from __future__ import absolute_import, division, print_function
import os
import sys
import glob
import numpy as np
import PIL.Image as pil
import matplotlib as mpl
import matplotlib.cm as cm
import torch
from torchvision import transforms, datasets
import networks
from layers import disp_to_depth
from utils import download_model_if_doesnt_exist
device = torch.device("cpu")
model_name = "mono+stereo_640x192"
model_path = os.path.join("models", "mono+stereo_640x192")
print("-> Loading model from ", model_path)
encoder_path = os.path.join(model_path, "encoder.pth")
depth_decoder_path = os.path.join(model_path, "depth.pth")
# LOADING PRETRAINED MODEL
print("   Loading pretrained encoder")
encoder = networks.ResnetEncoder(18, False)
loaded_dict_enc = torch.load(encoder_path, map_location=device)
# extract the height and width of image that this model was trained with
feed_height = loaded_dict_enc['height']
feed_width = loaded_dict_enc['width']
filtered_dict_enc = {k: v for k, v in loaded_dict_enc.items() if k in encoder.state_dict()}
encoder.load_state_dict(filtered_dict_enc)
encoder.to(device)
encoder.eval()
import onnx
import onnxruntime
def test_model_accuracy(export_model_name, raw_output, input):    
    ort_session = onnxruntime.InferenceSession(export_model_name)
    def to_numpy(tensor):
        return tensor.detach().cpu().numpy() if tensor.requires_grad else tensor.cpu().numpy()
    # compute ONNX Runtime output prediction
    ort_inputs = {ort_session.get_inputs()[0].name: to_numpy(input)}
    ort_outs = ort_session.run(None, ort_inputs)	
    # compare ONNX Runtime and PyTorch results
    np.testing.assert_allclose(to_numpy(raw_output), ort_outs[0], rtol=1e-03, atol=1e-05)
    print("Exported model has been tested with ONNXRuntime, and the result looks good!")		
def export_model(model, input, export_model_name):
    torch.onnx.export(model, input, export_model_name, verbose=False, export_params=True, opset_version=11)	
    onnx_model = onnx.load(export_model_name)    
    onnx.checker.check_model(onnx_model)
    graph_output = onnx.helper.printable_graph(onnx_model.graph)
    with open("graph_output.txt", mode="w") as fout:
        fout.write(graph_output)
export_model_name = "encoder_" + model_name + '.onnx'
input_image = pil.open("assets/runner.jpg").convert('RGB')
input_image = input_image.resize((feed_width, feed_height), pil.LANCZOS)
input_image = transforms.ToTensor()(input_image).unsqueeze(0)
features = encoder(input_image)
export_model(encoder, input_image, export_model_name)
test_model_accuracy(export_model_name, features, input_image)
#print("   Loading pretrained decoder")
depth_decoder = networks.DepthDecoder(
    num_ch_enc = encoder.num_ch_enc, scales=range(4))
loaded_dict = torch.load(depth_decoder_path, map_location=device)
depth_decoder.load_state_dict(loaded_dict)
depth_decoder.to(device)
depth_decoder.eval()
decoder_output = depth_decoder(features)
export_model_name = "decoder_" + model_name + '.onnx'
export_model(depth_decoder, features, export_model_name)
test_model_accuracy(export_model_name, decoder_output, features)

class ResnetEncoder(nn.Module):
    """Pytorch module for a resnet encoder
    """
    def __init__(self, num_layers, pretrained, num_input_images=1):
        super(ResnetEncoder, self).__init__()
        self.num_ch_enc = np.array([64, 64, 128, 256, 512])
        resnets = models.resnet18
        self.encoder = resnets(pretrained)	        
    def forward(self, input_image):         
        x = (input_image - 0.45) / 0.225
        x = self.encoder.conv1(x)
        x = self.encoder.bn1(x)		        
        output1 = self.encoder.relu(x)
        output2 = self.encoder.layer1(self.encoder.maxpool(output1))
        output3 = self.encoder.layer2(output2)
        output4 = self.encoder.layer3(output3)
        output5 = self.encoder.layer4(output4)        	
		
        return torch.cat((output1.flatten(), output2.flatten(), output3.flatten(), output4.flatten(), output5.flatten()))        

class DepthDecoder(nn.Module):
    def __init__(self, num_ch_enc, scales=range(4), num_output_channels=1, use_skips=True):
        super(DepthDecoder, self).__init__()
        self.num_output_channels = num_output_channels
        self.use_skips = use_skips
        self.upsample_mode = 'nearest'
        self.scales = scales
        self.num_ch_enc = num_ch_enc
        self.num_ch_dec = np.array([16, 32, 64, 128, 256])
        # decoder
        self.convs = OrderedDict()        
        for i in range(4, -1, -1):
            # upconv_0
            num_ch_in = self.num_ch_enc[-1] if i == 4 else self.num_ch_dec[i + 1]
            num_ch_out = self.num_ch_dec[i]
            self.convs[("upconv", i, 0)] = ConvBlock(num_ch_in, num_ch_out)
            # upconv_1
            num_ch_in = self.num_ch_dec[i]
            if self.use_skips and i > 0:
                num_ch_in += self.num_ch_enc[i - 1]
            num_ch_out = self.num_ch_dec[i]
            self.convs[("upconv", i, 1)] = ConvBlock(num_ch_in, num_ch_out)
        for s in self.scales:
            self.convs[("dispconv", s)] = Conv3x3(self.num_ch_dec[s], self.num_output_channels)
        self.decoder = nn.ModuleList(list(self.convs.values()))
        self.sigmoid = nn.Sigmoid()	        
    def forward(self, input_features):      
        #simplify the forward functions a lot, only post the codes which cause the issue  
        #the output processed by self.convs[("upconv", 4, 0)] will caused the error message
        return self.convs[("upconv", 4, 0)](input_features[2826240:2887680].reshape(1, 512, 6, 20))

std::string const model_folder("C:/Users/yyyy/programs/Qt/app_to_simplify_ml/"
                                   "cv_projects/clients/mobile_depth_estimation/monodepth2");
cv::dnn::Net net = cv::dnn::readNet(model_folder + "/decoder_mono+stereo_640x192.onnx");