import gradio import torch.nn as nn import torch from torch_geometric.loader import DataLoader import utils.clean_data as cd import utils.shape_features as sf import utils.node_features as nf import utils.edge_features as ef # from datetime import datetime # start_time = datetime.now() node_model_path = 'utils/emb_model/Node_64.pt' edge_model_path = 'utils/emb_model/Edge_64.pt' class InfoGraph(nn.Module): def __init__(self, hidden_dim, num_gc_layers, alpha=0.5, beta=1., gamma=.1): super(InfoGraph, self).__init__() self.alpha = alpha self.beta = beta self.gamma = gamma self.prior = False self.embedding_dim = mi_units = hidden_dim * num_gc_layers self.encoder = Encoder(dataset_num_features, hidden_dim, num_gc_layers) self.local_d = FF(self.embedding_dim) self.global_d = FF(self.embedding_dim) # self.local_d = MI1x1ConvNet(self.embedding_dim, mi_units) # self.global_d = MIFCNet(self.embedding_dim, mi_units) if self.prior: self.prior_d = PriorDiscriminator(self.embedding_dim) self.init_emb() def init_emb(self): initrange = -1.5 / self.embedding_dim for m in self.modules(): if isinstance(m, nn.Linear): torch.nn.init.xavier_uniform_(m.weight.data) if m.bias is not None: m.bias.data.fill_(0.0) def forward(self, x, edge_index, batch, num_graphs): # batch_size = data.num_graphs if x is None: x = torch.ones(batch.shape[0]).to(device) y, M = self.encoder(x, edge_index, batch) g_enc = self.global_d(y) l_enc = self.local_d(M) mode='fd' measure='JSD' local_global_loss = local_global_loss_(l_enc, g_enc, edge_index, batch, measure) if self.prior: prior = torch.rand_like(y) term_a = torch.log(self.prior_d(prior)).mean() term_b = torch.log(1.0 - self.prior_d(y)).mean() PRIOR = - (term_a + term_b) * self.gamma else: PRIOR = 0 return local_global_loss + PRIOR def outline_embedding(wkt, wall): wall_f, wkt_f = cd.read_wall_wkt(wall, wkt) apa_wall, apa_geo = cd.clean_geometry(wall_f, wkt_f) apa_geo = apa_geo apa_line = apa_geo.boundary apa_wall_O = cd.exterior_wall(apa_line, apa_wall) apa_coor = cd.geo_coor(apa_geo) xarr4cv, yarr4cv = apa_geo.exterior.coords.xy x4cv = xarr4cv.tolist() y4cv = yarr4cv.tolist() scale = 100000 xmin_abs = abs(min(x4cv)) ymin_abs = abs(min(y4cv)) p_4_cv = cd.points4cv(x4cv, y4cv, xmin_abs, ymin_abs, scale) grid_points = cd.gridpoints(apa_geo, 1) Dir_S_longestedge, Dir_N_longestedge, Dir_W_longestedge, Dir_E_longestedge, Dir_S_max, Dir_N_max, Dir_W_max, Dir_E_max, Facade_length, Facade_ratio = sf.wall_direction_ratio(apa_line, apa_wall) Perimeter = sf.apartment_perimeter(apa_geo) Area = sf.apartment_area(apa_geo) BBox_width_x, BBox_height_y, Aspect_ratio, Extent, ULC_x, ULC_y, LRC_x, LRC_y = sf.boundingbox_features(apa_geo) Max_diameter = sf.max_diameter(apa_geo) Fractality = sf.fractality(apa_geo) Circularity = sf.circularity(apa_geo) Outer_radius = sf.outer_radius(p_4_cv, xmin_abs, ymin_abs, scale) Inner_radius = sf.inner_radius(apa_geo, apa_line) Dist_mean, Dist_sigma, Roundness = sf.roundness_features(apa_line) Compactness = sf.compactness(apa_geo) Equivalent_diameter = sf.equivalent_diameter(apa_geo) Shape_membership_index = sf.shape_membership_index(apa_line) Convexity, Hull_geo = sf.convexity(p_4_cv, apa_geo, xmin_abs, ymin_abs, scale) Rectangularity, Rect_phi, Rect_width, Rect_height = sf.rectangle_features(p_4_cv, apa_geo, xmin_abs, ymin_abs, scale) Squareness = sf.squareness(apa_geo) Moment_index = sf.moment_index(apa_geo, Convexity, Compactness) nDetour_index = sf.ndetour_index(apa_geo, Hull_geo) nCohesion_index = sf.ncohesion_index(apa_geo, grid_points) nProximity_index, nSpin_index = sf.nproximity_nspin_index(apa_geo, grid_points) nExchange_index = sf.nexchange_index(apa_geo) nPerimeter_index = sf.nperimeter_index(apa_geo) nDepth_index = sf.ndepth_index(apa_geo, apa_line, grid_points) nGirth_index = sf.ngirth_index(apa_geo, Inner_radius) nRange_index = sf.nrange_index(apa_geo, Outer_radius) nTraversal_index = sf.ntraversal_index(apa_geo, apa_line) shape = [Dir_S_longestedge, Dir_N_longestedge, Dir_W_longestedge, Dir_E_longestedge, Dir_S_max, Dir_N_max, Dir_W_max, Dir_E_max, Facade_length, Facade_ratio, Perimeter, Area, BBox_width_x, BBox_height_y, Aspect_ratio, Extent, ULC_x, ULC_y, LRC_x, LRC_y, Max_diameter, Fractality, Circularity, Outer_radius, Inner_radius, Dist_mean, Dist_sigma, Roundness, Compactness, Equivalent_diameter, Shape_membership_index, Convexity, Rectangularity, Rect_phi, Rect_width, Rect_height, Squareness, Moment_index, nDetour_index, nCohesion_index, nProximity_index, nExchange_index, nSpin_index, nPerimeter_index, nDepth_index, nGirth_index, nRange_index, nTraversal_index] shape = [float(i) for i in shape] node_graph = nf.node_graph(apa_coor, apa_geo) node_model = torch.load(node_model_path) node_model.eval() node_dataloader = DataLoader(node_graph, batch_size=1) node_emb = node_model.encoder.get_embeddings(node_dataloader) node = node_emb[0].tolist() edge_graph = ef.edge_graph(apa_line, apa_wall) edge_model = torch.load(edge_model_path) edge_model.eval() edge_dataloader = DataLoader(edge_graph, batch_size=1) edge_emb = edge_model.encoder.get_embeddings(edge_dataloader) edge = edge_emb[0].tolist() json = {"edge": edge, "shape": shape, "node": node} return json gradio_interface = gradio.Interface(fn=outline_embedding, inputs = [gradio.Textbox(type="text", label="wkt", placeholder="wkt"), gradio.Textbox(type="text", label="wall", placeholder="wall")], outputs = "json", title="outline embedding") # end_time = datetime.now() # print('Duration: {}'.format(end_time - start_time)) # api_open=True, gradio_interface.queue(max_size=5, status_update_rate="auto") gradio_interface.launch(show_error=True, enable_queue=True)