Bläddra i källkod

generate_data_set.py

luojunhui 5 dagar sedan
förälder
incheckning
eba4d6ff64
1 ändrade filer med 3 tillägg och 3 borttagningar
  1. 3 3
      generate_data_set.py

+ 3 - 3
generate_data_set.py

@@ -206,7 +206,7 @@ def generate_push_dataset():
     filter_conversation = [i for i in data_set if len(json.loads(i['conversation'])) >= 20]
     print(len(filter_conversation))
 
-    samples =random.sample(filter_conversation, 100)
+    samples =random.sample(filter_conversation, 300)
 
     # init message push agent
     agent = MessagePushAgent()
@@ -228,7 +228,7 @@ def generate_push_dataset():
         )
         conversation = sorted(conversation, key=lambda i: i['timestamp'], reverse=False)
         last_timestamp = int(conversation[-1]["timestamp"])
-        push_time = last_timestamp + 48 * 3600
+        push_time = last_timestamp + 24 * 3600
         push_dt =  datetime.fromtimestamp(push_time).strftime('%Y-%m-%d %H:%M:%S')
         try:
             push_message = agent.generate_message(
@@ -255,7 +255,7 @@ def generate_push_dataset():
                 values (%s, %s, %s, %s, %s, %s, %s, %s, %s);
             """
             mysql_client.execute(insert_query, args=(
-                '3',
+                '2',
                 sample["staff_id"],
                 sample["user_id"],
                 '2025-06-16',