python爬蟲(chóng) 爬取58同城上所有城市的租房信息詳解

from fake_useragent import UserAgent
from lxml import etree
import requests, os
import time, re, datetime
import base64, json, pymysql
from fontTools.ttLib import TTFont

ua = UserAgent()


class CustomException(Exception):

  def __init__(self, status, msg):
    self.status = status
    self.msg = msg


class City_58:
  '''
  58同城的爬蟲(chóng)類，目前就寫這兩個(gè)
  出租房url: https://cd.58.com/chuzu/     cd代表成都縮寫
  二手房url: https://cd.58.com/ershoufang/
  '''

  font_dict = {
    "glyph00001": "0",
    "glyph00002": "1",
    "glyph00003": "2",
    "glyph00004": "3",
    "glyph00005": "4",
    "glyph00006": "5",
    "glyph00007": "6",
    "glyph00008": "7",
    "glyph00009": "8",
    "glyph00010": "9",
  }
  conn = None

  def __init__(self):
    self.session = requests.Session()
    self.session.headers = {
      "user-agent": ua.random
    }
    self.__init__all_city()

  def __init__all_city(self):
    '''獲取所有城市的名字及縮寫的對(duì)應(yīng)關(guān)系'''
    api = "https://www.58.com/changecity.html"
    headers = self.session.headers.copy()
    response = self.session.get(api, headers=headers)
    html = response.text
    res = re.findall("cityList = (.*?)</script>", html, re.S)[0]
    res = re.sub("\s", "", res)
    dic = json.loads(res)
    for k, v in dic.items():
      for k1, v1 in v.items():
        dic[k][k1] = v1.split("|")[0]
    city_dict = {}

    def traverse_dict(dic: dict):
      for k, v in dic.items():
        if k == "海外" or k == "其他":
          continue
        if isinstance(v, dict):
          traverse_dict(v)
        city_dict[k] = v

    traverse_dict(dic)

    other_city = re.findall("independentCityList = (.*?)var", html, re.S)[0]
    res = re.sub("\s", "", other_city)
    other_city_dic = json.loads(res)

    for k, v in other_city_dic.items():
      other_city_dic[k] = v.split("|")[0]

    city_dict.update(other_city_dic)
    self.all_city_dict = city_dict

  def spider_zufang(self, city: str = "成都", is_get_all: bool = True):
    '''爬取租房信息的爬蟲(chóng)方法'''
    assert self.all_city_dict is not None, "獲取所有城市信息失敗 !"
    format_city = self.all_city_dict.pop(city, None)
    assert format_city is not None, "{}該城市不在爬取城市之內(nèi)".format(city)
    while True:
      self.city = city
      # self.file = open("./house_info.json", "a", encoding="utf-8")
      start_url = self.__init_zufang(format_city)

      # 思路是什么，首先進(jìn)入?yún)^(qū)域的租房頁(yè)面，在該頁(yè)面中先提取出相應(yīng)的title，比如經(jīng)紀(jì)人，個(gè)人房源等等...
      # 我們需要構(gòu)建出相應(yīng)的url就可以了
      # start_url的格式為 https://cd.58.com/chuzu/ 我們需要轉(zhuǎn)為這樣的格式 https://cd.58.com/jintang/hezu/
      # 我們?cè)L問(wèn)轉(zhuǎn)化后的地址，再拿去到相應(yīng)的鏈接，比如經(jīng)紀(jì)人，個(gè)人房源等鏈接
      # 拿到該鏈接以后，這就是這個(gè)分類里的第一頁(yè)url，我們?cè)賹?duì)這個(gè)鏈接發(fā)生請(qǐng)求，
      # 拿到響應(yīng)體，這里可以寫一個(gè)while循環(huán)，因?yàn)槲覀儾恢烙卸嗌夙?yè)，其實(shí)也可以知道有多少頁(yè)，就是在這個(gè)響應(yīng)體中可拿到
      # 我的思路就是寫一個(gè)while循環(huán)，判斷是否有下一頁(yè)，有的繼續(xù)，沒(méi)有的話直接break

      for url_info_list in self.__get_url(start_url):
        # 這里的話，最好進(jìn)行判斷一下，因?yàn)槊總€(gè)title(值個(gè)人房源，品牌公寓等..)不一樣的話,可能爬取的策略也不太一樣
        title = url_info_list[1]
        if title in ["個(gè)人房源", "安選房源", "經(jīng)紀(jì)人", "熱租房源"] or "出租" in title:
          self.__spiders_v1(url_info_list)
          # pass
        elif title == "品牌公寓":
          self.__spiders_v2(url_info_list)
          pass
        elif title == "房屋求租":
          # 房屋求租不太想寫，數(shù)據(jù)也不是很多
          pass
        else:
          # 這種情況不在范圍內(nèi)，直接pass掉
          continue
      if not is_get_all:
        return
      try:
        city = list(self.all_city_dict.keys()).pop()
        format_city = self.all_city_dict.pop(city)
      except IndexError:
        print('全國(guó)出租房信息，爬取完畢')
        return

  def spider_ershoufang(self, city: str = "cd"):
    '''爬取二手房信息的爬蟲(chóng)方法'''
    pass

  def __spiders_v1(self, url_info_list):
    "負(fù)責(zé)處理個(gè)人房源，安選房源等等頁(yè)面的方法"
    url = url_info_list[2]
    page_num = 1
    while True:
      time.sleep(2)
      print("正在爬取{}-{}--第{}頁(yè)數(shù)據(jù)".format(url_info_list[0], url_info_list[1], page_num))
      response = self.__get_html_source(url)
      # 從html源碼中獲取到想要的數(shù)據(jù)
      for house_info_list in self.__deal_with_html_source_v1(response):
        self.__save_to_mysql(house_info_list, url_info_list)
      # 判斷是否還有下一頁(yè)
      next_page_url = self.__is_exist_next_page(response)
      if not next_page_url:
        print("{}-{}爬取完畢".format(url_info_list[0], url_info_list[1]))
        return
      url = next_page_url
      page_num += 1

  def __spiders_v2(self, url_info_list):
    '''處理品牌公寓的爬蟲(chóng)信息'''
    base_url = url_info_list[2]
    format_url = self.__format_url_v2(base_url)
    page_num = 1
    params = None
    while True:
      print("正在爬取{}--第{}頁(yè)數(shù)據(jù)...".format(url_info_list[1], page_num))
      time.sleep(2)
      url = format_url.format(page_num)
      response = self.__get_html_source(url, params)
      # 獲取到有用的數(shù)據(jù) deal_with_html_source_v2
      for house_info_list in self.__deal_with_html_source_v2(response):
        # self.__save_to_file_v2(house_info_list)
        self.__save_to_mysql(house_info_list)

      # 獲取到下一頁(yè)的encryptData
      encryptData = self.__get_html_encryptData(response)

      # 判斷是否還有下一頁(yè)，通過(guò)<div class="tip">信息不足，為您推薦附近房源</div>
      if not self.__is_exist_next_page_v2(response):
        print("{}爬取完畢".format(url_info_list[1]))
        return
      page_num += 1
      params = {
        "encryptData": encryptData or "",
        "segment": "true"
      }

  def __save_to_file_v2(self, house_info_list):
    '''
    :param house_info_list: 關(guān)于房子的信息的列表
    :param url_info_list: [區(qū)域，類型(個(gè)人房源，經(jīng)紀(jì)人等等...),url]
    :return:
    '''

    print("房間圖片地址>>:", file=self.file)
    print(json.dumps(house_info_list[0], ensure_ascii=False), file=self.file)
    print("房間描述>>:", file=self.file)
    print(json.dumps(house_info_list[1], ensure_ascii=False), file=self.file)
    print("房間詳情>>:", file=self.file)
    print(json.dumps(house_info_list[2], ensure_ascii=False), file=self.file)
    print("房間地理位置>>:", file=self.file)
    print(json.dumps(house_info_list[3], ensure_ascii=False), file=self.file)
    print("獲取房間的標(biāo)簽>>:", file=self.file)
    print(json.dumps(house_info_list[4], ensure_ascii=False), file=self.file)
    print("獲取房間的價(jià)格>>:", file=self.file)
    print(json.dumps(house_info_list[5], ensure_ascii=False), file=self.file)
    print(file=self.file)

  def __save_to_mysql(self, house_info_list, url_info_list=None):
    '''保存到數(shù)據(jù)庫(kù)'''
    if not self.conn:
      self.conn = pymysql.connect(host="127.0.0.1",
                    port=3306,
                    user="root",
                    password="root",
                    db="city_58")
      self.conn.cursor = self.conn.cursor(cursor=pymysql.cursors.DictCursor)
    if not url_info_list:
      sql = "insert into zu_house_copy (house_img_url,house_title,house_details,house_address,house_tags,hoouse_price,house_type,city) values (%s,%s,%s,%s,%s,%s,%s,%s)"
      house_info_list.append("品牌公寓")
    else:
      sql = "insert into zu_house_copy (house_img_url,house_title,house_details,house_address,house_tags,hoouse_price,area,house_type,city) values (%s,%s,%s,%s,%s,%s,%s,%s,%s)"
      house_info_list.append(url_info_list[0])
      house_info_list.append(url_info_list[1])
    house_info_list.append(self.city)
    row = self.conn.cursor.execute(sql, house_info_list)
    if not row:
      print("插入失敗")
    else:
      self.conn.commit()

  def __deal_with_html_source_v1(self, response):
    html = response.text
    self.__get_font_file(html)
    html = self.__format_html_source(html)
    for house_info_list in self.__parse_html_v1(html):
      yield house_info_list

  def __deal_with_html_source_v2(self, response):

    html = response.text
    # 源碼里的關(guān)于數(shù)字0123456789都是進(jìn)行處理過(guò)的，我們需要先獲取到字體文件
    # 我們先獲取到字體文件并且保存
    self.__get_font_file(html)

    # 對(duì)源碼中的字體進(jìn)行處理，得到瀏覽器顯示的數(shù)據(jù)
    html = self.__format_html_source(html)

    # 開(kāi)始從頁(yè)面中提取出想要的數(shù)據(jù)
    for house_info_list in self.__parse_html_v2(html):
      yield house_info_list

  def __parse_html_v1(self, html):
    xml = etree.HTML(html)

    li_xpath_list = xml.xpath("http://ul[@class='listUl']/li[@logr]")

    for li_xpath in li_xpath_list:
      house_info_list = []
      try:
        house_img_url = li_xpath.xpath("div[@class='img_list']/a/img/@lazy_src")[0]
      except IndexError:
        house_img_url = li_xpath.xpath("div[@class='img_list']/a/img/@src")[0]
      house_info_list.append(house_img_url)
      # 房間描述
      house_title = re.sub("\s", "", li_xpath.xpath("div[@class='des']/h3/a/text()")[0])
      house_info_list.append(house_title)
      # 房間詳情
      house_details = re.sub("\s", "",
                  li_xpath.xpath("div[@class='des']/p[@class='room strongbox']/text()")[0].strip())
      house_info_list.append(house_details)
      # 房間地理位置
      house_address = re.sub("\s", "",
                  li_xpath.xpath("div[@class='des']/p[@class='add']")[0].xpath("string(.)"))
      house_info_list.append(house_address)
      # 獲取房間的標(biāo)簽
      house_tags = "暫無(wú)標(biāo)簽"
      house_info_list.append(house_tags)
      # 獲取房間的價(jià)格
      hoouse_price = re.sub("\s", "",
                 li_xpath.xpath("div[@class='listliright']/div[@class='money']")[0].xpath("string(.)"))
      house_info_list.append(hoouse_price)

      yield house_info_list

  def __parse_html_v2(self, html):
    '''解析頁(yè)面，拿到數(shù)據(jù)'''
    xml = etree.HTML(html)
    li_xpath_list = xml.xpath("http://ul[@class='list']/li")
    for li_xpath in li_xpath_list:
      house_info_list = []
      # 房間圖片地址，這里只獲取了一張，我在想要不要獲取多張
      # 先空著。。。。。。。。。。。。。
      house_img_url = li_xpath.xpath("a/div[@class='img']/img/@lazy_src")[0]
      house_info_list.append(house_img_url)
      # 房間描述
      house_title = li_xpath.xpath("a/div[@class='des strongbox']/h3/text()")[0].strip()
      house_info_list.append(house_title)
      # 房間詳情
      house_details = re.sub("\s", "", li_xpath.xpath("a/div[@class='des strongbox']/p[@class='room']/text()")[0])
      # house_details = li_xpath.xpath("a/div[@class='des strongbox']/p[@class='room']/text()")[0]
      house_info_list.append(house_details)
      # 房間地理位置
      house_address = re.sub("\s", "", li_xpath.xpath(
        "a/div[@class='des strongbox']/p[@class='dist']")[0].xpath("string(.)")) or "暫無(wú)地址"
      # house_address = li_xpath.xpath( "a/div[@class='des strongbox']/p[@class='dist']/text()")[0]
      house_info_list.append(house_address)
      # 獲取房間的標(biāo)簽
      house_tags = ",".join(li_xpath.xpath("a/div[@class='des strongbox']/p[@class='spec']/span/text()"))
      house_info_list.append(house_tags)
      # 獲取房間的價(jià)格
      hoouse_price = re.sub("\s", "", li_xpath.xpath("a/div[@class='money']/span[@class='strongbox']")[0].xpath(
        "string(.)")) or "暫無(wú)價(jià)格"
      house_info_list.append(hoouse_price)

      yield house_info_list

  def __get_font_file(self, html):
    '''從源碼中獲取到字體文件，并且轉(zhuǎn)為保存，轉(zhuǎn)為TTFont對(duì)象'''
    try:
      b64 = re.findall(r"base64,(.*?)\'", html, re.S)[0]
      res = base64.b64decode(b64)
      with open("./online_font.ttf", "wb") as f:
        f.write(res)
      self.online_font = TTFont("./online_font.ttf")
      self.online_font.saveXML("./online.xml")
    except IndexError:
      return

  def __format_html_source(self, html):
    assert self.online_font, "必須創(chuàng)建字體對(duì)象"
    assert os.path.exists("./online.xml"), "請(qǐng)先獲取到字體文件。"

    with open("./online.xml", "rb") as f:
      file_data = f.read()

    online_uni_list = self.online_font.getGlyphOrder()[1:]
    file_selector = etree.HTML(file_data)
    for uni2 in online_uni_list:
      code = file_selector.xpath("http://cmap//map[@name='{}']/@code".format(uni2))[0]
      dd = "&#x" + code[2:].lower() + ";"
      if dd in html:
        html = html.replace(dd, self.font_dict[uni2])
    return html

  def __format_url_v2(self, url):
    '''
    :param url: https://cd.58.com/pinpaigongyu/?from=58_pc_zf_list_ppgy_tab_ppgy
    :return: https://cd.58.com/pinpaigongyu/pn/{}/?from=58_pc_zf_list_ppgy_tab_ppgy
    '''
    a = url.split("?")
    a[0] = a[0] + "pn/{}"
    format_url = "?".join(a)
    return format_url

  def __is_exist_next_page_v2(self, response):
    xml = self.__response_to_xml(response)
    try:
      _ = xml.xpath("http://div[@class='tip']")[0]
      return False
    except IndexError:
      return True

  def __get_html_encryptData(self, response):
    html = response.text
    encryptData = re.findall(r"encryptData\":\"(.*?)\"", html, re.S)[0]
    return encryptData

  def __get_url(self, start_url: str):
    url_set = set()
    for area, v in self.area_dict.items():
      url = self.__conversion_url(start_url, v)
      response = self.__get_html_source(url)
      title_dict = self.__get_title_info(response)
      for title_name, v in title_dict.items():
        # 對(duì)于求租、品牌公寓這個(gè)url，它是重復(fù)的,在這里進(jìn)行判斷判斷就好了
        if v in url_set:
          continue
        else:
          url_set.add(v)
          yield [area, title_name, v]

  def __conversion_url(self, url: str, area: str):
    '''
    :param url: https://cd.58.com/chuzu/
    :param area:
    :return: https://cd.58.com/區(qū)域縮寫/chuzu/
    '''
    lis = url.split("/")
    lis.insert(3, area)
    return "/".join(lis)

  def __init_zufang(self, format_city):
    '''首先將所需要的數(shù)據(jù)的獲取到'''
    start_url = "https://{}.58.com/chuzu/".format(format_city)
    headers = self.session.headers.copy()
    response = self.session.get(url=start_url, headers=headers)
    self.__get_area_info(response)
    return start_url

  def __get_html_source(self, url, params=None):
    '''通過(guò)get方式獲取到網(wǎng)頁(yè)的源碼'''
    time.sleep(1)
    headers = self.session.headers.copy()
    try:
      if not params:
        params = {}
      response = self.session.get(url=url, headers=headers, params=params)
      return response
    except Exception as e:
      with open("./url_log_error.txt", "a", encoding="utf-8") as f:
        f.write(str(datetime.datetime.now()) + "\n")
        f.write(str(e) + "\n")
        f.write("error_url>>:{}".format(url) + "\n")

  def __response_to_xml(self, response):
    try:
      xml = etree.HTML(response.text)
      return xml
    except AttributeError:
      raise CustomException(10000, "response對(duì)象轉(zhuǎn)換為xml失敗,錯(cuò)誤的鏈接地址為>>:{}".format(response))

  def __is_exist_next_page(self, response):
    '''判斷是否存在下一頁(yè),存在拿到下一頁(yè)的鏈接，不存在返回False'''
    xml = self.__response_to_xml(response)
    try:
      next_page_url = xml.xpath("http://a[@class='next']/@href")[0]
      return next_page_url
    except IndexError:
      return False

  def __get_area_info(self, response):
    '''獲取到當(dāng)前城市的區(qū)域'''
    xml = self.__response_to_xml(response)
    a_xpath_list = xml.xpath("http://dl[@class='secitem secitem_fist']//a[not(@class)]")
    area_key_list = []
    area_value_list = []
    for a_xpath in a_xpath_list:
      area_key_list.append(a_xpath.xpath("text()")[0])
      area_value_list.append(re.findall("com/(.*?)/", a_xpath.xpath("@href")[0])[0])
    assert len(area_key_list) == len(area_value_list), "數(shù)據(jù)不完整"

    self.area_dict = {k: v for k, v in zip(area_key_list, area_value_list)}

  def __get_title_info(self, response):
    '''獲取房屋的分類，比如個(gè)人房源，合租房，經(jīng)紀(jì)人，熱選房源...'''
    "listTitle"
    xml = self.__response_to_xml(response)
    a_xpath_list = xml.xpath("http://div[@class='listTitle']//a[not(@class)]")
    title_key_list = []
    title_value_list = []
    for a_xpath in a_xpath_list:
      title_key_list.append(a_xpath.xpath("span/text()")[0])
      title_value_list.append(a_xpath.xpath("@href")[0])
    assert len(title_key_list) == len(title_value_list), "數(shù)據(jù)不完整"
    return {k: v for k, v in zip(title_key_list, title_value_list)}
if __name__ == '__main__':
  city_58 = City_58()
  city_58.spider_zufang("重慶")