123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190 |
- #coding=utf-8
- #!/usr/bin/python
- import sys
- sys.path.append('..')
- from base.spider import Spider
- import json
- import time
- import base64
- class Spider(Spider): # 元类 默认的元类 type
- def getName(self):
- return "哔哩"
- def init(self,extend=""):
- print("============{0}============".format(extend))
- pass
- def isVideoFormat(self,url):
- pass
- def manualVideoCheck(self):
- pass
- def homeContent(self,filter):
- result = {}
- cateManual = {
- "7年级语文": "7年级语文",
- "7年级数学": "7年级数学",
- "7年级英语": "7年级英语",
- "7年级物理": "7年级物理",
- "7年级化学": "7年级化学",
- "7年级生物": "7年级生物",
- "7年级政治": "7年级政治",
- "7年级历史": "7年级历史",
- "7年级地理": "7年级地理",
- "8年级语文": "8年级语文",
- "8年级数学": "8年级数学",
- "8年级英语": "8年级英语",
- "8年级物理": "8年级物理",
- "8年级化学": "8年级化学",
- "8年级生物": "8年级生物",
- "8年级政治": "8年级政治",
- "8年级历史": "8年级历史",
- "8年级地理": "8年级地理",
- "9年级语文": "9年级语文",
- "9年级数学": "9年级数学",
- "9年级英语": "9年级英语",
- "9年级物理": "9年级物理",
- "9年级化学": "9年级化学",
- "9年级生物": "9年级生物",
- "9年级政治": "9年级政治",
- "9年级历史": "9年级历史",
- "9年级地理": "9年级地理"
- }
- classes = []
- for k in cateManual:
- classes.append({
- 'type_name':k,
- 'type_id':cateManual[k]
- })
- result['class'] = classes
- if(filter):
- result['filters'] = self.config['filter']
- return result
- def homeVideoContent(self):
- result = {
- 'list':[]
- }
- return result
- cookies = ''
- def getCookie(self):
- rsp = self.fetch("https://www.bilibili.com/")
- self.cookies = rsp.cookies
- return rsp.cookies
- def categoryContent(self,tid,pg,filter,extend):
- result = {}
- url = 'https://api.bilibili.com/x/web-interface/search/type?search_type=video&keyword={0}&duration=4&page={1}'.format(tid,pg)
- if len(self.cookies) <= 0:
- self.getCookie()
- rsp = self.fetch(url,cookies=self.cookies)
- content = rsp.text
- jo = json.loads(content)
- if jo['code'] != 0:
- rspRetry = self.fetch(url,cookies=self.getCookie())
- content = rspRetry.text
- jo = json.loads(content)
- videos = []
- vodList = jo['data']['result']
- for vod in vodList:
- aid = str(vod['aid']).strip()
- title = vod['title'].strip().replace("<em class=\"keyword\">","").replace("</em>","")
- img = 'https:' + vod['pic'].strip()
- remark = str(vod['duration']).strip()
- videos.append({
- "vod_id":aid,
- "vod_name":title,
- "vod_pic":img,
- "vod_remarks":remark
- })
- result['list'] = videos
- result['page'] = pg
- result['pagecount'] = 9999
- result['limit'] = 90
- result['total'] = 999999
- return result
- def cleanSpace(self,str):
- return str.replace('\n','').replace('\t','').replace('\r','').replace(' ','')
- def detailContent(self,array):
- aid = array[0]
- url = "https://api.bilibili.com/x/web-interface/view?aid={0}".format(aid)
- rsp = self.fetch(url,headers=self.header)
- jRoot = json.loads(rsp.text)
- jo = jRoot['data']
- title = jo['title'].replace("<em class=\"keyword\">","").replace("</em>","")
- pic = jo['pic']
- desc = jo['desc']
- typeName = jo['tname']
- vod = {
- "vod_id":aid,
- "vod_name":title,
- "vod_pic":pic,
- "type_name":typeName,
- "vod_year":"",
- "vod_area":"",
- "vod_remarks":"",
- "vod_actor":"",
- "vod_director":"",
- "vod_content":desc
- }
- ja = jo['pages']
- playUrl = ''
- for tmpJo in ja:
- cid = tmpJo['cid']
- part = tmpJo['part']
- playUrl = playUrl + '{0}${1}_{2}#'.format(part,aid,cid)
- vod['vod_play_from'] = 'B站'
- vod['vod_play_url'] = playUrl
- result = {
- 'list':[
- vod
- ]
- }
- return result
- def searchContent(self,key,quick):
- result = {
- 'list':[]
- }
- return result
- def playerContent(self,flag,id,vipFlags):
- # https://www.555dianying.cc/vodplay/static/js/playerconfig.js
- result = {}
- ids = id.split("_")
- url = 'https://api.bilibili.com:443/x/player/playurl?avid={0}&cid=%20%20{1}&qn=112'.format(ids[0],ids[1])
- rsp = self.fetch(url)
- jRoot = json.loads(rsp.text)
- jo = jRoot['data']
- ja = jo['durl']
-
- maxSize = -1
- position = -1
- for i in range(len(ja)):
- tmpJo = ja[i]
- if maxSize < int(tmpJo['size']):
- maxSize = int(tmpJo['size'])
- position = i
- url = ''
- if len(ja) > 0:
- if position == -1:
- position = 0
- url = ja[position]['url']
- result["parse"] = 0
- result["playUrl"] = ''
- result["url"] = url
- result["header"] = {
- "Referer":"https://www.bilibili.com",
- "User-Agent":"Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/100.0.4896.127 Safari/537.36"
- }
- result["contentType"] = 'video/x-flv'
- return result
- config = {
- "player": {},
- "filter": {}
- }
- header = {}
- def localProxy(self,param):
- return [200, "video/MP2T", action, ""]
|