Update app.py

app.py

@@ -29,25 +29,20 @@ if text:
     result = nlp(text+' '+nlp.tokenizer.mask_token)
     data_load_state.text('')
     
-    predicted_embeddings = model.encode(result['sequence'], convert_to_tensor=True)
-    semantic_history_embeddings = model.encode(semantic_text.spllit(','), convert_to_tensor=True)
-    
-    cosine_scores = util.cos_sim(embeddings1, embeddings2)
+    if len(semantic_text):   
+        predicted_embeddings = model.encode(result['sequence'], convert_to_tensor=True)
+        semantic_history_embeddings = model.encode(semantic_text.spllit(','), convert_to_tensor=True)
+        cosine_scores = util.cos_sim(predicted_embeddings, semantic_history_embeddings)
     
     for index, r in enumerate(result):
-        result[index]['score']=cosine_scores[index][index]
+        if len(semantic_text):
+            result[index]['score']*=cosine_scores[index][index]
         if r['token_str'].lower().strip() in history_keyword_text.lower().strip() and len(r['token_str'].lower().strip())>1:
             #found from history, then increase the score of tokens
             result[index]['score']*=HISTORY_WEIGHT
             
-    
-    
     #sort the results        
     df=pd.DataFrame(result).sort_values(by='score', ascending=False)
     
-    
-
-
-    
     #show the results as a table
     st.table(df)