bilibili.py 24.9 KB
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498
# -*- coding:utf-8 -*-
# @Time : 2019/12/11 18:58
# @Author : litao
import numpy as np
import random
import json, redis, re, requests
from selenium.webdriver import ActionChains
import time, datetime, copy
from selenium import webdriver
from PIL import Image
import os
from selenium.webdriver.support.ui import WebDriverWait
import cv2
from fontTools.ttLib import *
from crawler.crawler_sys.utils.trans_str_play_count_to_int import trans_play_count
from crawler.crawler_sys.utils.trans_duration_str_to_second import trans_duration
from concurrent.futures import ProcessPoolExecutor


rds_list = redis.StrictRedis(host='192.168.17.60', port=6379, db=1, decode_responses=True)
rds_single = redis.StrictRedis(host='192.168.17.60', port=6379, db=0, decode_responses=True)
# rds_get = redis.StrictRedis(host='192.168.17.60', port=6379, db=15, decode_responses=True)

class CrawlerMain(object):
    def __init__(self):
        self.chrome_options = webdriver.ChromeOptions()
        self.chrome_options.add_argument('--headless')
        self.chrome_options.add_argument('--disable-gpu')
        # self.chrome_options.add_argument("--start-maximized")
        self.chrome_options.add_argument("--no-sandbox")
        self.chrome_options.add_experimental_option('excludeSwitches', ['enable-automation'])
        self.timestamp = str(datetime.datetime.now().timestamp() * 1e3)
        prefs = {"profile.managed_default_content_settings.images": 2}
        self.chrome_options.add_experimental_option("prefs", prefs)
        # self.driver = webdriver.Chrome(options=self.chrome_options)
        self.one_video_dic = {
                "platform": "bilibili",
                "ID": "",
                "title": "",
                "url": "",
                "describe": "",
                "video_count": "",
                "sum_duration": "",
                "year": "",
                "provider": "",
                "style_tags": "",
                "project_tags":"",
                "language": "",
                "area": "",
                "if_pay": "",
                "play_count_sum": "",
                "play_heat": "",
                "rate": "",
                "favorite_count_sum": "",
                "comment_count_sum": "",
                "barrage_count": "",
        }
        self.single_video_dic = {
                "platform": "bilibili",
                "vid": "",
                "title": "",
                "video_title": "",
                "url": "",
                "video_url":"",
                "describe": "",
                "video_count": "",
                "duration": "",
                "year": "",
                "provider": "",
                "language": "",
                "area": "",
                "if_pay": "",
                "play_count": "",
                "play_heat": "",
                "rate": "",
                "favorite_count": "",
                "comment_count": "",
                "barrage_count": "",
        }

    def __exit__(self):
        self.driver.close()


    def list_page(self, url=None,
                  video_list_xpath=None,
                  play_count_xpath=None,
                  if_free=None,
                  title=None,
                  describe=None,
                  project_tag_xpath=None,
                  provider=None,
                  year=None,
                  next_page_xpath=None,
                  roll=None,
                  project_tag=None,
                  style_tags=None
                  ):

        # js = "var q=document.documentElement.scrollTop=%s" % roll
        # self.driver.execute_script(js)
        while True:
            # self.driver.get(url)
            # self.driver.implicitly_wait(5)
            time.sleep(0.2)
            # self.driver.execute_script(js)
            try:
                next_page_obj = self.driver.find_elements_by_xpath(next_page_xpath)
            except:
                # self.driver.get(self.driver.current_url)
                continue
            vidoe_list_obj = self.driver.find_elements_by_xpath(video_list_xpath)
            try:
                for one_video in vidoe_list_obj:
                    if_pay = ""
                    str_res_list = one_video.text.split("\n")
                    if len(str_res_list) == 4:
                        play_count_str, if_pay, title_str, video_count_str = str_res_list
                    elif len(str_res_list) == 3:
                        play_count_str, title_str, video_count_str = str_res_list
                    else:
                        play_count_str, title_str = str_res_list
                        video_count_str = ""
                    play_count = trans_play_count(play_count_str)
                    project_name = "bilibili_%s" % title_str
                    url = one_video.find_element_by_xpath("./a[1]").get_attribute('href')
                    data_dic = {
                            "play_count": play_count,
                            "url": url,
                            "title": title_str,
                            "video_count": video_count_str,
                            "if_pay": if_pay
                    }
                    if style_tags:
                        temp_dic = {}
                        temp_dic["style_tags"] = style_tags
                        data_dic.update(temp_dic)
                    if project_tag:
                        temp_dic = {}
                        temp_dic["project_tags"] = project_tag
                        data_dic.update(temp_dic)
                    if year:
                        temp_dic = {}
                        temp_dic["year"] = year
                        data_dic.update(temp_dic)
                    self.parse_data(data_dic, project_name)
                if next_page_obj:
                    # next_page_obj[0].click()
                    action = ActionChains(self.driver)
                    action.click(next_page_obj[0]).perform()
                else:
                    # self.driver.close()
                    break
            except Exception as e:
                print(e)
                self.driver.close()

    def style_tags(self, url=None,
                     video_list_xpath=None,
                     play_count_xpath=None,
                     if_free=None,
                     title=None,
                     describe=None,
                     style_tags_xpath=None,
                     style=None,
                     project_tag=None,
                     provider=None,
                     year=None,
                     next_page_xpath=None,
                     roll=None, ):
        self.driver.get(url)
        time.sleep(0.2)
        style_tags_obj = self.driver.find_elements_by_xpath(style_tags_xpath)
        for style_tag in style_tags_obj:
            style_tags = style_tag.text
            print(style_tags)
            action = ActionChains(self.driver)
            action.click(style_tag).perform()
            if style_tags == "全部":
                continue
            year=style_tags
        # if True:
            self.list_page(url=self.driver.current_url,
                           video_list_xpath=video_list_xpath,
                           play_count_xpath=play_count_xpath,
                           if_free=if_free,
                           title=title,
                           describe=describe,
                           project_tag_xpath=style_tags_xpath,
                           project_tag=project_tag,
                           style_tags=style,
                           provider=provider,
                           year=year,
                           next_page_xpath=next_page_xpath,
                           roll=roll,)

    def parse_data(self, data_dic, project_name):
        res = rds_list.hgetall(project_name)
        if res:
            if not res.get("project_tags"):
                res["project_tags"] = ""
            if data_dic.get("style_tags"):
                if data_dic.get("style_tags") not in res["style_tags"]:
                    data_dic["style_tags"] = res["style_tags"] + "," + data_dic.get("style_tags")
            if data_dic.get("project_tags"):
                if data_dic.get("project_tags") not in res["project_tags"]:
                    data_dic["project_tags"] = res["project_tags"] + "," + data_dic.get("project_tags")
            if data_dic.get("provider"):
                if data_dic.get("provider") not in res["provider"]:
                    data_dic["provider"] = res["provider"] + "," + data_dic.get("provider")
            res.update(data_dic)
            rds_list.hmset(project_name, res)
        else:
            data = copy.deepcopy(self.one_video_dic)
            data.update(data_dic)
            rds_list.hmset(project_name, data)


    def parse_single_data(self, data_dic, project_name):
        res = rds_single.hgetall(project_name)
        if res:
            if not res.get("project_tags"):
                res["project_tags"] = ""
            if data_dic.get("style_tags"):
                if data_dic.get("style_tags") not in res["style_tags"]:
                    data_dic["style_tags"] = res["style_tags"] + "," + data_dic.get("style_tags")
            if data_dic.get("project_tags"):
                if data_dic.get("project_tags") not in res["project_tags"]:
                    data_dic["project_tags"] = res["project_tags"] + "," + data_dic.get("project_tags")
            if data_dic.get("provider"):
                if data_dic.get("provider") not in res["provider"]:
                    data_dic["provider"] = res["provider"] + "," + data_dic.get("provider")
            res.update(data_dic)
            rds_single.hmset(project_name, res)
        else:
            data = copy.deepcopy(self.single_video_dic)
            data.update(data_dic)
            rds_single.hmset(project_name, data)


    def video_page(self,task=0):
        self.driver = webdriver.Chrome(options=self.chrome_options)
        self.driver.maximize_window()
        has_data = rds_get.dbsize()
        while has_data:
            keys = rds_get.randomkey()
            res = rds_get.hgetall(keys)

            has_data = rds_get.dbsize()
            # print(has_data)
            # print(res)
            # self.driver.execute_script('window.open("%s");' % res["url"])
              # 关闭到当前窗口
            # self.driver.close()
            time.sleep(0.2)
            # for handle in self.driver.window_handles:
            #     self.driver.switch_to.window(handle)
            self.driver.get(res["url"])
            self.driver.execute_script("window.scrollBy(0,1000)")
            self.driver.implicitly_wait(10)
            title = self.driver.find_element_by_xpath("//a[@class='media-title']").text
            print("task ",str(task),title)
            detail_page_url_obj = self.driver.find_element_by_xpath("//a[@class='media-title']")
            detail_page_url = detail_page_url_obj.get_attribute("href")
            print(detail_page_url)
            time.sleep(1)
            discribe_info = self.driver.find_element_by_xpath("//div[@class='media-count']")
            discribe_info_str = discribe_info.text
            play_count,barrage_count,favorite_count = discribe_info_str.split("  ·  ")
            # print(discribe_info_str)
            play_count=trans_play_count(play_count)
            if not play_count:
                print(discribe_info_str)
                play_count="--"
            rate_list = self.driver.find_elements_by_xpath("//div[@class='media-rating']")
            if rate_list:
                rate = rate_list[0].text
            else:
                rate = ""
            # print(rate)
            describe = self.driver.find_element_by_xpath("//a[@class='media-desc webkit-ellipsis']").text
            describe = describe.split("\n",-1)[0]
            # print(describe)
            dic = {"describe":describe,
                   "rate":rate,
                   "play_count":play_count,
                   "barrage_count":barrage_count,
                   "favorite_count":favorite_count,
                   }
            project_name = "bilibili_%s" % title
            self.parse_data(dic,project_name)
            self.one_video_page(title,res["url"])
            rds_get.delete(keys)


    def one_video_page(self,title,url):
        video_obj_list = self.driver.find_elements_by_xpath("//div[@id='eplist_module']//li")
        if video_obj_list:
            # time.sleep(0.1)
            action = ActionChains(self.driver)
            video_name_tags = self.driver.find_elements_by_xpath("//i[@class='mode-change iconfont icon-ep-list-simple']")
            if video_name_tags:
            # time.sleep(0.1)
                action.move_to_element(video_name_tags[0]).click().perform()
                del action
            time.sleep(0.1)
            video_obj_list = self.driver.find_elements_by_xpath("//div[@id='eplist_module']//li")
            video_obj= self.driver.find_element_by_xpath("//div[@id='eplist_module']//li")
            for video_count,video_obj in enumerate(video_obj_list):
                self.driver.implicitly_wait(10)
                action = ActionChains(self.driver)
                action.click(video_obj).perform()
                del action
                self.driver.execute_script("window.scrollBy(0,1000)")
                time.sleep(0.2)
                video_title = video_obj.text
                if_pay = ""
                # print(video_title)
                if "\n" in video_title:
                    video_title,if_pay = video_title.split("\n",-1)
                # print(self.driver.page_source)
                comment_count_list = self.driver.find_elements_by_xpath("//span[@class='results']")
                if comment_count_list:
                    comment_count = comment_count_list[0].text
                else:
                    comment_count = 0
                # print(comment_count)
                video_id = self.driver.find_element_by_xpath("//a[@class='av-link']")
                video_url = video_id.get_attribute("href")
                # print(video_url)
                # print(video_id.text)
                barrage_count_list = self.driver.find_elements_by_xpath("//span[@class='bilibili-player-video-info-danmaku-number']")
                if barrage_count_list:
                    barrage_count = barrage_count_list[0].text
                else:
                    barrage_count = "-"
                # print(barrage_count)
                duration = self.driver.find_elements_by_xpath('//*[@id="bilibiliPlayer"]/div[1]/div[1]/div[10]/div[2]/div[2]/div[1]/div[3]/div/span[3]')
                try:
                    duration = trans_duration(duration[0].text)
                    print(video_count,duration)
                except:
                    duration = ""
                # print(duration)
                project_name = "bilibili_%s_%s" % (title,video_title)
                dic = {
                        "title": title,
                        "video_title": video_title,
                        "if_pay":if_pay,
                        "comment_count":comment_count,
                        "url":url,
                        "video_url":video_url,
                        "video_id":video_id.text,
                        "barrage_count":barrage_count,
                        "duration":duration,
                        "video_count":video_count + 1
                }
                self.parse_single_data(dic,project_name)
        else:
            self.driver.execute_script("window.scrollBy(0,1000)")
            # time.sleep(0.4)
            video_title = self.driver.find_element_by_xpath('//*[@id="bilibiliPlayer"]/div[1]/div[1]/div[1]/div[1]').text
            if_pay = ""
            # print(video_title)
            if "\n" in video_title:
                video_title, if_pay = video_title.split("\n", -1)
            # print(self.driver.page_source)
            comment_count = self.driver.find_element_by_xpath("//span[@class='results']").text
            # print(comment_count)
            video_id = self.driver.find_element_by_xpath("//a[@class='av-link']")
            video_url = video_id.get_attribute("href")
            # print(video_url)
            # print(video_id.text)
            barrage_count = self.driver.find_element_by_xpath(
                "//span[@class='bilibili-player-video-info-danmaku-number']").text
            # print(barrage_count)
            duration = self.driver.find_elements_by_xpath(
                '//*[@id="bilibiliPlayer"]/div[1]/div[1]/div[10]/div[2]/div[2]/div[1]/div[3]/div/span[3]')
            try:
                duration = trans_duration(duration[0].text)
                print(duration)
            except:
                duration = 0
            project_name = "bilibili_%s_%s" % (title, video_title)
            dic = {
                    "title": title,
                    "video_title": video_title,
                    "if_pay": if_pay,
                    "comment_count": comment_count,
                    "url": url,
                    "video_url": video_url,
                    "video_id": video_id.text,
                    "barrage_count": barrage_count,
                    "duration": duration,
                    "video_count":1
            }
            self.parse_single_data(dic, project_name)


    def detail_page(self):
        pass

if __name__ == "__main__":
    # crlawler_bilbili = CrawlerMain()
    # crlawler_bilbili.list_page(url="https://www.bilibili.com/documentary/index/#style_id=-1&producer_id=-1&release_date=-1&season_status=1&order=2&st=3&sort=0&page=1",
    #                            video_list_xpath="//ul[@class='bangumi-list clearfix']//li",
    #                            next_page_xpath="//a[@class='p next-page']",
    #                            roll=1200
    #                            )
    # style_tags_dic = {
    #         "历史":"https://www.bilibili.com/documentary/index/?spm_id_from=333.6.b_7375626e6176.6#style_id=10033&producer_id=-1&release_date=-1&season_status=-1&order=2&st=3&sort=0&page=1",
    #         "美食":"https://www.bilibili.com/documentary/index/?spm_id_from=333.6.b_7375626e6176.6#style_id=10045&producer_id=-1&release_date=-1&season_status=-1&order=2&st=3&sort=0&page=1",
    #         "人文":"https://www.bilibili.com/documentary/index/?spm_id_from=333.6.b_7375626e6176.6#style_id=10065&producer_id=-1&release_date=-1&season_status=-1&order=2&st=3&sort=0&page=1",
    #         "科技":"https://www.bilibili.com/documentary/index/?spm_id_from=333.6.b_7375626e6176.6#style_id=10066&producer_id=-1&release_date=-1&season_status=-1&order=2&st=3&sort=0&page=1",
    #         "探险":"https://www.bilibili.com/documentary/index/?spm_id_from=333.6.b_7375626e6176.6#style_id=10067&producer_id=-1&release_date=-1&season_status=-1&order=2&st=3&sort=0&page=1",
    #         "宇宙":"https://www.bilibili.com/documentary/index/?spm_id_from=333.6.b_7375626e6176.6#style_id=10068&producer_id=-1&release_date=-1&season_status=-1&order=2&st=3&sort=0&page=1",
    #         "萌宠":"https://www.bilibili.com/documentary/index/?spm_id_from=333.6.b_7375626e6176.6#style_id=10069&producer_id=-1&release_date=-1&season_status=-1&order=2&st=3&sort=0&page=1",
    #         "社会":"https://www.bilibili.com/documentary/index/?spm_id_from=333.6.b_7375626e6176.6#style_id=10070&producer_id=-1&release_date=-1&season_status=-1&order=2&st=3&sort=0&page=1",
    #         "动物":"https://www.bilibili.com/documentary/index/?spm_id_from=333.6.b_7375626e6176.6#style_id=10071&producer_id=-1&release_date=-1&season_status=-1&order=2&st=3&sort=0&page=1",
    #         "自然":"https://www.bilibili.com/documentary/index/?spm_id_from=333.6.b_7375626e6176.6#style_id=10072&producer_id=-1&release_date=-1&season_status=-1&order=2&st=3&sort=0&page=1",
    #         "医疗":"https://www.bilibili.com/documentary/index/?spm_id_from=333.6.b_7375626e6176.6#style_id=10073&producer_id=-1&release_date=-1&season_status=-1&order=2&st=3&sort=0&page=1",
    #         "军事":"https://www.bilibili.com/documentary/index/?spm_id_from=333.6.b_7375626e6176.6#style_id=10074&producer_id=-1&release_date=-1&season_status=-1&order=2&st=3&sort=0&page=1",
    #         "灾难":"https://www.bilibili.com/documentary/index/?spm_id_from=333.6.b_7375626e6176.6#style_id=10064&producer_id=-1&release_date=-1&season_status=-1&order=2&st=3&sort=0&page=1",
    #         "罪案":"https://www.bilibili.com/documentary/index/?spm_id_from=333.6.b_7375626e6176.6#style_id=10075&producer_id=-1&release_date=-1&season_status=-1&order=2&st=3&sort=0&page=1",
    #         "神秘":"https://www.bilibili.com/documentary/index/?spm_id_from=333.6.b_7375626e6176.6#style_id=10076&producer_id=-1&release_date=-1&season_status=-1&order=2&st=3&sort=0&page=1",
    #         "旅行":"https://www.bilibili.com/documentary/index/?spm_id_from=333.6.b_7375626e6176.6#style_id=10077&producer_id=-1&release_date=-1&season_status=-1&order=2&st=3&sort=0&page=1",
    #         "运动":"https://www.bilibili.com/documentary/index/?spm_id_from=333.6.b_7375626e6176.6#style_id=10038&producer_id=-1&release_date=-1&season_status=-1&order=2&st=3&sort=0&page=1",
    #         "电影":"https://www.bilibili.com/documentary/index/?spm_id_from=333.6.b_7375626e6176.6#style_id=-10&producer_id=-1&release_date=-1&season_status=-1&order=2&st=3&sort=0&page=1",
    # }
    # for style in style_tags_dic:
    #     try:
    #         crlawler_bilbili = CrawlerMain()
    #         crlawler_bilbili.style_tags(
    #             url=style_tags_dic[style],
    #             video_list_xpath="//ul[@class='bangumi-list clearfix']//li",
    #             next_page_xpath="//a[@class='p next-page']",
    #             roll=1200,style=style
    #             )
    #     except:
    #         continue

#     project_dic = {
#         "央视":"https://www.bilibili.com/documentary/index/#style_id=-1&producer_id=4&release_date=-1&season_status=-1&order=2&st=3&sort=0&page=1",
#         "BBC":"https://www.bilibili.com/documentary/index/#style_id=-1&producer_id=1&release_date=-1&season_status=-1&order=2&st=3&sort=0&page=1",
#         "探索频道":"https://www.bilibili.com/documentary/index/#style_id=-1&producer_id=7&release_date=-1&season_status=-1&order=2&st=3&sort=0&page=1",
#         "NHK":"https://www.bilibili.com/documentary/index/#style_id=-1&producer_id=2&release_date=-1&season_status=-1&order=2&st=3&sort=0&page=1",
#         "历史频道":"https://www.bilibili.com/documentary/index/#style_id=-1&producer_id=6&release_date=-1&season_status=-1&order=2&st=3&sort=0&page=1",
#         "卫视":"https://www.bilibili.com/documentary/index/#style_id=-1&producer_id=8&release_date=-1&season_status=-1&order=2&st=3&sort=0&page=1",
#         "自制":"https://www.bilibili.com/documentary/index/#style_id=-1&producer_id=9&release_date=-1&season_status=-1&order=2&st=3&sort=0&page=1",
#         "ITV":"https://www.bilibili.com/documentary/index/#style_id=-1&producer_id=5&release_date=-1&season_status=-1&order=2&st=3&sort=0&page=1",
#         "SKY":"https://www.bilibili.com/documentary/index/#style_id=-1&producer_id=3&release_date=-1&season_status=-1&order=2&st=3&sort=0&page=1",
#         "ZDF":"https://www.bilibili.com/documentary/index/#style_id=-1&producer_id=10&release_date=-1&season_status=-1&order=2&st=3&sort=0&page=1",
#         "合作机构":"https://www.bilibili.com/documentary/index/#style_id=-1&producer_id=11&release_date=-1&season_status=-1&order=2&st=3&sort=0&page=1",
#         "国内其他":"https://www.bilibili.com/documentary/index/#style_id=-1&producer_id=12&release_date=-1&season_status=-1&order=2&st=3&sort=0&page=1",
#         "国外其他":"https://www.bilibili.com/documentary/index/#style_id=-1&producer_id=13&release_date=-1&season_status=-1&order=2&st=3&sort=0&page=1",
#
# }
#     for project in project_dic:
#         try:
#             crlawler_bilbili = CrawlerMain()
#             crlawler_bilbili.style_tags(
#                 url=project_dic[project],
#                 video_list_xpath="//ul[@class='bangumi-list clearfix']//li",
#                 next_page_xpath="//a[@class='p next-page']",
#                 roll=1200,project_tag=project
#                 )
#         except:
#             continue
#     project_dic = {
#         "2020":"https://www.bilibili.com/documentary/index/?spm_id_from=333.6.b_7375626e6176.6#style_id=-1&producer_id=-1&release_date=-1&season_status=-1&order=2&st=3&sort=0&page=1",
# }
#     crlawler_bilbili = CrawlerMain()
#     for project in project_dic:
#         crlawler_bilbili.style_tags(
#             url=project_dic[project],
#             video_list_xpath="//ul[@class='bangumi-list clearfix']//li",
#             next_page_xpath="//a[@class='p next-page']",
#             roll=1200,year=project,
#         style_tags_xpath="/html[1]/body[1]/div[2]/div[2]/div[2]/div[2]/div[3]/ul[1]/li"
#             )




    executor = ProcessPoolExecutor(max_workers=6)
    futures = []
    crlawler_bilbili = CrawlerMain()
    # process_task = crlawler_bilbili.video_page
    for one_scan in range(3):
        crlawler_bilbili = CrawlerMain()
        future = executor.submit(crlawler_bilbili.video_page,task=one_scan)
        futures.append(future)

        # crlawler_bilbili.video_page(one_scan)
    executor.shutdown(True)
    # crlawler_bilbili = CrawlerMain()
    # crlawler_bilbili.video_page()