梨视频视频爬取

# -*- coding:utf-8 -*-
# @time：2025/7/8 10:10
# @author:张旭辉
# @file : zhang.py
# @编辑器: PyCharm

import requests
def get_data(url):
    headers = {
        "user-agent":"Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/138.0.0.0 Safari/537.36",
    }
    response =  requests.get(url=url,headers=headers)
    response.encoding = response.apparent_encoding
    return response

def get_video_url(url,contId,params):
    headers = {
        "user-agent":"Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/138.0.0.0 Safari/537.36",
        "cookie":'PEAR_UUID=d6f6a867-a448-4561-9861-0170363cfa46; _uab_collina=175188032684858668947395; p_h5_u=EDC6CA6A-43EB-4E01-B161-70C22435000A; Hm_lvt_9707bc8d5f6bba210e7218b8496f076a=1751880327,1752203335; HMACCOUNT=15F47A8122133E49; Hm_lpvt_9707bc8d5f6bba210e7218b8496f076a=1752203336; tgw_l7_route=e0ec575606f95cc9ad003a83f9f8574f; JSESSIONID=7E3F76514FB7332116890EA82275AA69',
        "referer":f"https://www.pearvideo.com/video_{contId}"
    }
    # verify = False 发送get请求的时候 不去校验证书
    response =  requests.get(url=url,headers=headers,params=params)
    response.encoding = response.apparent_encoding
    return response

# 准备当前页面url地址
url = "https://www.pearvideo.com/video_1801228"
# 从url中获取当前视频的id
contId = url.split("_")[-1]
# 向指定服务器发送请求
video_url = "https://www.pearvideo.com/videoStatus.jsp"
response = get_video_url(video_url,contId,params={"contId":contId})
# 获取视频假地址
srcUrl = response.json()['videoInfo']['videos']['srcUrl']
# 假地址： https://video.pearvideo.com/mp4/short/20250709/1752204172726-16056591-hd.mp4
print(srcUrl)
# 真实地址：https://video.pearvideo.com/mp4/short/20250709/cont-1801228-16056591-hd.mp4
# 需要把 1752204172726 替换成 cont-1801228
# 1.根据srcUrl 提取对应的这串数字
fake = srcUrl.split('/')[-1].split('-')[0]
print(fake)
# 2.替换
real_video_url = srcUrl.replace(fake,f"cont-{contId}")
print(real_video_url)
# 向真实视频地址发送请求保存数据
response = get_data(real_video_url)
# 下载到本地
print("正在下载")
with open(f"{contId}.mp4","wb") as f:
    f.write(response.content)

print("下载完成")