Upload 2 files

dataset.py

@@ -149,7 +149,7 @@ class Graph_Regression_Dataset_test(object):
         adjoin_matrix.set_shape([None,None])
         y.set_shape([None])
         return x, adjoin_matrix , y
-    
+   
 class predict_smiles(object):
     def __init__(self,smiles ,normalize=False,max_len=1000,addH=True):
      
@@ -165,13 +165,15 @@ class predict_smiles(object):
             self.min = self.df[self.label_field].min()
             self.df[self.label_field] = (self.df[self.label_field]-self.min)/(self.max-self.min)-0.5
             self.value_range = self.max-self.min
+            
     def numerical_smiles(self, atoms_list,adj,label):
                 
         atom = np.array(atoms_list)            
         atoms_list = []
         for i in atom:
             if i not in [' ']:
-                atoms_list.append(i)
+                 
+                atoms_list.append(str(i,encoding='utf-8'))
         label = np.array(label)
        
         adj = np.array(adj)

utils.py

@@ -646,12 +646,12 @@ def gen_adj(shape,edges,length):
     
     adj=edges
     e = shape
-    ones = np.zeros([e,e])
+    ones = np.eye(e)
 
     #for i in range(e):
     for i in range (len(length)):
         if adj[i,0] != adj[i,1]:
-            ones[adj[i,0],adj[i,1]]=format(float(length[i] ), '.3f')
+            ones[adj[i,0],adj[i,1]]=(float(length[i] ))
                    
     return ones
 
@@ -659,38 +659,40 @@ def gen_adj(shape,edges,length):
 if __name__ == "__main__":
     import pandas as pd 
     from tqdm import tqdm
-    f = pd.read_csv (r"data/reg/train3.csv")
-    re = []
-    pce = f['PCE']
-    for ind,smile in enumerate ( f.iloc[:,1]):
-        print(ind)
-        atom,adj = mol_to_geognn_graph_data_MMFF3d(smile)
-        np.save('data/reg/train/adj'+str(ind)+'.npy',np.array(adj))
-        re.append([atom,'data/reg/train/adj'+str(ind)+'.npy',pce[ind] ])
-    r = pd.DataFrame(re)
-    r.to_csv('data/reg/train/train.csv')
-    re = []
-
-    f = pd.read_csv(r'data/reg/test3.csv')
-    re = []
-    pce = f['PCE']
+    f = pd.read_csv (r"J:\screenacc\new4.csv")
+    # re = []
+    # pce = f['PCE']
+    # for ind,smile in enumerate ( f.iloc[:,1]):
+    #     print(ind)
+    #     atom,adj = mol_to_geognn_graph_data_MMFF3d(smile)
+    #     np.save('data/reg/train/adj'+str(ind)+'.npy',np.array(adj))
+    #     re.append([atom,'data/reg/train/adj'+str(ind)+'.npy',pce[ind] ])
+    # r = pd.DataFrame(re)
+    # r.to_csv('data/reg/train/train.csv')
+    # re = []
+
+    # f = pd.read_csv(r'data/reg/test3.csv')
+    # re = []
+    # pce = f['PCE']
    
-    for ind,smile in enumerate ( f.iloc[:,1]):
-        print(ind)
-        atom,adj = mol_to_geognn_graph_data_MMFF3d(smile)
-        np.save('data/reg/test/adj'+str(ind)+'.npy',np.array(adj))
-        re.append([atom,'data/reg/test/adj'+str(ind)+'.npy',pce[ind] ])
-    r = pd.DataFrame(re)
-    r.to_csv('data/reg/test/test.csv')
+    # for ind,smile in enumerate ( f.iloc[:,1]):
+    #     print(ind)
+    #     atom,adj = mol_to_geognn_graph_data_MMFF3d(smile)
+    #     np.save('data/reg/test/adj'+str(ind)+'.npy',np.array(adj))
+    #     re.append([atom,'data/reg/test/adj'+str(ind)+'.npy',pce[ind] ])
+    # r = pd.DataFrame(re)
+    # r.to_csv('data/reg/test/test.csv')
     
-    f = pd.read_csv(r'val.csv')
+    # f = pd.read_csv(r'val.csv')
     re = []
     pce = f['PCE']        
    
-    for ind,smile in enumerate ( f.iloc[:,1]):
+    for ind,smile in enumerate ( f.iloc[ 22000: ,0]):
+        
+        ind = ind + 22000
         print(ind)
         atom,adj = mol_to_geognn_graph_data_MMFF3d(smile)
         np.save('data/reg/val/adj'+str(ind)+'.npy',np.array(adj))
         re.append([atom,'data/reg/val/adj'+str(ind)+'.npy',pce[ind] ])
     r = pd.DataFrame(re)
-    r.to_csv('data/reg/val/val.csv')
+    r.to_csv('data/reg/val/val22000.csv')