Update to topN.

results/dis_dru_the_20_table.csv

-,dis,dis name,dru,dru name,pred
-0,C0751137,Craniofacial Pain,CHEMBL1490,TRIHEXYPHENIDYL,1.0
-1,C0086237,"Epilepsy, Cryptogenic",CHEMBL641,ATOMOXETINE,1.0
-2,C0270824,Visual epilepsy,CHEMBL3833361,PROMETHAZINE TEOCLATE,1.0
-3,C3151568,"NEPHROTIC SYNDROME, TYPE 4",CHEMBL121,ROSIGLITAZONE,1.0
-4,C0086237,"Epilepsy, Cryptogenic",CHEMBL640,PROCAINAMIDE,1.0
-5,C0270824,Visual epilepsy,CHEMBL3833362,ZINC OLEATE,1.0
-6,C0270824,Visual epilepsy,CHEMBL3833364,HYDRARGAPHEN,1.0
-7,C0270824,Visual epilepsy,CHEMBL3833368,RUCAPARIB CAMSYLATE,1.0
-8,C0270824,Visual epilepsy,CHEMBL3833369,FISH OIL,1.0
-9,C0270824,Visual epilepsy,CHEMBL3833373,AVELUMAB,1.0
-10,C0270824,Visual epilepsy,CHEMBL3833381,MERBROMIN,1.0
-11,C0270824,Visual epilepsy,CHEMBL3833382,PRAJMALIUM,1.0
-12,C0270824,Visual epilepsy,CHEMBL3833383,GALLIUM DOTATATE GA-68,1.0
-13,C0270824,Visual epilepsy,CHEMBL3833311,DEXTROMORAMIDE TARTRATE,1.0
-14,C0270824,Visual epilepsy,CHEMBL3833388,PRAJMALIUM BITARTRATE,1.0
-15,C0270824,Visual epilepsy,CHEMBL3833389,DEXTROMORAMIDE,1.0
-16,C0270824,Visual epilepsy,CHEMBL3833393,EMICIZUMAB,1.0
-17,C0270824,Visual epilepsy,CHEMBL3833401,ALUMINUM CHLORIDE,1.0
-18,C0338656,Cognitive Dysfunction,CHEMBL3039594,GENTAMICIN SULFATE,1.0
-19,C0162323,Polyarthritis,CHEMBL1003,CLAVULANATE POTASSIUM,1.0

testRepoDB.py

@@ -147,9 +147,9 @@ def metrics(model):
     pure_predictions = [item for sublist in [preds, predsN] for item in sublist]
     labels = torch.tensor([item for sublist in [labels1, labels2] for item in sublist])
 
-    fpr, tpr, label1 = plot_roc(labels, pure_predictions, ('disorder', 'dis_dru_the', 'drug'), "dmsr-f/",
+    fpr, tpr, label1 = plot_roc(labels, pure_predictions, ('disorder', 'dis_dru_the', 'drug'), "dmsr/",
                                 "RepoDB")
-    recall, precision, label2 = plot_prc(labels, pure_predictions, ('disorder', 'dis_dru_the', 'drug'), "dmsr-f/",
+    recall, precision, label2 = plot_prc(labels, pure_predictions, ('disorder', 'dis_dru_the', 'drug'), "dmsr/",
                                          "RepoDB")
 
     plotMetrics(fpr, tpr, label1, recall, precision, label2)

testRepoDBWeightsAndBiases.py

@@ -159,9 +159,9 @@ def metrics(model):
     pure_predictions = [item for sublist in [preds, predsN] for item in sublist]
     labels = torch.tensor([item for sublist in [labels1, labels2] for item in sublist])
 
-    fpr, tpr, label1 = plot_roc(labels, pure_predictions, ('disorder', 'dis_dru_the', 'drug'), "dmsr-f/",
+    fpr, tpr, label1 = plot_roc(labels, pure_predictions, ('disorder', 'dis_dru_the', 'drug'), "dmsr/",
                                 "RepoDB")
-    recall, precision, label2 = plot_prc(labels, pure_predictions, ('disorder', 'dis_dru_the', 'drug'), "dmsr-f/",
+    recall, precision, label2 = plot_prc(labels, pure_predictions, ('disorder', 'dis_dru_the', 'drug'), "dmsr/",
                                          "RepoDB")
 
     plotMetrics(fpr, tpr, label1, recall, precision, label2)

topN.py

@@ -7,8 +7,8 @@ from deepsnap.dataset import GraphDataset
 from datetime import datetime
 from deepsnap.hetero_gnn import HeteroSAGEConv
 
-
 device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')  # It defines whether to execute on cpu or gpu.
+device = 'cpu'
 constructor = heterograph_construction.DISNETConstructor(device=device)  # Graph constructor.
 edge = ('disorder', 'dis_dru_the', 'drug')  # Graph edge type to study.
 n = 200000  # Number of new predictions.
@@ -19,8 +19,10 @@ Input:
     original: Graph.
     pred: Edge predictions.
 """
+
+
 def filterPreds(original, pred):
-    headsO = original.edge_index[edge][0, :].long() # Heads of the original edges of the graph.
+    headsO = original.edge_index[edge][0, :].long()  # Heads of the original edges of the graph.
 
     new = []
     for i, elem in enumerate(pred):
@@ -28,11 +30,11 @@ def filterPreds(original, pred):
         head = i
         tail = torch.arange(0, len(pred_labels))  # All tails.
         indexH = ((headsO == head).nonzero(as_tuple=True)[0])  # Index of those heads originally present in the graph.
-        #Check
-        print(len(tail))
-        for index in indexH:
-            tail = tail[tail != index]  # Just get those tails not present in the original graph.
-        print(len(tail))
+
+        # Check
+        for t in original.edge_index[edge][1, indexH]:
+            tail = tail[tail != t]  # Just get those tails not present in the original graph.
+
         new.append([head, tail, pred_labels[tail].cpu().detach().numpy()])  # New predictions are appended.
 
     return new
@@ -47,24 +49,29 @@ Input:
 Output:
     Dataframe containing the top n predictions ordered and decoded.
 """
+
+
 def getTopN(model, dataloader, n):
     print("     Looking for new edges.")
     for batch in zip(dataloader):
-        batch.to(device)
+        batch = batch[0]
         pred = model.predict_all(batch)  # Predict all edges.
-        
+
         new = filterPreds(batch, pred)  # Filter those edges present in the original graph.
-      
+
     print("     Decoding predictions, this may take a while.")
     return constructor.decodePredictions(new, edge[1], n)
 
+
 """
 It gets the heterograph object and its conversion to dataloader.
 Output:
     The heterograph and its dataloader.
 """
+
+
 def getOriginal():
-    hetero, _ = constructor.DISNETHeterograph()
+    hetero, _ = constructor.DISNETHeterograph(full=True, withoutRepoDB=False)
     dataset = GraphDataset(
         [hetero],
         task='link_pred',
@@ -76,17 +83,20 @@ def getOriginal():
     )
     return dataset_loader, hetero
 
+
 """
 It wraps all the necessary calls to get the top n predictions of the DMSR model.
 """
+
+
 def dmsr():
     # Necessary instantiations.
     original, hetero = getOriginal()
-    convs = generate_convs_link_pred_layers(hetero, HeteroSAGEConv, 107)
-    model = HeteroGNN(convs, hetero, 107, 0.8).to(device).to(device)
+    convs = generate_convs_link_pred_layers(hetero, HeteroSAGEConv, 31)
+    model = HeteroGNN(convs, hetero, 31, 0.5).to(device)
 
     # Load and prepare for inference model.
-    model.load_state_dict(torch.load("./models/dmsr", map_location=torch.device(device)))
+    model.load_state_dict(torch.load("./models/dmsrC", map_location=torch.device(device)))
     model = model.to(device)
     model.eval()
 
@@ -95,6 +105,7 @@ def dmsr():
     _ = getTopN(model, original, n)
     print("Finished getting top", n, "at", datetime.now().strftime("%H:%M:%S"))
 
+
 if __name__ == '__main__':
     dmsr()