import torch
import argparse
import torch.nn as nn
import torch.nn.functional as F
import torch.optim as optim
from torch.optim.lr_scheduler import StepLR

class CudaEventTimer(object):
    def __init__(self, start_event: torch.cuda.Event, end_event: torch.cuda.Event):
        self.start_event = start_event
        self.end_event = end_event

    def get_elapsed_msec(self):
        torch.cuda.current_stream().wait_event(self.end_event)
        self.end_event.synchronize()
        return self.start_event.elapsed_time(self.end_event)


class Timer:
    """Timer."""
    def __init__(self):
        self.started_ = False
        self.event_timers = []
        self.start_event = None
        self.elapsed_records = None

    def start(self):
        """Start the timer."""
        self.start_event = torch.cuda.Event(enable_timing=True)
        self.start_event.record()
        self.started_ = True

    def stop(self, reset=False, record=False):
        """Stop the timer."""
        assert self.started_, "timer is not started"
        end_event = torch.cuda.Event(enable_timing=True)
        end_event.record()
        self.event_timers.append(CudaEventTimer(self.start_event, end_event))
        self.start_event = None
        self.started_ = False

    def _get_elapsed_msec(self):
        self.elapsed_records = [et.get_elapsed_msec() for et in self.event_timers]
        self.event_timers.clear()
        return sum(self.elapsed_records)

    def reset(self):
        """Reset timer."""
        self.started_ = False
        self.start_event = None
        self.elapsed_records = None
        self.event_timers.clear()

    def elapsed(self, reset=True):
        """Calculate the elapsed time."""
        started_ = self.started_
        # If the timing in progress, end it first.
        if self.started_:
            self.stop()
        # Get the elapsed time.
        elapsed_ = self._get_elapsed_msec()
        # Reset the elapsed time
        if reset:
            self.reset()
        # If timing was in progress, set it back.
        if started_:
            self.start()
        return elapsed_

args = None
epoch_fwd_times = []

class Net(nn.Module):
    def __init__(self):
        super(Net, self).__init__()
        self.conv1 = nn.Conv2d(1, 32, 3, 1)
        self.conv2 = nn.Conv2d(32, 64, 3, 1)
        self.dropout1 = nn.Dropout(0.25)
        self.dropout2 = nn.Dropout(0.5)
        self.fc1 = nn.Linear(9216, 128)
        self.fc2 = nn.Linear(128, 10)
        self.timer = Timer()

    def forward(self, x):
        global epoch_fwd_times
        self.timer.start()
        x = self.conv1(x)
        x = F.relu(x)
        x = self.conv2(x)
        x = F.relu(x)
        x = F.max_pool2d(x, 2)
        x = self.dropout1(x)
        x = torch.flatten(x, 1)
        x = self.fc1(x)
        x = F.relu(x)
        x = self.dropout2(x)
        x = self.fc2(x)
        output = F.log_softmax(x, dim=1)
        self.timer.stop()
        fwd_time = self.timer.elapsed(reset=True)
        epoch_fwd_times.append(fwd_time)
        return output


def train(args, model, device, train_loader, optimizer, epoch):
    global epoch_fwd_times
    model.train()
    target = torch.LongTensor(args.batch_size).random_(10).to(device)
    for batch_idx, images in enumerate(train_loader):
        # data, target = data.to(device), target.to(device)
        data = images.to(device, non_blocking=True)
        optimizer.zero_grad()
        output = model(data)
        loss = F.nll_loss(output, target)
        loss.backward()
        optimizer.step()
        if batch_idx % args.log_interval == 0:
            print('Train Epoch: {} [{}/{} ({:.0f}%)]\tLoss: {:.6f}'.format(
                epoch, batch_idx * len(data), len(train_loader.dataset),
                100. * batch_idx / len(train_loader), loss.item()))
            if args.dry_run:
                break
    print(f"Epoch avg fwd_time: {sum(epoch_fwd_times) / len(epoch_fwd_times)}")
    epoch_fwd_times = []


def test(model, device, test_loader):
    model.eval()
    test_loss = 0
    correct = 0
    with torch.no_grad():
        for data, target in test_loader:
            data, target = data.to(device), target.to(device)
            output = model(data)
            test_loss += F.nll_loss(output, target, reduction='sum').item()  # sum up batch loss
            pred = output.argmax(dim=1, keepdim=True)  # get the index of the max log-probability
            correct += pred.eq(target.view_as(pred)).sum().item()

    test_loss /= len(test_loader.dataset)

    print('\nTest set: Average loss: {:.4f}, Accuracy: {}/{} ({:.0f}%)\n'.format(
        test_loss, correct, len(test_loader.dataset),
        100. * correct / len(test_loader.dataset)))


def main():
    global args
    # Training settings
    parser = argparse.ArgumentParser(description='PyTorch MNIST Example')
    parser.add_argument('--batch-size', type=int, default=64, metavar='N',
                        help='input batch size for training (default: 64)')
    parser.add_argument('--test-batch-size', type=int, default=1000, metavar='N',
                        help='input batch size for testing (default: 1000)')
    parser.add_argument('--epochs', type=int, default=14, metavar='N',
                        help='number of epochs to train (default: 14)')
    parser.add_argument('--lr', type=float, default=1.0, metavar='LR',
                        help='learning rate (default: 1.0)')
    parser.add_argument('--iterations', type=int, default=1000)
    parser.add_argument('--gamma', type=float, default=0.7, metavar='M',
                        help='Learning rate step gamma (default: 0.7)')
    parser.add_argument('--no-cuda', action='store_true', default=False,
                        help='disables CUDA training')
    parser.add_argument('--dry-run', action='store_true', default=False,
                        help='quickly check a single pass')
    parser.add_argument('--seed', type=int, default=1, metavar='S',
                        help='random seed (default: 1)')
    parser.add_argument('--log-interval', type=int, default=10, metavar='N',
                        help='how many batches to wait before logging training status')
    parser.add_argument('--save-model', action='store_true', default=False,
                        help='For Saving the current Model')
    args = parser.parse_args()
    use_cuda = not args.no_cuda and torch.cuda.is_available()

    torch.manual_seed(args.seed)

    if use_cuda:
        device = torch.device("cuda")
    else:
        device = torch.device("cpu")

    train_kwargs = {'batch_size': args.batch_size}
    test_kwargs = {'batch_size': args.test_batch_size}
    if use_cuda:
        cuda_kwargs = {'num_workers': 1,
                       'pin_memory': True,
                       'shuffle': True}
        train_kwargs.update(cuda_kwargs)
        test_kwargs.update(cuda_kwargs)

    # transform=transforms.Compose([
    #     transforms.ToTensor(),
    #     transforms.Normalize((0.1307,), (0.3081,))
    #     ])
    # dataset1 = datasets.MNIST('../data', train=True, download=True,
    #                    transform=transform)
    # dataset2 = datasets.MNIST('../data', train=False,
    #                    transform=transform)
    # train_loader = torch.utils.data.DataLoader(dataset1,**train_kwargs)
    # test_loader = torch.utils.data.DataLoader(dataset2, **test_kwargs)

    from torch.utils.data import Dataset, DataLoader
    class RandomDataset(Dataset):
        def __init__(self, length):
            self.len = length
            self.data = torch.randn(1, 28, 28, length)

        def __getitem__(self, index):
            return self.data[:, :, :, index]

        def __len__(self):
            return self.len
    train_dataset = RandomDataset(args.batch_size * args.iterations)
    train_loader = torch.utils.data.DataLoader(train_dataset, **train_kwargs)

    model = Net().to(device)
    optimizer = optim.Adadelta(model.parameters(), lr=args.lr)

    scheduler = StepLR(optimizer, step_size=1, gamma=args.gamma)
    for epoch in range(1, args.epochs + 1):
        train(args, model, device, train_loader, optimizer, epoch)
        # test(model, device, test_loader)
        scheduler.step()

    if args.save_model:
        torch.save(model.state_dict(), "mnist_cnn.pt")


if __name__ == '__main__':
    main()