/*
 * Copyright (c) 2017 ARM Limited.
 *
 * SPDX-License-Identifier: MIT
 *
 * Permission is hereby granted, free of charge, to any person obtaining a copy
 * of this software and associated documentation files (the "Software"), to
 * deal in the Software without restriction, including without limitation the
 * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
 * sell copies of the Software, and to permit persons to whom the Software is
 * furnished to do so, subject to the following conditions:
 *
 * The above copyright notice and this permission notice shall be included in all
 * copies or substantial portions of the Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
 * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
 * OUT OF OR IN CONCLCTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
 * SOFTWARE.
 */
#include "arm_compute/core/Types.h"
#include "arm_compute/runtime/NEON/functions/NEArithmeticSubtraction.h"
#include "arm_compute/runtime/Tensor.h"
#include "arm_compute/runtime/TensorAllocator.h"
#include "tests/NEON/Accessor.h"
#include "tests/PaddingCalculator.h"
#include "tests/datasets/ConvertPolicyDataset.h"
#include "tests/datasets/ShapeDatasets.h"
#include "tests/framework/Asserts.h"
#include "tests/framework/Macros.h"
#include "tests/framework/datasets/Datasets.h"
#include "tests/validation/Validation.h"
#include "tests/validation/fixtures/ArithmeticSubtractionFixture.h"

namespace arm_compute
{
namespace test
{
namespace validation
{
namespace
{
/** Input data sets **/
const auto ArithmeticSubtractionU8Dataset = combine(combine(framework::dataset::make("DataType", DataType::U8), framework::dataset::make("DataType", DataType::U8)),
                                                    framework::dataset::make("DataType",
                                                                             DataType::U8));
const auto ArithmeticSubtractionS16Dataset = combine(combine(framework::dataset::make("DataType", { DataType::U8, DataType::S16 }), framework::dataset::make("DataType", DataType::S16)),
                                                     framework::dataset::make("DataType", DataType::S16));
const auto ArithmeticSubtractionU8U8S16Dataset = combine(combine(framework::dataset::make("DataType", DataType::U8), framework::dataset::make("DataType", DataType::U8)),
                                                         framework::dataset::make("DataType", DataType::S16));
const auto ArithmeticSubtractionS16U8S16Dataset = combine(combine(framework::dataset::make("DataType", DataType::S16), framework::dataset::make("DataType", DataType::U8)),
                                                          framework::dataset::make("DataType", DataType::S16));
const auto ArithmeticSubtractionU8S16S16Dataset = combine(combine(framework::dataset::make("DataType", DataType::U8), framework::dataset::make("DataType", DataType::S16)),
                                                          framework::dataset::make("DataType", DataType::S16));
const auto ArithmeticSubtractionQS8Dataset = combine(combine(framework::dataset::make("DataType", DataType::QS8), framework::dataset::make("DataType", DataType::QS8)),
                                                     framework::dataset::make("DataType", DataType::QS8));
const auto ArithmeticSubtractionQS16Dataset = combine(combine(framework::dataset::make("DataType", DataType::QS16), framework::dataset::make("DataType", DataType::QS16)),
                                                      framework::dataset::make("DataType", DataType::QS16));
#ifdef __ARM_FEATURE_FP16_VECTOR_ARITHMETIC
const auto ArithmeticSubtractionFP16Dataset = combine(combine(framework::dataset::make("DataType", DataType::F16), framework::dataset::make("DataType", DataType::F16)),
                                                      framework::dataset::make("DataType", DataType::F16));
#endif /* __ARM_FEATURE_FP16_VECTOR_ARITHMETIC */
const auto ArithmeticSubtractionFP32Dataset = combine(combine(framework::dataset::make("DataType", DataType::F32), framework::dataset::make("DataType", DataType::F32)),
                                                      framework::dataset::make("DataType", DataType::F32));
} // namespace

TEST_SUITE(NEON)
TEST_SUITE(ArithmeticSubtraction)

// *INDENT-OFF*
// clang-format off
DATA_TEST_CASE(Validate, framework::DatasetMode::ALL, zip(zip(zip(
               framework::dataset::make("Input1Info", { TensorInfo(TensorShape(32U, 13U, 2U), 1, DataType::U8),
                                                        TensorInfo(TensorShape(32U, 13U, 2U), 1, DataType::U8),
                                                        TensorInfo(TensorShape(27U, 13U, 2U), 1, DataType::U8),      // Window shrink
                                                        TensorInfo(TensorShape(32U, 13U, 2U), 1, DataType::U8),      // Invalid data type combination
                                                        TensorInfo(TensorShape(32U, 13U, 2U), 1, DataType::F32),     // Mismatching shapes
                                                        TensorInfo(TensorShape(32U, 13U, 2U), 1, DataType::QS8, 2),  // Mismatching fixed point
                                                        TensorInfo(TensorShape(32U, 13U, 2U), 1, DataType::QS8, 2),
                                                      }),
               framework::dataset::make("Input2Info",{ TensorInfo(TensorShape(32U, 13U, 2U), 1, DataType::U8),
                                                       TensorInfo(TensorShape(32U, 13U, 2U), 1, DataType::U8),
                                                       TensorInfo(TensorShape(27U, 13U, 2U), 1, DataType::U8),
                                                       TensorInfo(TensorShape(32U, 13U, 2U), 1, DataType::S16),
                                                       TensorInfo(TensorShape(48U, 11U, 2U), 1, DataType::F32),
                                                       TensorInfo(TensorShape(32U, 13U, 2U), 1, DataType::QS8, 3),
                                                       TensorInfo(TensorShape(32U, 13U, 2U), 1, DataType::QS8, 2),
                                                     })),
               framework::dataset::make("OutputInfo",{ TensorInfo(TensorShape(32U, 13U, 2U), 1, DataType::S16),
                                                       TensorInfo(TensorShape(32U, 13U, 2U), 1, DataType::U8),
                                                       TensorInfo(TensorShape(27U, 13U, 2U), 1, DataType::U8),
                                                       TensorInfo(TensorShape(32U, 13U, 2U), 1, DataType::U8),
                                                       TensorInfo(TensorShape(48U, 11U, 2U), 1, DataType::F32),
                                                       TensorInfo(TensorShape(32U, 13U, 2U), 1, DataType::QS8, 3),
                                                       TensorInfo(TensorShape(32U, 13U, 2U), 1, DataType::QS8, 2),
                                                     })),
               framework::dataset::make("Expected", { true, true, false, false, false, false, true })),
               input1_info, input2_info, output_info, expected)
{
    ARM_COMPUTE_EXPECT(bool(NEArithmeticSubtraction::validate(&input1_info.clone()->set_is_resizable(false), &input2_info.clone()->set_is_resizable(false), &output_info.clone()->set_is_resizable(false), ConvertPolicy::WRAP)) == expected, framework::LogLevel::ERRORS);
}
// clang-format on
// *INDENT-ON*

template <typename T1, typename T2 = T1, typename T3 = T1>
using NEArithmeticSubtractionFixture = ArithmeticSubtractionValidationFixture<Tensor, Accessor, NEArithmeticSubtraction, T1, T2, T3>;

TEST_SUITE(U8)
DATA_TEST_CASE(Configuration, framework::DatasetMode::ALL, combine(framework::dataset::concat(datasets::SmallShapes(), datasets::LargeShapes()), framework::dataset::make("ConvertPolicy", { ConvertPolicy::SATURATE, ConvertPolicy::WRAP })),
               shape, policy)
{
    // Create tensors
    Tensor ref_src1 = create_tensor<Tensor>(shape, DataType::U8);
    Tensor ref_src2 = create_tensor<Tensor>(shape, DataType::U8);
    Tensor dst      = create_tensor<Tensor>(shape, DataType::U8);

    // Create and Configure function
    NEArithmeticSubtraction sub;
    sub.configure(&ref_src1, &ref_src2, &dst, policy);

    // Validate valid region
    const ValidRegion valid_region = shape_to_valid_region(shape);
    validate(dst.info()->valid_region(), valid_region);

    // Validate padding
    const PaddingSize padding = PaddingCalculator(shape.x(), 16).required_padding();
    validate(ref_src1.info()->padding(), padding);
    validate(ref_src2.info()->padding(), padding);
    validate(dst.info()->padding(), padding);
}

FIXTURE_DATA_TEST_CASE(RunSmall, NEArithmeticSubtractionFixture<uint8_t>, framework::DatasetMode::PRECOMMIT, combine(combine(datasets::SmallShapes(), ArithmeticSubtractionU8Dataset),
                                                                                                                     framework::dataset::make("ConvertPolicy", { ConvertPolicy::SATURATE, ConvertPolicy::WRAP })))
{
    // Validate output
    validate(Accessor(_target), _reference);
}
TEST_SUITE_END()

template <typename T1, typename T2 = T1>
using NEArithmeticSubtractionToS16Fixture = NEArithmeticSubtractionFixture<T1, T2, int16_t>;

TEST_SUITE(S16)
DATA_TEST_CASE(Configuration, framework::DatasetMode::ALL, combine(combine(combine(framework::dataset::concat(datasets::SmallShapes(), datasets::LargeShapes()),
                                                                                   framework::dataset::make("DataType", { DataType::U8, DataType::S16 })),
                                                                           framework::dataset::make("DataType", { DataType::U8, DataType::S16 })),
                                                                   framework::dataset::make("ConvertPolicy", { ConvertPolicy::SATURATE, ConvertPolicy::WRAP })),
               shape, data_type1, data_type2, policy)
{
    // Create tensors
    Tensor ref_src1 = create_tensor<Tensor>(shape, data_type1);
    Tensor ref_src2 = create_tensor<Tensor>(shape, data_type2);
    Tensor dst      = create_tensor<Tensor>(shape, DataType::S16);

    // Create and Configure function
    NEArithmeticSubtraction sub;
    sub.configure(&ref_src1, &ref_src2, &dst, policy);

    // Validate valid region
    const ValidRegion valid_region = shape_to_valid_region(shape);
    validate(dst.info()->valid_region(), valid_region);

    // Validate padding
    const PaddingSize padding = PaddingCalculator(shape.x(), 16).required_padding();
    validate(ref_src1.info()->padding(), padding);
    validate(ref_src2.info()->padding(), padding);
    validate(dst.info()->padding(), padding);
}

TEST_SUITE(S16_S16_S16)
FIXTURE_DATA_TEST_CASE(RunSmall, NEArithmeticSubtractionToS16Fixture<int16_t>, framework::DatasetMode::PRECOMMIT, combine(combine(datasets::SmallShapes(), ArithmeticSubtractionS16Dataset),
                       framework::dataset::make("ConvertPolicy", { ConvertPolicy::SATURATE, ConvertPolicy::WRAP })))
{
    // Validate output
    validate(Accessor(_target), _reference);
}

FIXTURE_DATA_TEST_CASE(RunLarge, NEArithmeticSubtractionToS16Fixture<int16_t>, framework::DatasetMode::NIGHTLY, combine(combine(datasets::LargeShapes(), ArithmeticSubtractionS16Dataset),
                                                                                                                        framework::dataset::make("ConvertPolicy", { ConvertPolicy::SATURATE, ConvertPolicy::WRAP })))
{
    // Validate output
    validate(Accessor(_target), _reference);
}
TEST_SUITE_END()

TEST_SUITE(U8_U8_S16)
FIXTURE_DATA_TEST_CASE(RunSmall, NEArithmeticSubtractionToS16Fixture<uint8_t>, framework::DatasetMode::PRECOMMIT, combine(combine(datasets::SmallShapes(),
                       ArithmeticSubtractionU8U8S16Dataset),
                       framework::dataset::make("ConvertPolicy", { ConvertPolicy::SATURATE, ConvertPolicy::WRAP })))
{
    // Validate output
    validate(Accessor(_target), _reference);
}

FIXTURE_DATA_TEST_CASE(RunLarge, NEArithmeticSubtractionToS16Fixture<uint8_t>, framework::DatasetMode::NIGHTLY, combine(combine(datasets::LargeShapes(),
                                                                                                                        ArithmeticSubtractionU8U8S16Dataset),
                                                                                                                        framework::dataset::make("ConvertPolicy", { ConvertPolicy::SATURATE, ConvertPolicy::WRAP })))
{
    // Validate output
    validate(Accessor(_target), _reference);
}
TEST_SUITE_END()

TEST_SUITE(S16_U8_S16)
using NEAriSubS16U8ToS16Fixture = NEArithmeticSubtractionToS16Fixture<int16_t, uint8_t>;
FIXTURE_DATA_TEST_CASE(RunSmall, NEAriSubS16U8ToS16Fixture, framework::DatasetMode::PRECOMMIT, combine(combine(datasets::SmallShapes(),
                                                                                                               ArithmeticSubtractionS16U8S16Dataset),
                                                                                                       framework::dataset::make("ConvertPolicy", { ConvertPolicy::SATURATE, ConvertPolicy::WRAP })))
{
    // Validate output
    validate(Accessor(_target), _reference);
}

FIXTURE_DATA_TEST_CASE(RunLarge, NEAriSubS16U8ToS16Fixture, framework::DatasetMode::NIGHTLY, combine(combine(datasets::LargeShapes(),
                                                                                                             ArithmeticSubtractionS16U8S16Dataset),
                                                                                                     framework::dataset::make("ConvertPolicy", { ConvertPolicy::SATURATE, ConvertPolicy::WRAP })))
{
    // Validate output
    validate(Accessor(_target), _reference);
}
TEST_SUITE_END()

TEST_SUITE(U8_S16_S16)
using NEAriSubU8S16ToS16Fixture = NEArithmeticSubtractionToS16Fixture<uint8_t, int16_t>;
FIXTURE_DATA_TEST_CASE(RunSmall, NEAriSubU8S16ToS16Fixture, framework::DatasetMode::PRECOMMIT, combine(combine(datasets::SmallShapes(),
                                                                                                               ArithmeticSubtractionU8S16S16Dataset),
                                                                                                       framework::dataset::make("ConvertPolicy", { ConvertPolicy::SATURATE, ConvertPolicy::WRAP })))
{
    // Validate output
    validate(Accessor(_target), _reference);
}

FIXTURE_DATA_TEST_CASE(RunLarge, NEAriSubU8S16ToS16Fixture, framework::DatasetMode::NIGHTLY, combine(combine(datasets::LargeShapes(),
                                                                                                             ArithmeticSubtractionU8S16S16Dataset),
                                                                                                     framework::dataset::make("ConvertPolicy", { ConvertPolicy::SATURATE, ConvertPolicy::WRAP })))
{
    // Validate output
    validate(Accessor(_target), _reference);
}
TEST_SUITE_END()
TEST_SUITE_END()

template <typename T1, typename T2 = T1, typename T3 = T1>
using NEArithmeticSubtractionFixedPointFixture = ArithmeticSubtractionValidationFixedPointFixture<Tensor, Accessor, NEArithmeticSubtraction, T1, T2, T3>;

TEST_SUITE(Quantized)
TEST_SUITE(QS8)
FIXTURE_DATA_TEST_CASE(RunSmall, NEArithmeticSubtractionFixedPointFixture<int8_t>, framework::DatasetMode::PRECOMMIT, combine(combine(combine(datasets::SmallShapes(),
                       ArithmeticSubtractionQS8Dataset),
                       framework::dataset::make("ConvertPolicy", { ConvertPolicy::SATURATE, ConvertPolicy::WRAP })),
                       framework::dataset::make("FractionalBits", 1, 7)))
{
    // Validate output
    validate(Accessor(_target), _reference);
}

FIXTURE_DATA_TEST_CASE(RunLarge, NEArithmeticSubtractionFixedPointFixture<int8_t>, framework::DatasetMode::NIGHTLY, combine(combine(combine(datasets::LargeShapes(),
                       ArithmeticSubtractionQS8Dataset),
                       framework::dataset::make("ConvertPolicy", { ConvertPolicy::SATURATE, ConvertPolicy::WRAP })),
                       framework::dataset::make("FractionalBits", 1, 7)))
{
    // Validate output
    validate(Accessor(_target), _reference);
}
TEST_SUITE_END()

TEST_SUITE(QS16)
FIXTURE_DATA_TEST_CASE(RunSmall, NEArithmeticSubtractionFixedPointFixture<int16_t>, framework::DatasetMode::PRECOMMIT, combine(combine(combine(datasets::SmallShapes(),
                       ArithmeticSubtractionQS16Dataset),
                       framework::dataset::make("ConvertPolicy", { ConvertPolicy::SATURATE, ConvertPolicy::WRAP })),
                       framework::dataset::make("FractionalBits", 1, 15)))
{
    // Validate output
    validate(Accessor(_target), _reference);
}

FIXTURE_DATA_TEST_CASE(RunLarge, NEArithmeticSubtractionFixedPointFixture<int16_t>, framework::DatasetMode::NIGHTLY, combine(combine(combine(datasets::LargeShapes(),
                       ArithmeticSubtractionQS16Dataset),
                       framework::dataset::make("ConvertPolicy", { ConvertPolicy::SATURATE, ConvertPolicy::WRAP })),
                       framework::dataset::make("FractionalBits", 1, 15)))
{
    // Validate output
    validate(Accessor(_target), _reference);
}
TEST_SUITE_END()
TEST_SUITE_END()

TEST_SUITE(Float)
#ifdef __ARM_FEATURE_FP16_VECTOR_ARITHMETIC
TEST_SUITE(FP16)
FIXTURE_DATA_TEST_CASE(RunSmall, NEArithmeticSubtractionFixture<half>, framework::DatasetMode::ALL, combine(combine(datasets::SmallShapes(), ArithmeticSubtractionFP16Dataset),
                                                                                                            framework::dataset::make("ConvertPolicy", { ConvertPolicy::SATURATE, ConvertPolicy::WRAP })))
{
    // Validate output
    validate(Accessor(_target), _reference);
}
TEST_SUITE_END()
#endif /* __ARM_FEATURE_FP16_VECTOR_ARITHMETIC */

TEST_SUITE(FP32)
DATA_TEST_CASE(Configuration, framework::DatasetMode::ALL, combine(framework::dataset::concat(datasets::SmallShapes(), datasets::LargeShapes()), framework::dataset::make("ConvertPolicy", { ConvertPolicy::SATURATE, ConvertPolicy::WRAP })),
               shape, policy)
{
    // Create tensors
    Tensor ref_src1 = create_tensor<Tensor>(shape, DataType::F32);
    Tensor ref_src2 = create_tensor<Tensor>(shape, DataType::F32);
    Tensor dst      = create_tensor<Tensor>(shape, DataType::F32);

    // Create and Configure function
    NEArithmeticSubtraction sub;
    sub.configure(&ref_src1, &ref_src2, &dst, policy);

    // Validate valid region
    const ValidRegion valid_region = shape_to_valid_region(shape);
    validate(dst.info()->valid_region(), valid_region);

    // Validate padding
    const PaddingSize padding = PaddingCalculator(shape.x(), 16).required_padding();
    validate(ref_src1.info()->padding(), padding);
    validate(ref_src2.info()->padding(), padding);
    validate(dst.info()->padding(), padding);
}

FIXTURE_DATA_TEST_CASE(RunSmall, NEArithmeticSubtractionFixture<float>, framework::DatasetMode::PRECOMMIT, combine(combine(datasets::SmallShapes(), ArithmeticSubtractionFP32Dataset),
                                                                                                                   framework::dataset::make("ConvertPolicy", { ConvertPolicy::SATURATE, ConvertPolicy::WRAP })))
{
    // Validate output
    validate(Accessor(_target), _reference);
}

FIXTURE_DATA_TEST_CASE(RunLarge, NEArithmeticSubtractionFixture<float>, framework::DatasetMode::NIGHTLY, combine(combine(datasets::LargeShapes(), ArithmeticSubtractionFP32Dataset),
                                                                                                                 framework::dataset::make("ConvertPolicy", { ConvertPolicy::SATURATE, ConvertPolicy::WRAP })))
{
    // Validate output
    validate(Accessor(_target), _reference);
}
TEST_SUITE_END()
TEST_SUITE_END()

TEST_SUITE_END()
TEST_SUITE_END()
} // namespace validation
} // namespace test
} // namespace arm_compute