Update app.py

app.py

@@ -17,18 +17,6 @@ dataset2 = fs.get_feature_group(name="daily_document_info").read()
 df2 = dataset2
 topics = df['topic'].unique()
 
-readable_topics_dic = dict()
-readable_topics = []
-for t in topics.tolist():
-    selected_data = df[df['topic'] == t]
-    keywords = selected_data['keywords'][selected_data.index[0]]
-    input = "With an answer of only few words (less than 5) give me a word or expression that charaterise the best this set of words as if it was the description of a theme: " + str(keywords)
-    new_topic = gpt_predict(input) 
-    readable_topics += new_topic
-    readable_topics_dic[new_topic] = t
-    break
-print(readable_topics)
-
 def gpt_predict(inputs, request:gr.Request=gr.State([]), top_p = 1, temperature = 1, chat_counter = 0,history =[]):
     payload = {
         "model": MODEL,
@@ -111,6 +99,20 @@ def gpt_predict(inputs, request:gr.Request=gr.State([]), top_p = 1, temperature
         print (f'error found: {e}')
     return partial_words
 
+readable_topics_dic = dict()
+readable_topics = []
+for t in topics.tolist():
+    selected_data = df[df['topic'] == t]
+    keywords = selected_data['keywords'][selected_data.index[0]]
+    input = "With an answer of only few words (less than 5) give me a word or expression that charaterise the best this set of words as if it was the description of a theme: " + str(keywords)
+    new_topic = gpt_predict(input) 
+    readable_topics += new_topic
+    readable_topics_dic[new_topic] = t
+    break
+print(readable_topics)
+
+
+
 def display_topics(topic):
     topic = readable_topics_dic[topic]
     # Filter DataFrame based on the selected topic