Update app.py

app.py

@@ -131,14 +131,14 @@ def predict(query,data):
     scores = np.matmul(query_emb, doc_emb.transpose(1,0))[0].tolist()
     doc_score_pairs = list(zip(doc_text, scores, file_names))
     doc_score_pairs = sorted(doc_score_pairs, key=lambda x: x[1], reverse=True)
-    k = 5
+    
     probs_sum = 0
     probs = softmax(sorted(scores,reverse = True)[:k])
     table = {"Passage":[],"Answer":[],"Probabilities":[],"Source":[]}
     
     for i, (passage, _, names) in enumerate(doc_score_pairs[:k]):
         passage = passage.replace("\n","")
-        passage = passage.replace(" . "," ")
+        #passage = passage.replace(" . "," ")
         
         if probs[i] > 0.1 or (i < 3 and probs[i] > 0.05): #generate answers for more likely passages but no less than 2
             QA = {'question':query,'context':passage}
@@ -153,7 +153,7 @@ def predict(query,data):
             table["Passage"].append(passage)
             table["Answer"].append("no_answer_calculated")
             table["Probabilities"].append("P(p|q): {}".format(round(probs[i],5)))
-    
+    print(table)
     df = pd.DataFrame(table)
     print("time: "+ str(time.time()-start))
     
@@ -174,7 +174,7 @@ def predict(query,data):
   
         list_outputs.append(text)
     
-    return df
+    return list_outputs
 
 iface = gr.Interface(examples = [
         ["How high is the highest mountain?","China.pdf"],