mirror of
https://github.com/NanmiCoder/MediaCrawler.git
synced 2026-06-06 09:57:25 +08:00
support bilibili level two comment
This commit is contained in:
@@ -101,10 +101,12 @@ class BilibiliClient(AbstractApiClient):
|
||||
check_login_uri = "/x/web-interface/nav"
|
||||
response = await self.get(check_login_uri)
|
||||
if response.get("isLogin"):
|
||||
utils.logger.info("[BilibiliClient.pong] Use cache login state get web interface successfull!")
|
||||
utils.logger.info(
|
||||
"[BilibiliClient.pong] Use cache login state get web interface successfull!")
|
||||
ping_flag = True
|
||||
except Exception as e:
|
||||
utils.logger.error(f"[BilibiliClient.pong] Pong bilibili failed: {e}, and try to login again...")
|
||||
utils.logger.error(
|
||||
f"[BilibiliClient.pong] Pong bilibili failed: {e}, and try to login again...")
|
||||
ping_flag = False
|
||||
return ping_flag
|
||||
|
||||
@@ -185,13 +187,21 @@ class BilibiliClient(AbstractApiClient):
|
||||
|
||||
result = []
|
||||
is_end = False
|
||||
next_page =0
|
||||
next_page = 0
|
||||
while not is_end:
|
||||
comments_res = await self.get_video_comments(video_id, CommentOrderType.DEFAULT, next_page)
|
||||
curson_info: Dict = comments_res.get("cursor")
|
||||
cursor_info: Dict = comments_res.get("cursor")
|
||||
comment_list: List[Dict] = comments_res.get("replies", [])
|
||||
is_end = curson_info.get("is_end")
|
||||
next_page = curson_info.get("next")
|
||||
is_end = cursor_info.get("is_end")
|
||||
next_page = cursor_info.get("next")
|
||||
if is_fetch_sub_comments:
|
||||
for comment in comment_list:
|
||||
comment_id = comment['rpid']
|
||||
if (comment.get("rcount", 0) > 0):
|
||||
{
|
||||
await self.get_video_all_level_two_comments(
|
||||
video_id, comment_id, CommentOrderType.DEFAULT, 10, crawl_interval, callback)
|
||||
}
|
||||
if callback: # 如果有回调函数,就执行回调函数
|
||||
await callback(video_id, comment_list)
|
||||
await asyncio.sleep(crawl_interval)
|
||||
@@ -200,3 +210,61 @@ class BilibiliClient(AbstractApiClient):
|
||||
continue
|
||||
# todo handle get sub comments
|
||||
return result
|
||||
|
||||
async def get_video_all_level_two_comments(self,
|
||||
video_id: str,
|
||||
level_one_comment_id: int,
|
||||
order_mode: CommentOrderType,
|
||||
ps: int = 10,
|
||||
crawl_interval: float = 1.0,
|
||||
callback: Optional[Callable] = None,
|
||||
) -> Dict:
|
||||
"""
|
||||
get video all level two comments for a level one comment
|
||||
:param video_id: 视频 ID
|
||||
:param level_one_comment_id: 一级评论 ID
|
||||
:param order_mode:
|
||||
:param ps: 一页评论数
|
||||
:param crawl_interval:
|
||||
:param callback:
|
||||
:return:
|
||||
"""
|
||||
|
||||
pn = 0
|
||||
while True:
|
||||
result = await self.get_video_level_two_comments(
|
||||
video_id, level_one_comment_id, 0, ps, order_mode)
|
||||
comment_list: List[Dict] = result.get("replies", [])
|
||||
if callback: # 如果有回调函数,就执行回调函数
|
||||
await callback(video_id, comment_list)
|
||||
await asyncio.sleep(crawl_interval)
|
||||
if (int(result["page"]["count"]) <= (pn+1) * ps):
|
||||
break
|
||||
|
||||
pn += 1
|
||||
|
||||
async def get_video_level_two_comments(self,
|
||||
video_id: str,
|
||||
level_one_comment_id: int,
|
||||
pn: int,
|
||||
ps: int,
|
||||
order_mode: CommentOrderType,
|
||||
) -> Dict:
|
||||
"""get video level two comments
|
||||
:param video_id: 视频 ID
|
||||
:param level_one_comment_id: 一级评论 ID
|
||||
:param order_mode: 排序方式
|
||||
|
||||
:return:
|
||||
"""
|
||||
uri = "/x/v2/reply/reply"
|
||||
post_data = {
|
||||
"oid": video_id,
|
||||
"mode": order_mode.value,
|
||||
"type": 1,
|
||||
"ps": ps,
|
||||
"pn": pn,
|
||||
"root": level_one_comment_id,
|
||||
}
|
||||
result = await self.get(uri, post_data)
|
||||
return result
|
||||
|
||||
Reference in New Issue
Block a user