Fork me on GitHub

Go语言学习——如何实现一个过滤器

1、过滤器使用场景

  做业务的时候我们经常要使用过滤器或者拦截器(听这口音就是从Java过来的)。常见的场景如一个HTTP请求,需要经过鉴权过滤器、白名单校验过滤、参数验证过滤器等重重关卡最终拿到数据。

  Java使用过滤器很简单。XML时代,只要添加一个过滤器配置再新建一个实现了Filter接口的xxxFilter实现类;Java Configuration时代,只要在xxxConfiguration配置类中声明一个Filter注解,如果想设置Filter的执行顺序,加上Order注解就行了。

  Java的过滤器实在太方便也太好用了。

  以至于在Java有关过滤器的面试题中,只有类似于“过滤器的使用场景有哪些?”,“过滤器和拦截器有什么区别?“,几乎很少听到”你知道过滤器是怎么实现的吗?“,”如果让你实现一个过滤器,你会怎么做?“这样的题目。

 

2、使用过滤器的场景特征

如同上面过滤器的例子,我们发现过滤器有一些特征:

  1、入参一样,比如HTTP请求的过滤器的入参就是ServletRequest对象

  2、返回值类型相同,比如都是true或者false,或者是链接到下一个过滤器或者return。

如下是Java实现的CORS过滤器

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
import org.springframework.http.HttpStatus;
import org.springframework.util.StringUtils;
 
import javax.servlet.*;
import javax.servlet.http.HttpServletRequest;
import javax.servlet.http.HttpServletResponse;
import java.io.IOException;
 
public class CORSFilter implements Filter {
 
    @Override
    public void doFilter(ServletRequest reserRealmq, ServletResponse res, FilterChain chain) throws IOException, ServletException {
        HttpServletRequest request = (HttpServletRequest) reserRealmq;
        HttpServletResponse response = (HttpServletResponse) res;
 
        String currentOrigin= request.getHeader("Origin");
        if (!StringUtils.isEmpty(currentOrigin)) {
            response.setHeader("Access-Control-Allow-Origin", currentOrigin);
            response.setHeader("Access-Control-Allow-Methods", "POST, GET, OPTIONS, DELETE, PUT");
            response.setHeader("Access-Control-Allow-Credentials", "true");
            response.setHeader("Access-Control-Allow-Headers", "Origin, No-Cache, X-Requested-With, If-Modified-Since, Cache-Control, Expires, Content-Type, X-E4M-With, Index-Url");
        }
 
        // return http status 204 if OPTIONS requst
        if ("OPTIONS".equals(request.getMethod())){
            response.setStatus(HttpStatus.NO_CONTENT.value());
        }else {
            chain.doFilter(reserRealmq, res);
        }
    }
 
    @Override
    public void init(FilterConfig filterConfig) throws ServletException {
 
    }
 
    @Override
    public void destroy() {
 
    }
}

  

 

凡是具有这种特征的需求,我们都可以抽象为过滤器进行实现(Java里面称为责任链模式)。

 

下面就来说说,基于Go语言如何实现一个过滤器。

 

3、简单实现

  过滤器本质就是一堆条件判定,最直观的过滤方案就是创建几个方法,针对每个方法的返回结果判定,如果返回为false则终止请求,如果为true则继续执行下一个过滤器。

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
package main
 
import (
    "context"
)
 
func main() {
    ctx := context.TODO()
   
  if continued := F1(ctx); !continued {
    ...
    return
  }
   
  if continued := F2(ctx); !continued {
    ...
    return
  }
   
  if continued := F3(ctx); !continued {
    ...
    return
  }
}
 
func F1(ctx context.Context) bool {
  ...
  return true
}
 
func F2(ctx context.Context) bool {
  ...
  return true
}
 
func F3(ctx context.Context) bool {
  ...
  return false
}

  

该版本从功能上说,完全符合过滤器的要求。

但是从代码层面来说,有几个问题:

  1、复用性较差。main函数中对于各个过滤器的判定,除了函数名不一样,其他逻辑都一样,可以考虑抽象重用。

  2、可扩展性较差。因为有些代码复用性差,导致代码不好扩展,如果这时候添加、删除过滤器或者调整过滤器执行顺序,代码都需要较大改动才能实现。

  3、难以维护。不用多说。

 

4、重构实现

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
package main
 
import (
    "context"
    "fmt"
)
 
type MyContext struct {
    context.Context
    KeyValue map[string]bool
}
 
type FilterFunc func(*MyContext) bool
 
type FilterFuncChain []FilterFunc
 
type CombinedFunc struct {
    CF    FilterFuncChain
    MyCtx *MyContext
}
 
func main() {
    myContext := MyContext{Context: context.TODO(), KeyValue: map[string]bool{"key": false}}
 
    cf := CombinedFilter(&myContext, F1, F2, F3);
    DoFilter(cf)
}
 
func DoFilter(cf *CombinedFunc) {
    for _, f := range cf.CF {
        res := f(cf.MyCtx)
        fmt.Println("result:", res)
        if res == false {
            fmt.Println("stopped")
            return
        }
    }
}
 
func CombinedFilter(ctx *MyContext, ff ...FilterFunc) *CombinedFunc {
    return &CombinedFunc{
        CF:    ff,
        MyCtx: ctx,
    }
}
 
func F1(ctx *MyContext) bool {
    ctx.KeyValue["key"] = true
    fmt.Println(ctx.KeyValue["key"])
 
    return ctx.KeyValue["key"]
}
 
func F2(ctx *MyContext) bool {
    ctx.KeyValue["key"] = false
    fmt.Println(ctx.KeyValue["key"])
 
    return ctx.KeyValue["key"]
}
 
func F3(ctx *MyContext) bool {
    ctx.KeyValue["key"] = false
    fmt.Println(ctx.KeyValue["key"])
 
    return ctx.KeyValue["key"]
}

代码不长,我们一块块分析。

 

4.1 自定义的Context

  这里我使用了自定义的Context,重新定义一个MyContext的结构体,其中组合了标准库中的Context,即具备标准库Context的能力。

  这里MyContext是作为数据载体在各个过滤器之间传递。没有用标准库的Context,采用自定义的Context主要是为了说明我们可以根据需要扩展MyContext,通过扩展MyContext添加任何我们需要的参数。这里添加的是一个map键值对。我们可以将每个过滤器处理的结果存入这个map中,再传递到下一个过滤器。

1
myContext := MyContext{Context: context.TODO(), KeyValue: map[string]bool{"key": false}}

上面的等价写法还可以是

1
2
ctx := context.TODO()
myContext := context.WithValue(ctx, "key", "value")

这里充分利用了Context的WithValue的用法,有兴趣可以去看下,这是Context创建map键值对的方式。

 

4.2 充分利用Go的type的特性

 

1
type FilterFunc func(*MyContext) bool

 

  前面在使用过滤的场景特种中提到,过滤器的入参和返回值都是一样的。所以这里我们利用Go的type特性,将这种过滤器函数定义为一个变量FilterFunc

  这一特性对于精简代码起到了关键性的作用。且看

1
2
3
4
5
6
7
8
cf := CombinedFilter(&myContext, F1, F2, F3);
 
func CombinedFilter(ctx *MyContext, ff ...FilterFunc) *CombinedFunc {
    return &CombinedFunc{
        CF:    ff,
        MyCtx: ctx,
    }
}

因为这里的F1、F2和F3都有相同入参和返回值,所以抽象为FilterFunc,并使用变长参数的FilterFunc统一接收。

CombinedFilter不仅可以加F1、F2和F3,后面还可以有F4、F5...

 

1
type FilterFuncChain []FilterFunc

  这里的抽象也是同样的道理。

  如果之前写过Java,这里是不是已经看到了Filter接口的影子。其实这里的FilterFunc可以等价于Java里面的Filter接口,接口是一种约束一种契约,Filter定义了如果要实现该接口必须要实现接口定义的方法。

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
package javax.servlet;
 
import java.io.IOException;
 
/**
 * A FilterChain is an object provided by the servlet container to the developer
 * giving a view into the invocation chain of a filtered request for a resource.
 * Filters use the FilterChain to invoke the next filter in the chain, or if the
 * calling filter is the last filter in the chain, to invoke the resource at the
 * end of the chain.
 *
 * @see Filter
 * @since Servlet 2.3
 **/
 
public interface FilterChain {
 
    /**
     * Causes the next filter in the chain to be invoked, or if the calling
     * filter is the last filter in the chain, causes the resource at the end of
     * the chain to be invoked.
     *
     * @param request
     *            the request to pass along the chain.
     * @param response
     *            the response to pass along the chain.
     *
     * @throws IOException if an I/O error occurs during the processing of the
     *                     request
     * @throws ServletException if the processing fails for any other reason
 
     * @since 2.3
     */
    public void doFilter(ServletRequest request, ServletResponse response)
            throws IOException, ServletException;
 
}

  

4.3 遍历执行过滤器

  因为有了上面的特性,我们才能将这些过滤器存入切片然后依次执行,如下

1
2
3
4
5
6
7
8
9
10
func DoFilter(cf *CombinedFunc) {
    for _, f := range cf.CF {
        res := f(cf.MyCtx)
        fmt.Println("result:", res)
        if res == false {
            fmt.Println("stopped")
            return
        }
    }
}

在执行的过程中,如果我们发现如果返回值为false,则表示没有通过某个过滤器校验,则退出也不会继续执行后面的过滤器。

 

5、继续改进

既然MyContext中的map集合可以存储各个Filter的执行情况,而且可以在各个过滤器之间传递,我们甚至可以省略FilterFunc函数的返回值,改进后如下

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
package main
 
import (
    "context"
    "fmt"
)
 
type MyContext struct {
    context.Context
    KeyValue map[string]bool
}
 
type FilterFunc func(*MyContext)
 
type FilterFuncChain []FilterFunc
 
type CombinedFunc struct {
    CF    FilterFuncChain
    MyCtx *MyContext
}
 
func main() {
    myContext := MyContext{Context: context.TODO(), KeyValue: map[string]bool{"key": false}}
 
    cf := CombinedFilter(&myContext, F1, F2, F3);
    DoFilter(cf)
}
 
func DoFilter(cf *CombinedFunc) {
    for _, f := range cf.CF {
        f(cf.MyCtx)
        continued :=  cf.MyCtx.KeyValue["key"]
        fmt.Println("result:", continued)
        if !continued {
            fmt.Println("stopped")
            return
        }
    }
}
 
func CombinedFilter(ctx *MyContext, ff ...FilterFunc) *CombinedFunc {
    return &CombinedFunc{
        CF:    ff,
        MyCtx: ctx,
    }
}
 
func F1(ctx *MyContext) {
    ctx.KeyValue["key"] = true
    fmt.Println(ctx.KeyValue["key"])
    //return ctx.KeyValue["key"]
}
 
func F2(ctx *MyContext) {
    ctx.KeyValue["key"] = false
    fmt.Println(ctx.KeyValue["key"])
    //return ctx.KeyValue["key"]
}
 
func F3(ctx *MyContext) {
    ctx.KeyValue["key"] = false
    fmt.Println(ctx.KeyValue["key"])
    //return ctx.KeyValue["key"]
}

  

6、总结

基于Go语言造轮子实现一个过滤器的雏形,通过实现一个相对优雅可扩展的过滤器熟悉了type的用法,Context.WithValue的作用。

 

如果您觉得阅读本文对您有帮助,请点一下“推荐”按钮,您的“推荐”将是我最大的写作动力!如果您想持续关注我的文章,请扫描二维码,关注JackieZheng的微信公众号,我会将我的文章推送给您,并和您一起分享我日常阅读过的优质文章。

posted @   JackieZheng  阅读(3627)  评论(0编辑  收藏  举报
编辑推荐:
· 后端思维之高并发处理方案
· 理解Rust引用及其生命周期标识(下)
· 从二进制到误差:逐行拆解C语言浮点运算中的4008175468544之谜
· .NET制作智能桌面机器人:结合BotSharp智能体框架开发语音交互
· 软件产品开发中常见的10个问题及处理方法
阅读排行:
· 2025成都.NET开发者Connect圆满结束
· 后端思维之高并发处理方案
· 千万级大表的优化技巧
· 在 VS Code 中,一键安装 MCP Server!
· 10年+ .NET Coder 心语 ── 继承的思维:从思维模式到架构设计的深度解析
历史上的今天:
2017-08-30 Spring集成RabbitMQ-必须知道的几个概念
点击右上角即可分享
微信分享提示