drf中的jwt使用与手动签发token、校验用户

jwt认证

1）session存储token，需要数据库参与，耗服务器资源、低效
2）缓存存token，需要缓存参与，高效，不易集群
3）客户端存token，服务器存签发与交易token的算法，高效，易集群

drf-jwt使用：

安装：djangorestframework-jwt
视图类签发token - username,password => token
认证类校验token - token => user

jwt格式：基本信息(头base64).用户信息过期时间(载荷base64).所有信息+秘钥(签名HS256)

签发token

源码入口

# 前提：给一个局部禁用了所有 认证与权限 的视图类发送用户信息得到token，其实就是登录接口

# 1）rest_framework_jwt.views.ObtainJSONWebToken 的 父类 JSONWebTokenAPIView 的 post 方法
#        接受有username、password的post请求
# 2）post方法将请求数据交给 rest_framework_jwt.serializer.JSONWebTokenSerializer 处理
#        完成数据的校验，会走序列化类的 全局钩子校验规则，校验得到登录用户并签发token存储在序列化对象中

核心源码：rest_framework_jwt.serializer.JSONWebTokenSerializer的validate(self, attrs)方法

def validate(self, attrs):
    # 账号密码字典
    credentials = {
        self.username_field: attrs.get(self.username_field),
        'password': attrs.get('password')
    }
    if all(credentials.values()):
        # 签发token第1步：用账号密码得到user对象
        user = authenticate(**credentials)
        if user:
            if not user.is_active:
                msg = _('User account is disabled.')
                raise serializers.ValidationError(msg)
            # 签发token第2步：通过user得到payload，payload包含着用户信息与过期时间
            payload = jwt_payload_handler(user)
            # 在视图类中，可以通过 序列化对象.object.get('user'或者'token') 拿到user和token 
            return {
                # 签发token第3步：通过payload签发出token
                'token': jwt_encode_handler(payload),
                'user': user
            }
        else:
            msg = _('Unable to log in with provided credentials.')
            raise serializers.ValidationError(msg)
    else:
        msg = _('Must include "{username_field}" and "password".')
        msg = msg.format(username_field=self.username_field)
        raise serializers.ValidationError(msg)

根据源码内容我们可以简要的推导出手动签发token的逻辑

# 1）通过username、password得到user对象
# 2）通过user对象生成payload：jwt_payload_handler(user) => payload
#        from rest_framework_jwt.serializers import jwt_payload_handler
# 3）通过payload签发token：jwt_encode_handler(payload) => token
#        from rest_framework_jwt.serializers import jwt_encode_handler

校验token

源码入口

# 前提：访问一个配置了jwt认证规则的视图类，就需要提交认证字符串token，在认证类中完成token的校验

# 1）rest_framework_jwt.authentication.JSONWebTokenAuthentication 的 父类 BaseJSONWebTokenAuthentication 的 authenticate 方法
#        请求头拿认证信息jwt-token => 通过反爬小规则确定有用的token => payload => user

核心源码：rest_framework_jwt.authentication.BaseJSONWebTokenAuthentication的authenticate(self, request)方法

def authenticate(self, request):
    """
    Returns a two-tuple of `User` and token if a valid signature has been
    supplied using JWT-based authentication.  Otherwise returns `None`.
    """
    # 带有反爬小规则的获取token：前台必须按 "jwt token字符串" 方式提交
    # 校验user第1步：从请求头 HTTP_AUTHORIZATION 中拿token，并提取
    jwt_value = self.get_jwt_value(request)
    # 游客
    if jwt_value is None:
        return None
    # 校验
    try:
        # 校验user第2步：token => payload
        payload = jwt_decode_handler(jwt_value)
    except jwt.ExpiredSignature:
        msg = _('Signature has expired.')
        raise exceptions.AuthenticationFailed(msg)
    except jwt.DecodeError:
        msg = _('Error decoding signature.')
        raise exceptions.AuthenticationFailed(msg)
    except jwt.InvalidTokenError:
        raise exceptions.AuthenticationFailed()
    # 校验user第3步：token => payload
    user = self.authenticate_credentials(payload)

    return (user, jwt_value)

手动校验token逻辑

# 1）从请求头中获取token
# 2）根据token解析出payload：jwt_decode_handler(token) => payloay
#        from rest_framework_jwt.authentication import jwt_decode_handler
# 3）根据payload解析出user：self.authenticate_credentials(payload) => user
#        继承drf-jwt的BaseJSONWebTokenAuthentication，拿到父级的authenticate_credentials方法

案例：实现多方式登陆签发token

models.py

from django.db import models

from django.contrib.auth.models import AbstractUser
class User(AbstractUser):
    mobile = models.CharField(max_length=11, unique=True)

    class Meta:
        db_table = 'api_user'
        verbose_name = '用户表'
        verbose_name_plural = verbose_name

    def __str__(self):
        return self.username

serializers.py

from rest_framework import serializers
from . import models
import re

# 拿到前台token的两个函数： user => payload => token
# from rest_framework_jwt.settings import api_settings
# jwt_payload_handler = api_settings.JWT_PAYLOAD_HANDLER
# jwt_encode_handler = api_settings.JWT_ENCODE_HANDLER
from rest_framework_jwt.serializers import jwt_payload_handler
from rest_framework_jwt.serializers import jwt_encode_handler


# 1) 前台提交多种登录信息都采用一个key，所以后台可以自定义反序列化字段进行对应
# 2) 序列化类要处理序列化与反序列化，要在fields中设置model绑定的Model类所有使用到的字段
# 3) 区分序列化字段与反序列化字段 read_only | write_only
# 4) 在自定义校验规则中（局部钩子、全局钩子）校验数据是否合法、确定登录的用户、根据用户签发token
# 5) 将登录的用户与签发的token保存在序列化类对象中
class UserModelSerializer(serializers.ModelSerializer):
    # 自定义反序列字段：一定要设置write_only，只参与反序列化，不会与model类字段映射
    usr = serializers.CharField(write_only=True)
    pwd = serializers.CharField(write_only=True)
    class Meta:
        model = models.User
        fields = ['usr', 'pwd', 'username', 'mobile', 'email']
        # 系统校验规则
        extra_kwargs = {
            'username': {
                'read_only': True
            },
            'mobile': {
                'read_only': True
            },
            'email': {
                'read_only': True
            },
        }

    def validate(self, attrs):
        usr = attrs.get('usr')
        pwd = attrs.get('pwd')

        # 多方式登录：各分支处理得到该方式下对应的用户
        if re.match(r'.+@.+', usr):
            user_query = models.User.objects.filter(email=usr)
        elif re.match(r'1[3-9][0-9]{9}', usr):
            user_query = models.User.objects.filter(mobile=usr)
        else:
            user_query = models.User.objects.filter(username=usr)
        user_obj = user_query.first()

        # 签发：得到登录用户，签发token并存储在实例化对象中
        if user_obj and user_obj.check_password(pwd):
            # 签发token，将token存放到 实例化类对象的token 名字中
            payload = jwt_payload_handler(user_obj)
            token = jwt_encode_handler(payload)
            # 将当前用户与签发的token都保存在序列化对象中
            self.user = user_obj
            self.token = token
            return attrs

        raise serializers.ValidationError({'data': '数据有误'})

views.py

#实现多方式登陆签发token：账号、手机号、邮箱等登陆
# 1) 禁用认证与权限组件
# 2) 拿到前台登录信息，交给序列化类
# 3) 序列化类校验得到登录用户与token存放在序列化对象中
# 4) 取出登录用户与token返回给前台
import re
from . import serializers, models
from utils.response import APIResponse

from rest_framework_jwt.serializers import jwt_payload_handler
from rest_framework_jwt.serializers import jwt_encode_handler

class LoginAPIView(APIView):
    # 1) 禁用认证与权限组件
    authentication_classes = []
    permission_classes = []
    def post(self, request, *args, **kwargs):
        # 2) 拿到前台登录信息，交给序列化类，规则：账号用usr传，密码用pwd传
        user_ser = serializers.UserModelSerializer(data=request.data)
        # 3) 序列化类校验得到登录用户与token存放在序列化对象中
        user_ser.is_valid(raise_exception=True)
        # 4) 取出登录用户与token返回给前台
        return APIResponse(token=user_ser.token, results=serializers.UserModelSerializer(user_ser.user).data)

    # "一根筋" 思考方式：所有逻辑都在视图类中处理
    def my_post(self, request, *args, **kwargs):
        usr = request.data.get('usr')
        pwd = request.data.get('pwd')
        if re.match(r'.+@.+', usr):
            user_query = models.User.objects.filter(email=usr)
        elif re.match(r'1[3-9][0-9]{9}', usr):
            user_query = models.User.objects.filter(mobile=usr)
        else:
            user_query = models.User.objects.filter(username=usr)
        user_obj = user_query.first()
        if user_obj and user_obj.check_password(pwd):
            payload = jwt_payload_handler(user_obj)
            token = jwt_encode_handler(payload)
            return APIResponse(results={'username': user_obj.username}, token=token)
        return APIResponse(data_msg='不可控错误')

案例：自定义认证反爬规则的认证类

authentications.py

import jwt
from rest_framework_jwt.authentication import BaseJSONWebTokenAuthentication
from rest_framework_jwt.authentication import jwt_decode_handler
from rest_framework.exceptions import AuthenticationFailed
class JWTAuthentication(BaseJSONWebTokenAuthentication):
    def authenticate(self, request):
        jwt_token = request.META.get('HTTP_AUTHORIZATION')

        # 自定义校验规则：auth token jwt
        token = self.parse_jwt_token(jwt_token)

        if token is None:
            return None

        try:
            # token => payload
            payload = jwt_decode_handler(token)
        except jwt.ExpiredSignature:
            raise AuthenticationFailed('token已过期')
        except:
            raise AuthenticationFailed('非法用户')
        # payload => user
        user = self.authenticate_credentials(payload)

        return (user, token)

    # 自定义校验规则：auth token jwt，auth为前盐，jwt为后盐
    def parse_jwt_token(self, jwt_token):
        tokens = jwt_token.split()
        if len(tokens) != 3 or tokens[0].lower() != 'auth' or tokens[2].lower() != 'jwt':
            return None
        return tokens[1]

views.py

from rest_framework.views import APIView
from utils.response import APIResponse
# 必须登录后才能访问 - 通过了认证权限组件
from rest_framework.permissions import IsAuthenticated
# 自定义jwt校验规则
from .authentications import JWTAuthentication
class UserDetail(APIView):
    authentication_classes = [JWTAuthentication]
    permission_classes = [IsAuthenticated]
    def get(self, request, *args, **kwargs):
        return APIResponse(results={'username': request.user.username})

admin使用自定义User表：新增用户密码密文

from django.contrib import admin
from . import models

# 自定义User表，admin后台管理，采用密文密码
from django.contrib.auth.admin import UserAdmin

class MyUserAdmin(UserAdmin):
    add_fieldsets = (
        (None, {
            'classes': ('wide',),
            'fields': ('username', 'password1', 'password2', 'mobile', 'email'),
        }),
    )

admin.site.register(models.User, MyUserAdmin)

群查接口各种筛选组件数据准备

models.py

class Car(models.Model):
    name = models.CharField(max_length=16, unique=True, verbose_name='车名')
    price = models.DecimalField(max_digits=10, decimal_places=2, verbose_name='价格')
    brand = models.CharField(max_length=16, verbose_name='品牌')

    class Meta:
        db_table = 'api_car'
        verbose_name = '汽车表'
        verbose_name_plural = verbose_name

    def __str__(self):
        return self.name

admin.py

admin.site.register(models.Car)

serializers.py

class CarModelSerializer(serializers.ModelSerializer):
    class Meta:
        model = models.Car
        fields = ['name', 'price', 'brand']

views.py

# Car的群查接口
from rest_framework.generics import ListAPIView

class CarListAPIView(ListAPIView):
    queryset = models.Car.objects.all()
    serializer_class = serializers.CarModelSerializer

urls.py

url(r'^cars/$', views.CarListAPIView.as_view()),

drf搜索过滤组件

views.py

from rest_framework.generics import ListAPIView

# 第一步：drf的SearchFilter - 搜索过滤
from rest_framework.filters import SearchFilter

class CarListAPIView(ListAPIView):
    queryset = models.Car.objects.all()
    serializer_class = serializers.CarModelSerializer

    # 第二步：局部配置 过滤类 们（全局配置用DEFAULT_FILTER_BACKENDS）
    filter_backends = [SearchFilter]

    # 第三步：SearchFilter过滤类依赖的过滤条件 => 接口：/cars/?search=...
    search_fields = ['name', 'price']
    # eg：/cars/?search=1，name和price中包含1的数据都会被查询出

drf排序过滤组件

views.py

from rest_framework.generics import ListAPIView

# 第一步：drf的OrderingFilter - 排序过滤
from rest_framework.filters import OrderingFilter

class CarListAPIView(ListAPIView):
    queryset = models.Car.objects.all()
    serializer_class = serializers.CarModelSerializer

    # 第二步：局部配置 过滤类 们（全局配置用DEFAULT_FILTER_BACKENDS）
    filter_backends = [OrderingFilter]

    # 第三步：OrderingFilter过滤类依赖的过滤条件 => 接口：/cars/?ordering=...
    ordering_fields = ['pk', 'price']
    # eg：/cars/?ordering=-price,pk，先按price降序，如果出现price相同，再按pk升序

drf基础分页组件

pahenations.py

from rest_framework.pagination import PageNumberPagination

class MyPageNumberPagination(PageNumberPagination):
    # ?page=页码
    page_query_param = 'page'
    # ?page=页面 下默认一页显示的条数
    page_size = 3
    # ?page=页面&page_size=条数 用户自定义一页显示的条数
    page_size_query_param = 'page_size'
    # 用户自定义一页显示的条数最大限制：数值超过5也只显示5条
    max_page_size = 5

views.py

from rest_framework.generics import ListAPIView

class CarListAPIView(ListAPIView):
    # 如果queryset没有过滤条件，就必须 .all()，不然分页会出问题
    queryset = models.Car.objects.all()
    serializer_class = serializers.CarModelSerializer
    
    # 分页组件 - 给视图类配置分页类即可 - 分页类需要自定义，继承drf提供的分页类即可
    pagination_class = pagenations.MyPageNumberPagination