使用scrapy+selenium爬取淘宝网
–***2019-3-27测试有效***—-
第一步:
打开cmd,输入scrapy startproject taobao_s新建一个项目。
接着cd 进入我们的项目文件夹内输入scrapy genspider taobao www.taobao.com新建一个爬虫
文件内是这样的,tools是我建的一个工具模块,里面有一个处理数据的函数和selenium登录的函数。
class TaobaoSpider(scrapy.Spider): name = \'taobao\' # allowed_domains = [\'www.taobao.com\'] base_url = [\'https://s.taobao.com/search?q=\'] pages = 100 re_headers = { \'user-agent\': \'Mozilla/5.0 (Windows NT 6.1; WOW64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/65.0.3325.181 Safari/537.36\', \'referer\': \'https://www.taobao.com/\', \'accept-encoding\': \'gzip, deflate, b\', } i = 1 def start_requests(self): keys = self.settings.get(\'KEYS\')#获取要搜索的关键词 self.browser,list = register()#这里调用selenium登录的方法并返回browser和一个cookies self.browser.get(self.base_url[0]+keys)#使用browser登录淘宝商品搜索页面 self.browser.execute_script("window.scrollTo(0, document.body.scrollHeight)")#使用execute_script执行js操作,这里是下拉到最底下 url_i = self.browser.current_url#获取selenium界面当前的url用来错误处理 html = self.browser.page_source#获取源代码 yield scrapy.Request(url=self.base_url[0]+keys,headers=self.re_headers,cookies=list,callback=self.parse,meta={\'html\':html,\'i\':self.i,\'url\':url_i}) def parse(self, response): time.sleep(5)#等待时间,可调 html = response.meta.get(\'html\') i = response.meta.get("i") url_i = response.meta.get("url") i +=1 if i > 100:#因为翻一百页,到了之后就不在执行循环 return try: soup = BeautifulSoup(html,\'html.parser\') lists = soup.select(\'#mainsrp-itemlist > div > div > div > div\') for list in lists:#这一段是解析数据 item = TaobaoSItem() url = list.select(\'a[class="pic-link J_ClickStat J_ItemPicA"]\')[0].attrs.get(\'href\',\'\') name = list.select("a[class=\'J_ClickStat\']")[0].get_text().strip() name = data_cleaning(name) price = list.select(\'div[class="price g_price g_price-highlight"] strong\')[0].get_text() num = list.select(\'div[class="deal-cnt"]\')[0].get_text() shop_name = list.select("a[class=\'shopname J_MouseEneterLeave J_ShopInfo\']")[0].get_text().strip() shop_name = data_cleaning(shop_name) item[\'url\'] = url item[\'name\'] = name item[\'price\'] = price item[\'num\'] = num item[\'shop_name\'] = shop_name yield item button = self.browser.find_elements(By.XPATH,\'//a[@class="J_Ajax num icon-tag"]\')[-1]#这里是获取点击下一页的,因为到第二页以后会有二个一样class的,一个是上一页,一个是下一页。 button.click()#点击进入下一页 time.sleep(random.random()*2) self.browser.execute_script("window.scrollTo(0, document.body.scrollHeight)")#下拉操作 html = self.browser.page_source yield scrapy.Request(url=response.url,headers=self.re_headers,callback=self.parse,meta={\'html\':html,\'i\':i,\'url\':url_i},dont_filter=True) except Exception as e:#如果被淘宝抓到就重新登录,用保存的url在接着获取数据 time.sleep(10) print(e) self.browser.close() self.browser,list = register() self.browser.get(url=url_i) time.sleep(random.random()*2) self.browser.execute_script("window.scrollTo(0, document.body.scrollHeight)") html = self.browser.page_source yield scrapy.Request(url=response.url,headers=self.re_headers,callback=self.parse,meta={\'html\':html,\'i\':i,\'url\':url_i},dont_filter=True) def close(spider, reason):#这是结束时执行的函数,用来关掉开启的浏览器进程 spider.browser.close()
这是tools
def data_cleaning(data):#这是清洗数据的 if \' \' in data: data = re.sub(\' \', \'\', data) if "\'" in data: data = re.sub("\'", \'\', data) if r\'\n\' in data: data = re.sub(r\'\\n\', \'\', data) return data def register():#这是登录的函数,主要 while True: #因为淘宝能够识别出selenium,有时我们会登录失败,会重新登录 browser = webdriver.FirefoxOptions() browser.add_argument(\'-headless\') #无头浏览器 browser = webdriver.Firefox(firefox_options=browser) # browser = webdriver.Firefox() browser.get(\'https://login.taobao.com/member/login.jhtml\')#进入登录页面try: input = WebDriverWait(browser, 10).until( EC.presence_of_element_located((By.CLASS_NAME, \'forget-pwd.J_Quick2Static\'))) #因为登录页面有时候是扫码登录,使用需要我们点击切换到密码登录 input.click() except Exception as e: #因为页面有时是直接密码登录,使用如果直接是密码登录就不需要点击 print(e) user = browser.find_element(By.ID, \'TPL_username_1\')#找到账号输入框 password = browser.find_element(By.ID, \'TPL_password_1\')#密码输入框 user.send_keys(USER) #输入账号并等待一下 time.sleep(random.random() * 2) password.send_keys(PASSWORD)#输入密码并等待一下 time.sleep(random.random() * 1) browser.execute_script("Object.defineProperties(navigator,{webdriver:{get:() => false}})") #淘宝对selenium的识别主要是通过navigator.webdriver,使用selenium的浏览器api显示的是True,所有我们改成fALSE就可以过淘宝的检测 action = ActionChains(browser) time.sleep(random.random() * 1) butt = browser.find_element(By.ID, \'nc_1_n1z\') browser.switch_to.frame(browser.find_element(By.ID, \'_oid_ifr_\')) browser.switch_to.default_content() action.click_and_hold(butt).perform() action.reset_actions() action.move_by_offset(285, 0).perform()#输入账号密码后会有一个滑动验证 time.sleep(random.random() * 1) button = browser.find_element(By.ID, \'J_SubmitStatic\')#登录按钮 time.sleep(random.random() * 2) button.click() time.sleep(random.random() * 2) # browser.get(\'https://www.taobao.com/\') cookie = browser.get_cookies()#获取cookies,原本想selenium实现登录,其他使用scrapy来,但是淘宝的商品搜索页的js找不到加上时间不够就没写了。 list = {}#scrapy携带的cookies需要字典类型的 for cookiez in cookie: name = cookiez[\'name\'] value = cookiez[\'value\'] list[name] = value if len(list) > 10: break else: browser.close() return browser,list
然后是数据保存
class TaobaoSPipeline(object): def open_spider(self,spider): #scrapy打开时启动,这里是打开或者新建一个txt文件,文件路径是当前路径 self.f = open(\'淘宝店铺数据.txt\',\'w\') def process_item(self, item, spider):#数据保存以字典的形式,也可以改成数据库或者csv data = {} data[\'url\'] = item[\'url\'] data[\'name\'] = item[\'name\'] data[\'price\'] = item[\'price\'] data[\'num\'] = item[\'num\'] data[\'shop_name\'] = item[\'shop_name\'] self.f.write(str(data)+\'\n\') return item def close_spider(self,spider):#scrapy结束时启动,用来关掉文件。 self.f.close()
代码地址 https://github.com/18370652038/taobao.git