修改, 但搜索功能完全无法使用了

This commit is contained in:
Twisuki 2024-12-11 18:09:08 +08:00
parent ccb12f0f63
commit 086f21272f

View File

@ -2,24 +2,34 @@ from nonebot.log import logger
import re
import httpx
import urllib.parse
from bs4 import BeautifulSoup
from watchfiles import awatch
async def get_async_data (url):
async with httpx.AsyncClient() as client:
return await client.get(url)
async def search(msg : str, num : int):
logger.info(f"搜索 : \"{msg}\"")
result = ""
headers = {
'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/100.0.4896.127 Safari/537.36'
}
url = "https://mzh.moegirl.org.cn/index.php?search=" + msg
async with httpx.AsyncClient() as client:
response = await client.get(url, headers = headers)
logger.info(response.headers.get('Location'))
logger.info(f"连接萌娘百科中, 状态码 : {response.status_code}")
# headers = {
# 'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/100.0.4896.127 Safari/537.36'
# }
url = "https://mzh.moegirl.org.cn/index.php?search=" + urllib.parse.quote_plus(msg)
# client = httpx.AsyncClient()
# async with client:
# response = await client.post(url, headers = headers)
response = await get_async_data(url)
logger.info(f"连接\"{response.headers.get('location')}\"中, 状态码 : {response.status_code}")
# 正常搜索
if response.status_code == 200:
"""
萌娘百科搜索页面结构
div.searchresults # 若无, 证明页面已重定向
div.searchresults
p ...
ul.mw-search-results # 若无, 证明无搜索结果
li # 一个搜索结果
@ -29,11 +39,8 @@ async def search(msg : str, num : int):
li ...
li ...
"""
if response.status_code == 200:
soup = BeautifulSoup(response.text, 'html.parser')
# 检测div.searchresults, 是否已重定向
if soup.find('div', class_='searchresults'):
# 检测ul.mw-search-results, 是否有结果
if soup.find('ul', class_='mw-search-results'):
ul_tag = soup.select('ul.mw-search-results')[0]
@ -62,9 +69,13 @@ async def search(msg : str, num : int):
logger.info("无结果")
return "无结果"
# 无div.searchresults, 重定向
else:
# 重定向
elif response.status_code == 302:
logger.info(f"\"{msg}\"已被重定向")
# 读取重定向结果
soup = BeautifulSoup(response.text, 'html.parser')
logger.success("重定向成功")
logger.info(response)
num = 0
"""
@ -83,14 +94,15 @@ async def search(msg : str, num : int):
p = str(p_tag)
p = re.sub(r'<.*?>', '', p)
if p != '':
result += str(p) + "/n"
result += str(p) + "\n"
num += 1
if num >= 5:
break
return result
# 状态码非200
# 状态码非200或302
else:
logger.error(f"网络错误, 状态码 : {response.status_code}")
return f"网络错误, 状态码 : {response.status_code}"