selenium获取aiqiyi信息写入mysql数据库，在django admin后台管理中显示出来

先创建django项目，创建应用index，修改settings中的数据库把原本的sqlite改成mysql的连接方式，注意数据库名，要和获取爱奇艺信息的连接的数据库名相同；然后在应用index的models.py中写入代码，执行迁移数据，在数据库可视化软件中刷新一下，就可以看见一个表名为“应用名_index/models.py中类名”，比如这里应用名为index，类名为Infos，生成的表名就叫index_Infos；跑一下selenium的程序，数据就在index_Infos表中了。接下来创建admin管理员，在执行下runserver，就可以打开网页看到后台数据了。

 django-admin startproject MyDjango

cd MyDjango

python manage.py startapp index

# settings.py
INSTALLED_APPS = [
    'django.contrib.admin',
    'django.contrib.auth',
    'django.contrib.contenttypes',
    'django.contrib.sessions',
    'django.contrib.messages',
    'django.contrib.staticfiles',
    'index'
]

数据库

# settings
DATABASES = {
    # 第一个数据库,default是默认的数据库不可删除，可留空{}
    'default': {
        'ENGINE': 'django.db.backends.mysql',
        'NAME': 'aiqiyi_db',
        'USER': 'root',
        'PASSWORD': '1234',
        'HOST': '127.0.0.1',
        'PORT': '3306',
    }
}




LANGUAGE_CODE = 'zh-hans'

TIME_ZONE = 'Asia/Shanghai'

新建一个数据库aiqiyi_db，然后数据迁移过去

# index/models.py
from django.db import models

class Infos(models.Model):
    id = models.AutoField(primary_key=True)
    name = models.CharField(max_length=255)
    performer = models.CharField(max_length=255)
    comments = models.CharField(max_length=1000)
    source = models.CharField(max_length=255)
    # 数字:models.IntegerField()

    def __str__(self):
        # 返回名
        return self.name

python manage.py makemigrations

python manage.py migrate

创建管理员python manage.py createsuperuser

运行python manage.py runserver 8002

打开网页http://127.0.0.1:8002/admin/ 

上面创建完Django表之后的任一步，都可以运行下面获取爱奇艺的信息。运行完刷新一下admin网页。

 models.py

from sqlalchemy import *
from sqlalchemy.orm import sessionmaker
from sqlalchemy.ext.declarative import declarative_base
from datetime import datetime
database = 'mysql+pymysql://root:1234@localhost/aiqiyi_db?charset=utf8mb4'

Base = declarative_base()
# 创建数据库连接对象
engine = create_engine(database)
DBSession = sessionmaker(bind=engine)
SQLsession = DBSession()

# ORM
class Infos(Base):
    __tablename__ = 'index_infos'    #这个表是Django生成的表index_infos
    id = Column(Integer(), primary_key=True)
    name = Column(String(255))
    performer = Column(String(255))
    comments = Column(String(1000))
    source = Column(String(255))
    created = Column(DateTime, default=datetime.now())
    updated = Column(DateTime, default=datetime.now(), onupdate=datetime.now())

Base.metadata.create_all(engine)

aiqiyi.py 获取的信息分电视剧和电影两种

import time
from selenium import webdriver
from models import *

driver = webdriver.Chrome()

driver.get('https://www.iqiyi.com/ranks1/1/0?vfrm=rank_list&vfrmblk=channel.1&vfrmrst=rank.1')
# driver.get('https://www.iqiyi.com/ranks1/1/0?vfrm=rank_list&vfrmblk=channel.1&vfrmrst=rank.1')
time.sleep(2)
movie_list=driver.find_elements_by_class_name('rvi__con')
datalist=[]
for i in movie_list:
    name = i.find_element_by_class_name('rvi__tit1').text
    performer = i.find_element_by_css_selector("span[class='rvi__type__txt']").text
    comments = i.find_element_by_css_selector("p[class='rvi__des2']").text
    print(name,performer,comments)
    time.sleep(1)
    #写入字典
    datalist.append(dict(name=name, performer=performer, comments=comments, source="TV_play"))
# datalist.append(dict(name=name, performer=performer, comments=comments, source="film"))
    #遍历字典，数据库不存在该name就写入数据库
    for d in datalist:
        find_data = SQLsession.query(Infos).filter_by(name=d['name']).first()
        if not find_data:
            SQLsession.add(Infos(**d))
    SQLsession.commit()
SQLsession.close()
time.sleep(2)
driver.quit()

显示中文

from django.db import models

# ORM
class Infos(models.Model):
    # class Meta:
    #     db_table='spiders_infos'
    id = models.AutoField(primary_key=True)
    name = models.CharField(max_length=255,verbose_name=u'影视剧名')
    performer = models.CharField(max_length=255,verbose_name=u'演员')
    comments = models.CharField(max_length=1000,verbose_name=u'简介')
    source = models.CharField(max_length=255,verbose_name=u'分类')
    # 数字:models.IntegerField()

    class Meta:
        db_table=u'index_infos'
##单复数
        # verbose_name=u'影视'    
        verbose_name_plural=u'影视' 

    def __str__(self):
        # 返回名
        return self.name