import torch

import torch.nn as nn

import torchvision.transforms.functional as TF

class DoubleConv(nn.Module):

    def __init__(self, in_channels, out_channels):

        super(DoubleConv, self).__init__()

        self.conv = nn.Sequential(

            # kernel_size=3

            # stride=1(evry rown a nd column is affected)

            # padding=1(same conv)input height i width isti nakon konv

            # bias=false

            # batchnomr-normalizacija

            # relu aktivacijska funk

            nn.Conv2d(in_channels, out_channels, 3, 1, 1, bias=False),

            nn.BatchNorm2d(out_channels),

            nn.ReLU(inplace=True),

            nn.Conv2d(out_channels, out_channels, 3, 1, 1, bias=False),

            nn.BatchNorm2d(out_channels),

            ##inplace means that it will modify the input directly, without allocating any additional output

            nn.ReLU(inplace=True),

        )

    def forward(self, x):

        return self.conv(x)

class UNET(nn.Module):

    ##moguce da broj out kanala promjenis,ovjde 1 jer je binary image segemntation

    def __init__(self, in_channels=3, out_channels=2, features=[64, 128, 256, 512]):

        super(UNET, self).__init__()

        # jer hocemo evaluirat i sve to imamo tu listu zbog layera,spremamo sve te konvolucije

        self.downs = nn.ModuleList()

        self.ups = nn.ModuleList()

        self.pool = nn.MaxPool2d(kernel_size=2, stride=2)

        # Down part

        for feature in features:

            self.downs.append(DoubleConv(in_channels, feature))

            in_channels = feature

        # Up part

        for feature in reversed(features):

            self.ups.append(nn.ConvTranspose2d(feature * 2, feature, kernel_size=2,

                                               stride=2))  ##kernel size is tride ce poduplat ovdje height i width slike

            self.ups.append(DoubleConv(feature * 2, feature))

        self.bottleneck = DoubleConv(features[-1], features[-1] * 2)  # onaj zasebni dio

        self.final_conv = nn.Conv2d(features[0], out_channels, kernel_size=1)

    def forward(self, x):

        skip_connections = []

        for down in self.downs:

            x = down(x)

            skip_connections.append(x)

            x = self.pool(x)

        x = self.bottleneck(x)

        skip_connections = skip_connections[::-1]  # obrnuti redoslijed

        for idx in range(0, len(self.ups), 2):

            x = self.ups[idx](x)

            skip_connection = skip_connections[idx // 2]

            if x.shape != skip_connection.shape:

                x = TF.resize(x, size=skip_connection.shape[2:])  ##ako input nije djeljiv s 2, resize se

            concat_skip = torch.cat((skip_connection, x), dim=1)

            x = self.ups[idx + 1](concat_skip)

        return self.final_conv(x)