import os

import argparse

import torch

from networks.vnet_sdf import VNet

from utils.test_patch_sass import test_all_case

parser = argparse.ArgumentParser()

parser.add_argument('--dataset_name', type=str,  default='LA', help='dataset_name')

parser.add_argument('--root_path', type=str, default='/data/omnisky/postgraduate/Yb/data_set/LASet/data', help='Name of Experiment')

parser.add_argument('--exp', type=str,  default='vnet', help='exp_name')

parser.add_argument('--model', type=str,  default='vnet_DTC', help='model_name')

parser.add_argument('--gpu', type=str,  default='1', help='GPU to use')

parser.add_argument('--labelnum', type=int, default=11, help='labeled data')

parser.add_argument('--iter', type=int,  default=6000, help='model iteration')

parser.add_argument('--detail', type=int,  default=1, help='print metrics for every samples?')

parser.add_argument('--nms', type=int, default=0, help='apply NMS post-procssing?')

FLAGS = parser.parse_args()

os.environ['CUDA_VISIBLE_DEVICES'] = FLAGS.gpu

snapshot_path = "../model/{}".format(FLAGS.model)

num_classes = 2

test_save_path = "model/{}_{}_{}_labeled/{}_predictions/".format(FLAGS.dataset_name, FLAGS.exp, FLAGS.labelnum, FLAGS.model)

if not os.path.exists(test_save_path):

    os.makedirs(test_save_path)

print(test_save_path)

with open(FLAGS.root_path + '/../test.list', 'r') as f:

    image_list = f.readlines()

image_list = [FLAGS.root_path + "/" + item.replace('\n', '') + "/mri_norm2.h5" for item in

              image_list]

def test_calculate_metric(epoch_num):

    net = VNet(n_channels=1, n_classes=num_classes-1, normalization='batchnorm', has_dropout=False).cuda()

    save_mode_path = 'model/LA_vnet_12_labeled/sassnet_label12/iter_5200_dice_0.8954771273472677.pth'

    net.load_state_dict(torch.load(save_mode_path))

    print("init weight from {}".format(save_mode_path))

    net.eval()

    avg_metric = test_all_case(net, image_list, num_classes=num_classes,

                           patch_size=(112, 112, 80), stride_xy=18, stride_z=4,

                           save_result=False, test_save_path=test_save_path,

                           metric_detail=FLAGS.detail, nms=FLAGS.nms)

    return avg_metric

if __name__ == '__main__':

    metric = test_calculate_metric(FLAGS.iter) #6000

    print(metric)

# python test_LA.py --model 0214_re01 --gpu 0