fix(weibo): 修复 RFC2822 时间转换丢弃时区偏移导致时间戳晚 8 小时

rfc2822_to_timestamp 先用 %z 解析出带偏移的 datetime,再用
replace(tzinfo=timezone.utc) 处理。replace 只是覆盖 tzinfo 而不做换算,
+0800 的时间被当成 UTC,算出的时间戳比真实时间晚 8 小时。

改用 astimezone(timezone.utc) 做真正的时区换算,与紧邻上方的
rfc2822_to_china_datetime 保持一致。

影响 store/weibo 落库的 create_time 字段(帖子与评论各一处)。同一条记录里
create_time 与 create_date_time 会相差 8 小时,跨日内容的日期还会整体偏移
一天。

补两条离线单测:一条断言 +0800 输入对应的时间戳,一条断言 create_time 与
create_date_time 指向同一时刻。两条测试在修复前均失败。
This commit is contained in:
ztcools
2026-07-30 15:07:07 +08:00
parent 17f66121e0
commit 04631ba8e4
2 changed files with 14 additions and 1 deletions

View File

@@ -47,3 +47,16 @@ async def test_convert_browser_context_cookies_uses_url_filter():
browser_context.cookies.assert_awaited_once_with(urls=["https://www.douyin.com"])
assert cookie_str == "sessionid=abc"
assert cookie_dict == {"sessionid": "abc"}
def test_rfc2822_to_timestamp_converts_utc_offset():
# 2023-12-23 17:12:54 +0800 == 2023-12-23 09:12:54 UTC == 1703322774
assert utils.rfc2822_to_timestamp("Sat Dec 23 17:12:54 +0800 2023") == 1703322774
def test_rfc2822_to_timestamp_agrees_with_china_datetime():
rfc2822_time = "Sat Dec 23 17:12:54 +0800 2023"
assert utils.rfc2822_to_timestamp(rfc2822_time) == int(
utils.rfc2822_to_china_datetime(rfc2822_time).timestamp()
)

View File

@@ -118,7 +118,7 @@ def rfc2822_to_timestamp(rfc2822_time):
dt_object = datetime.strptime(rfc2822_time, rfc2822_format)
# Convert datetime object to UTC time
dt_utc = dt_object.replace(tzinfo=timezone.utc)
dt_utc = dt_object.astimezone(timezone.utc)
# Calculate Unix timestamp from UTC time
timestamp = int(dt_utc.timestamp())