Spaces:
Sleeping
Sleeping
Sasha commited on
Commit ·
81cc161
1
Parent(s): ae9e59e
fix: resolve timezone offset in VOD end_time calculations
Browse files
local_worker/backfill_chats_only.py
ADDED
|
@@ -0,0 +1,133 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import os
|
| 2 |
+
import sys
|
| 3 |
+
import time
|
| 4 |
+
import requests
|
| 5 |
+
from dotenv import load_dotenv
|
| 6 |
+
|
| 7 |
+
# Load local environment settings
|
| 8 |
+
load_dotenv()
|
| 9 |
+
|
| 10 |
+
# Import helper functions and config from vod_backfiller
|
| 11 |
+
sys.path.append(os.path.dirname(os.path.abspath(__file__)))
|
| 12 |
+
from vod_backfiller import get_vod_details, download_vod_chat, headers, API_URL, TWITCH_CHANNEL
|
| 13 |
+
|
| 14 |
+
def backfill_single_vod(vod_id):
|
| 15 |
+
print(f"\n=========================================================")
|
| 16 |
+
print(f" Начинаем обработку VOD: {vod_id}...")
|
| 17 |
+
print(f"=========================================================")
|
| 18 |
+
|
| 19 |
+
# 1. Fetch metadata
|
| 20 |
+
print("[Sync] Получение метаданных VOD...")
|
| 21 |
+
details = get_vod_details(vod_id)
|
| 22 |
+
if not details:
|
| 23 |
+
print(f"[Error] Не удалось получить информацию о VOD {vod_id}. Пропускаем.")
|
| 24 |
+
return False
|
| 25 |
+
|
| 26 |
+
title = details.get("title", "Архивный стрим")
|
| 27 |
+
created_at_raw = details.get("createdAt") # e.g. "2026-06-09T17:15:30Z"
|
| 28 |
+
length_seconds = details.get("lengthSeconds", 0)
|
| 29 |
+
|
| 30 |
+
# timezone-safe calculation using datetime
|
| 31 |
+
from datetime import datetime, timedelta
|
| 32 |
+
created_at = created_at_raw.replace("+00:00", "").replace("Z", "") + "Z"
|
| 33 |
+
dt_start = datetime.fromisoformat(created_at_raw.replace("Z", "+00:00"))
|
| 34 |
+
dt_end = dt_start + timedelta(seconds=length_seconds)
|
| 35 |
+
end_time = dt_end.strftime("%Y-%m-%dT%H:%M:%SZ")
|
| 36 |
+
|
| 37 |
+
print(f"Название: {title}")
|
| 38 |
+
print(f"Начало: {created_at}")
|
| 39 |
+
print(f"Длительность: {length_seconds // 3600}ч {(length_seconds % 3600) // 60}м")
|
| 40 |
+
|
| 41 |
+
# 2. Sync stream record in DB
|
| 42 |
+
print("\n[Sync] Регистрация сессии стрима на бэкенде...")
|
| 43 |
+
try:
|
| 44 |
+
url = f"{API_URL}/api/streams/sync-vod-direct"
|
| 45 |
+
payload = {
|
| 46 |
+
"twitchStreamId": f"vod-{vod_id}",
|
| 47 |
+
"title": title,
|
| 48 |
+
"category": "Архив",
|
| 49 |
+
"startTime": created_at,
|
| 50 |
+
"endTime": end_time
|
| 51 |
+
}
|
| 52 |
+
res = requests.post(url, json=payload, headers=headers, timeout=5)
|
| 53 |
+
if res.status_code != 200:
|
| 54 |
+
print(f"[Error] Не удалось синхронизировать сессию стрима: {res.status_code} - {res.text}")
|
| 55 |
+
return False
|
| 56 |
+
print("-> Сессия стрима успешно зарегистрирована в базе!")
|
| 57 |
+
except Exception as e:
|
| 58 |
+
print(f"[Error] Ошибка сети при обращении к бэкенду: {e}")
|
| 59 |
+
return False
|
| 60 |
+
|
| 61 |
+
# 3. Check for existing backfill status
|
| 62 |
+
chat_start_offset = 0
|
| 63 |
+
try:
|
| 64 |
+
status_url = f"{API_URL}/api/streams/backfill-status?twitchStreamId=vod-{vod_id}"
|
| 65 |
+
res = requests.get(status_url, headers=headers, timeout=5)
|
| 66 |
+
if res.status_code == 200:
|
| 67 |
+
status_data = res.json()
|
| 68 |
+
msg_count = status_data.get("messageCount", 0)
|
| 69 |
+
latest_msg_ts = status_data.get("latestMessageTimestamp")
|
| 70 |
+
if msg_count > 0 and latest_msg_ts:
|
| 71 |
+
from datetime import datetime
|
| 72 |
+
clean_ts = latest_msg_ts.replace('Z', '+00:00')
|
| 73 |
+
latest_epoch = datetime.fromisoformat(clean_ts).timestamp()
|
| 74 |
+
|
| 75 |
+
clean_created = created_at.replace('Z', '+00:00')
|
| 76 |
+
created_epoch = datetime.fromisoformat(clean_created).timestamp()
|
| 77 |
+
|
| 78 |
+
diff = int(latest_epoch - created_epoch)
|
| 79 |
+
if diff > 0:
|
| 80 |
+
chat_start_offset = diff
|
| 81 |
+
print(f"[Sync] Найден существующий чат в базе ({msg_count} сообщений). Возобновляем загрузку с секунды {chat_start_offset}...")
|
| 82 |
+
except Exception as e:
|
| 83 |
+
print(f"[Sync Warning] Не удалось получить статус дозаписи: {e}")
|
| 84 |
+
|
| 85 |
+
# 4. Download and Upload Chat comments
|
| 86 |
+
chat_comments = download_vod_chat(vod_id, start_offset=chat_start_offset)
|
| 87 |
+
if chat_comments:
|
| 88 |
+
print("\n[Sync] Отправка чата на сервер (пакетами по 100 сообщений)...")
|
| 89 |
+
batch_size = 100
|
| 90 |
+
for i in range(0, len(chat_comments), batch_size):
|
| 91 |
+
batch = chat_comments[i:i+batch_size]
|
| 92 |
+
try:
|
| 93 |
+
url = f"{API_URL}/api/log/messages"
|
| 94 |
+
res = requests.post(url, json={"messages": batch}, headers=headers, timeout=5)
|
| 95 |
+
if res.status_code == 200:
|
| 96 |
+
print(f"-> Отправлено сообщений: {i + len(batch)} / {len(chat_comments)}", end="\r")
|
| 97 |
+
else:
|
| 98 |
+
print(f"\n-> Ошибка отправки пакета: {res.status_code}")
|
| 99 |
+
except Exception as e:
|
| 100 |
+
print(f"\n-> Ошибка сети при отправке пакета: {e}")
|
| 101 |
+
print("\n-> Синхронизация чата успешно завершена!")
|
| 102 |
+
else:
|
| 103 |
+
print("\n-> Новых сообщений для загрузки не найдено.")
|
| 104 |
+
|
| 105 |
+
# 5. Mark Completed on Server
|
| 106 |
+
try:
|
| 107 |
+
url_mark = f"{API_URL}/api/streams/mark-backfilled"
|
| 108 |
+
requests.post(url_mark, json={"twitchStreamId": f"vod-{vod_id}"}, headers=headers, timeout=5)
|
| 109 |
+
print("-> Статус импорта успешно обновлен на сервере!")
|
| 110 |
+
except Exception as e:
|
| 111 |
+
print(f"[Warning] Не удалось отметить стрим как импортированный: {e}")
|
| 112 |
+
|
| 113 |
+
print(f"\n[Done] Обработка VOD {vod_id} завершена!")
|
| 114 |
+
return True
|
| 115 |
+
|
| 116 |
+
if __name__ == "__main__":
|
| 117 |
+
vod_ids = ['2798809640']
|
| 118 |
+
|
| 119 |
+
print("=========================================================")
|
| 120 |
+
print(" Автоматический импорт чатов для последних VOD ")
|
| 121 |
+
print("=========================================================")
|
| 122 |
+
print(f"Список VOD: {', '.join(vod_ids)}")
|
| 123 |
+
print(f"Канал: {TWITCH_CHANNEL}")
|
| 124 |
+
print(f"Бэкенд: {API_URL}")
|
| 125 |
+
print("=========================================================")
|
| 126 |
+
|
| 127 |
+
for vid in vod_ids:
|
| 128 |
+
try:
|
| 129 |
+
backfill_single_vod(vid)
|
| 130 |
+
except Exception as e:
|
| 131 |
+
print(f"[Error] Критическая ошибка при обработке VOD {vid}: {e}")
|
| 132 |
+
|
| 133 |
+
print("\nВсе три VOD успешно обработаны!")
|
local_worker/vod_backfiller.py
CHANGED
|
@@ -368,11 +368,12 @@ if __name__ == "__main__":
|
|
| 368 |
created_at_raw = details.get("createdAt") # e.g. "2026-06-09T17:15:30Z"
|
| 369 |
length_seconds = details.get("lengthSeconds", 0)
|
| 370 |
|
| 371 |
-
#
|
|
|
|
| 372 |
created_at = created_at_raw.replace("+00:00", "").replace("Z", "") + "Z"
|
| 373 |
-
|
| 374 |
-
|
| 375 |
-
end_time =
|
| 376 |
|
| 377 |
print(f"\nНазвание: {title}")
|
| 378 |
print(f"Начало: {created_at}")
|
|
|
|
| 368 |
created_at_raw = details.get("createdAt") # e.g. "2026-06-09T17:15:30Z"
|
| 369 |
length_seconds = details.get("lengthSeconds", 0)
|
| 370 |
|
| 371 |
+
# timezone-safe calculation using datetime
|
| 372 |
+
from datetime import datetime, timedelta
|
| 373 |
created_at = created_at_raw.replace("+00:00", "").replace("Z", "") + "Z"
|
| 374 |
+
dt_start = datetime.fromisoformat(created_at_raw.replace("Z", "+00:00"))
|
| 375 |
+
dt_end = dt_start + timedelta(seconds=length_seconds)
|
| 376 |
+
end_time = dt_end.strftime("%Y-%m-%dT%H:%M:%SZ")
|
| 377 |
|
| 378 |
print(f"\nНазвание: {title}")
|
| 379 |
print(f"Начало: {created_at}")
|