农产品信息智能推荐平台(3)

农产品信息智能推荐平台(3)

昨天在摄图网上爬取的图片信息并不理想，所以又找到了百度的图片搜索功能，因此采用百度图片来代替之前的图片数据。
代码：

print('导入数据库')
    connect = pymysql.connect(host='localhost', user='root', password='xxxxx', db='nongchanpin', port=3306)
    cursor = connect.cursor()
    print("连接数据库成功")
    jiage=pandas.read_csv('2019.csv',encoding='gbk')
    jiage2=pandas.read_csv('pic2.csv',encoding='gbk')
    name2=jiage2['name']
    name=jiage['名称']
    namelist=[]
    namelist2=[]
    liu=[]
    for i in range(len(name)):
        if name[i] in namelist:
            continue
        else:
            namelist.append(name[i])
    for i in range(len(name2)):
        if name2[i] in namelist2:
            continue
        else:
            namelist2.append(name2[i])
    for i in range(len(namelist)):
        if namelist[i] in namelist2:
            continue
        else:
            liu.append(namelist[i])
    print(liu)
    url='https://image.baidu.com/search/index?tn=baiduimage&ps=1&ct=201326592&lm=-1&cl=2&nc=1&ie=utf-8&word=%E5%9B%BE%E7%89%87'
    srclist=[]
    driver.get(url)#请求首页面
    driver.maximize_window()
    driver.implicitly_wait(3)
    n=1
    for i in range(len(namelist)):
        if n==len(srclist):
            liu.append(namelist[i-1])
        n = len(srclist)
        sleep(1)
        driver.find_element_by_xpath('//*[@id="kw"]').clear()
        driver.find_element_by_xpath('//*[@id="kw"]').send_keys(namelist[i])
        driver.find_element_by_xpath('//*[@id="imgMainSearch"]/div[1]/form/span[2]/input').click()
        sleep(2)
        try:
            shuju = driver.find_element_by_xpath('//*[@id="imgid"]/div/ul/li[2]/div/a/img')
            ip = shuju.get_attribute("src")
            if ip in srclist:
                continue
            else:
                cursor.execute(
                    'insert into pic2(name,src)VALUES ("{}","{}")'.format(namelist[i], ip))
                connect.commit()
                srclist.append(ip)
        except:
            continue
    for i in range(len(liu)):
        sleep(1)
        driver.find_element_by_xpath('//*[@id="kw"]').clear()
        driver.find_element_by_xpath('//*[@id="kw"]').send_keys(liu[i])
        driver.find_element_by_xpath('//*[@id="imgMainSearch"]/div[1]/form/span[2]/input').click()
        sleep(2)
        try:
            shuju = driver.find_element_by_xpath('//*[@id="imgid"]/div/ul/li[5]/div/a/img')
            ip = shuju.get_attribute("src")
            if ip in srclist:
                continue
            else:
                cursor.execute(
                    'insert into pic2(name,src)VALUES ("{}","{}")'.format(liu[i], ip))
                connect.commit()
                srclist.append(ip)
        except:
            continue
    cursor.close()
    connect.close()

看了结果，确实比之前的图片信息好了太多。