousSupport(const aclTensor* self, const DataType promoteType, const aclIntArray* dims)
{
    if (!IsRegBase()) {
        return false;
    }
    if (!CheckType(self->GetDataType(), NON_CONTIOUS_SUPPORT_DTYPE_SUPPORT_LIST) ||
        self->GetDataType() != promoteType) {
        return false;
    }
    if (!op::IsReduceNonContiguousSupport(self, dims)) {
        return false;
    }
    return true;
}

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

aclnnStatus aclnnReduceSumGetWorkspaceSize(
    const aclTensor* self, const aclIntArray* dims, bool keepDims, aclDataType dtype, aclTensor* out,
    uint64_t* workspaceSize, aclOpExecutor** executor)
{
    L2_DFX_PHASE_1(aclnnReduceSum, DFX_IN(self, dims, keepDims, dtype), 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, dims, dtype, out);
    CHECK_RET(ret == ACLNN_SUCCESS, ret);

    // 检查self的format是否支持
    CheckFormat(self);
        
    // 输入self为空tensor时，直接返回dtype类型的空tensor
    if (self->IsEmpty()) {
        ret = FillScalar(out, 0.0f, uniqueExecutor.get());
        if (ret == ACLNN_SUCCESS) {
            *workspaceSize = uniqueExecutor->GetWorkspaceSize();
            uniqueExecutor.ReleaseTo(executor);
        }
        return ret;
    }

    // 当输入tensor是0维时，直接将输入tensor作为输出返回
    if (self->GetViewShape().GetDimNum() == 0) {
        // 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);

        // 将计算结果转换成输出out的数据类型
        auto selfCasted = l0op::Cast(selfContiguous, out->GetDataType(), uniqueExecutor.get());
        CHECK_RET(selfCasted != nullptr, ACLNN_ERR_INNER_NULLPTR);

        auto viewCopyResult = l0op::ViewCopy(selfCasted, out, uniqueExecutor.get());
        CHECK_RET(viewCopyResult != nullptr, ACLNN_ERR_INNER_NULLPTR);
        *workspaceSize = uniqueExecutor->GetWorkspaceSize();
        uniqueExecutor.ReleaseTo(executor);
        return ACLNN_SUCCESS;
    }

    // 数据类型转换处理
    op::DataType selfType = self->GetDataType();
    op::DataType dataType = op::ToOpDataType(dtype);
    op::DataType promoteType = dataType;
    promoteType = (selfType == promoteType) ? selfType : promoteType;

    // 空dims处理
    op::Shape shape = self->GetViewShape();
    if (dims->Size() == 0) {
        size_t dimDum = shape.GetDimNum();
        int64_t appendDim[dimDum];
        for (uint64_t i = 0; i < dimDum; i++) {
            appendDim[i] = i;
        }
        dims = uniqueExecutor.get()->AllocIntArray(appendDim, dimDum);
    }

    // self为bool，dtype为int64处理
    if (selfType == op::DataType::DT_BOOL && dataType == op::DataType::DT_INT64) {
        static const size_t maxDim = 16777216; // 2^24
        size_t reduceDims = 1;
        for (int i = 0; i < static_cast<int>(dims->Size()); i++) {
            reduceDims *= shape[(*dims)[i]];
        }
        promoteType = (reduceDims < maxDim) ? op::DataType::DT_FLOAT : op::DataType::DT_INT64;
    }
    if (IsNonContiguousSupport(self, promoteType, dims)) {
        OP_LOGD("Enter NonContigous");
        auto selfContiguous = uniqueExecutor.get()->CreateView(
            self, self->GetViewShape(), self->GetStorageShape(), self->GetViewStrides(), self->GetViewOffset());
        CHECK_RET(selfContiguous != nullptr, ACLNN_ERR_INNER_NULLPTR);
        // 调用ReduceSum算子kernel,将输入self的数据类型转换成指定的数据类型
        const aclTensor* reduceSumOut = nullptr;
        reduceSumOut = l0op::ReduceSumOp(selfContiguous, dims, keepDims, uniqueExecutor.get());
        CHECK_RET(reduceSumOut != nullptr, ACLNN_ERR_INNER_NULLPTR);
        CHECK_RET(CheckShapeAndScalarSame(reduceSumOut, out), ACLNN_ERR_PARAM_INVALID);
        reduceSumOut = l0op::Cast(reduceSumOut, out->GetDataType(), uniqueExecutor.get());
        auto viewCopyResult = l0op::ViewCopy(reduceSumOut, out, uniqueExecutor.get());
        CHECK_RET(viewCopyResult != nullptr, ACLNN_ERR_INNER_NULLPTR);
    } else {
        OP_LOGD("Enter Contigous");
        // 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);

        // 将输入self的数据类型进行转换
        auto selfContiguousCasted = l0op::Cast(selfContiguous, promoteType, uniqueExecutor.get());
        CHECK_RET(selfContiguousCasted != nullptr, ACLNN_ERR_INNER_NULLPTR);

        // 调用ReduceSum算子kernel,将输入self的数据类型转换成指定的数据类型
        const aclTensor* reduceSumOut = nullptr;
        if (dataType == op::DataType::DT_BOOL) {
            reduceSumOut = l0op::ReduceAny(selfContiguousCasted, dims, keepDims, uniqueExecutor.get());
        } else {
            reduceSumOut = l0op::ReduceSumOp(selfContiguousCasted, dims, keepDims, uniqueExecutor.get());
        }
        CHECK_RET(reduceSumOut != nullptr, ACLNN_ERR_INNER_NULLPTR);
        CHECK_RET(CheckShapeAndScalarSame(reduceSumOut, out), ACLNN_ERR_PARAM_INVALID);

        // 将计算结果转换成输出out的数据类型
        reduceSumOut = l0op::Cast(reduceSumOut, out->GetDataType(), uniqueExecutor.get());

        // 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(reduceSumOut, 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);
    return ACLNN_SUCCESS;
}

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

#ifdef __cplusplus
}
#endif