# Copyright 2022 Huawei Technologies Co., Ltd
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# ============================================================================

"""Define the Resnet50"""

import numpy as np
import mindspore.nn as nn
import mindspore as ms
import mindspore.ops as ops


def weight_variable_0(shape):
    """weight_variable_0"""
    zeros = np.zeros(shape).astype(np.float32)
    return ms.Tensor(zeros)


def weight_variable_1(shape):
    """weight_variable_1"""
    ones = np.ones(shape).astype(np.float32)
    return ms.Tensor(ones)


def conv3x3(in_channels, out_channels, stride=1, padding=0):
    """3x3 convolution """
    return nn.Conv2d(in_channels, out_channels,
                     kernel_size=3, stride=stride, padding=padding, weight_init='XavierUniform',
                     has_bias=False, pad_mode="same")


def conv1x1(in_channels, out_channels, stride=1, padding=0):
    """1x1 convolution"""
    return nn.Conv2d(in_channels, out_channels,
                     kernel_size=1, stride=stride, padding=padding, weight_init='XavierUniform',
                     has_bias=False, pad_mode="same")


def conv7x7(in_channels, out_channels, stride=1, padding=0):
    """1x1 convolution"""
    return nn.Conv2d(in_channels, out_channels,
                     kernel_size=7, stride=stride, padding=padding, weight_init='XavierUniform',
                     has_bias=False, pad_mode="same")


def bn_with_initialize(out_channels):
    """bn_with_initialize"""
    shape = (out_channels)
    mean = weight_variable_0(shape)
    var = weight_variable_1(shape)
    beta = weight_variable_0(shape)
    bn = nn.BatchNorm2d(out_channels, momentum=0.99, eps=0.00001, gamma_init='Uniform',
                        beta_init=beta, moving_mean_init=mean, moving_var_init=var)
    return bn


def bn_with_initialize_last(out_channels):
    """bn_with_initialize_last"""
    shape = (out_channels)
    mean = weight_variable_0(shape)
    var = weight_variable_1(shape)
    beta = weight_variable_0(shape)
    bn = nn.BatchNorm2d(out_channels, momentum=0.99, eps=0.00001, gamma_init='Uniform',
                        beta_init=beta, moving_mean_init=mean, moving_var_init=var)
    return bn


def fc_with_initialize(input_channels, out_channels):
    """fc_with_initialize"""
    return nn.Dense(input_channels, out_channels, weight_init='XavierUniform', bias_init='Uniform')


class ResidualBlock(nn.Cell):
    """ResidualBlock"""
    expansion = 4

    def __init__(self,
                 in_channels,
                 out_channels,
                 stride=1):
        """init block"""
        super(ResidualBlock, self).__init__()

        out_chls = out_channels // self.expansion
        self.conv1 = conv1x1(in_channels, out_chls, stride=stride, padding=0)
        self.bn1 = bn_with_initialize(out_chls)

        self.conv2 = conv3x3(out_chls, out_chls, stride=1, padding=0)
        self.bn2 = bn_with_initialize(out_chls)

        self.conv3 = conv1x1(out_chls, out_channels, stride=1, padding=0)
        self.bn3 = bn_with_initialize_last(out_channels)

        self.relu = ops.ReLU()
        self.add = ops.Add()

    def construct(self, x):
        """construct"""
        identity = x

        out = self.conv1(x)
        out = self.bn1(out)
        out = self.relu(out)

        out = self.conv2(out)
        out = self.bn2(out)
        out = self.relu(out)

        out = self.conv3(out)
        out = self.bn3(out)

        out = self.add(out, identity)
        out = self.relu(out)

        return out


class ResidualBlockWithDown(nn.Cell):
    """ResidualBlockWithDown"""
    expansion = 4

    def __init__(self,
                 in_channels,
                 out_channels,
                 stride=1,
                 down_sample=False):
        """init block with down"""
        super(ResidualBlockWithDown, self).__init__()

        out_chls = out_channels // self.expansion
        self.conv1 = conv1x1(in_channels, out_chls, stride=stride, padding=0)
        self.bn1 = bn_with_initialize(out_chls)

        self.conv2 = conv3x3(out_chls, out_chls, stride=1, padding=0)
        self.bn2 = bn_with_initialize(out_chls)

        self.conv3 = conv1x1(out_chls, out_channels, stride=1, padding=0)
        self.bn3 = bn_with_initialize_last(out_channels)

        self.relu = ops.ReLU()
        self.down_sample = down_sample

        self.conv_down_sample = conv1x1(in_channels, out_channels, stride=stride, padding=0)
        self.bn_down_sample = bn_with_initialize(out_channels)
        self.add = ops.Add()

    def construct(self, x):
        """construct"""
        identity = x

        out = self.conv1(x)
        out = self.bn1(out)
        out = self.relu(out)

        out = self.conv2(out)
        out = self.bn2(out)
        out = self.relu(out)

        out = self.conv3(out)
        out = self.bn3(out)

        identity = self.conv_down_sample(identity)
        identity = self.bn_down_sample(identity)

        out = self.add(out, identity)
        out = self.relu(out)

        return out


class MakeLayer0(nn.Cell):
    """MakeLayer0"""

    def __init__(self, block, in_channels, out_channels, stride):
        """init"""
        super(MakeLayer0, self).__init__()
        self.a = ResidualBlockWithDown(in_channels, out_channels, stride=1, down_sample=True)
        self.b = block(out_channels, out_channels, stride=stride)
        self.c = block(out_channels, out_channels, stride=1)

    def construct(self, x):
        """construct"""
        x = self.a(x)
        x = self.b(x)
        x = self.c(x)

        return x


class MakeLayer1(nn.Cell):
    """MakeLayer1"""

    def __init__(self, block, in_channels, out_channels, stride):
        """init"""
        super(MakeLayer1, self).__init__()
        self.a = ResidualBlockWithDown(in_channels, out_channels, stride=stride, down_sample=True)
        self.b = block(out_channels, out_channels, stride=1)
        self.c = block(out_channels, out_channels, stride=1)
        self.d = block(out_channels, out_channels, stride=1)

    def construct(self, x):
        """construct"""
        x = self.a(x)
        x = self.b(x)
        x = self.c(x)
        x = self.d(x)

        return x


class MakeLayer2(nn.Cell):
    """MakeLayer2"""

    def __init__(self, block, in_channels, out_channels, stride):
        """init"""
        super(MakeLayer2, self).__init__()
        self.a = ResidualBlockWithDown(in_channels, out_channels, stride=stride, down_sample=True)
        self.b = block(out_channels, out_channels, stride=1)
        self.c = block(out_channels, out_channels, stride=1)
        self.d = block(out_channels, out_channels, stride=1)
        self.e = block(out_channels, out_channels, stride=1)
        self.f = block(out_channels, out_channels, stride=1)

    def construct(self, x):
        """construct"""
        x = self.a(x)
        x = self.b(x)
        x = self.c(x)
        x = self.d(x)
        x = self.e(x)
        x = self.f(x)

        return x


class MakeLayer3(nn.Cell):
    """MakeLayer3"""

    def __init__(self, block, in_channels, out_channels, stride):
        """init"""
        super(MakeLayer3, self).__init__()
        self.a = ResidualBlockWithDown(in_channels, out_channels, stride=stride, down_sample=True)
        self.b = block(out_channels, out_channels, stride=1)
        self.c = block(out_channels, out_channels, stride=1)

    def construct(self, x):
        """construct"""
        x = self.a(x)
        x = self.b(x)
        x = self.c(x)

        return x


class Head(nn.Cell):
    """Head"""
    def __init__(self):
        super(Head, self).__init__()
        self.conv1 = conv7x7(3, 64, stride=2, padding=0)
        self.bn1 = bn_with_initialize(64)
        self.relu = ops.ReLU()
        self.maxpool = nn.MaxPool2d(kernel_size=3, stride=2, pad_mode="same")

    def construct(self, x):
        x = self.conv1(x)
        x = self.bn1(x)
        x = self.relu(x)
        x = self.maxpool(x)
        return x


class ResNet(nn.Cell):
    """ResNet"""

    def __init__(self, block, num_classes=100, batch_size=32):
        """init"""
        super(ResNet, self).__init__()
        self.batch_size = batch_size
        self.num_classes = num_classes
        self.head = Head()

        self.layer1 = MakeLayer0(block, in_channels=64, out_channels=256, stride=1)
        self.layer2 = MakeLayer1(block, in_channels=256, out_channels=512, stride=2)
        self.layer3 = MakeLayer2(block, in_channels=512, out_channels=1024, stride=2)
        self.layer4 = MakeLayer3(block, in_channels=1024, out_channels=2048, stride=2)

        self.pool = ops.ReduceMean(keep_dims=True)
        self.squeeze = ops.Squeeze(axis=(2, 3))
        self.fc = fc_with_initialize(512 * block.expansion, num_classes)

        # pipeline parallel config
        self.head.pipeline_stage = 0
        self.layer1.pipeline_stage = 0
        self.layer2.pipeline_stage = 0
        self.layer3.pipeline_stage = 1
        self.layer4.pipeline_stage = 1
        self.fc.pipeline_stage = 1

    def construct(self, x):
        """construct"""
        x = self.head(x)
        x = self.layer1(x)
        x = self.layer2(x)
        x = self.layer3(x)
        x = self.layer4(x)

        x = self.pool(x, (2, 3))
        x = self.squeeze(x)
        x = self.fc(x)
        return x


def resnet50(batch_size, num_classes):
    """create resnet50"""
    return ResNet(ResidualBlock, num_classes, batch_size)