parent
73478149ed
commit
3b8b572b09
@ -0,0 +1,22 @@
|
||||
import requests
|
||||
from bs4 import BeautifulSoup
|
||||
|
||||
|
||||
def get_video_ids(keyword, num_videos):
|
||||
video_ids = []
|
||||
page_num = (num_videos - 1) // 20 + 1
|
||||
base_url = "https://search.bilibili.com/all?keyword={}&order=totalrank&duration=0&tids_1=0&page={}"
|
||||
for page in range(1, page_num + 1):
|
||||
url = base_url.format(keyword, page)
|
||||
response = requests.get(url)
|
||||
soup = BeautifulSoup(response.text, 'html.parser')
|
||||
video_tags = soup.find_all('li', {'class': 'video-item matrix'})
|
||||
for tag in video_tags:
|
||||
video_id = tag.get('data - aid')
|
||||
if video_id:
|
||||
video_ids.append(video_id)
|
||||
if len(video_ids) >= num_videos:
|
||||
break
|
||||
if len(video_ids) >= num_videos:
|
||||
break
|
||||
return video_ids
|
Loading…
Reference in new issue