布隆过滤器用Redisson实现的例子

女爷i 2023-02-28 01:20 3阅读 0赞

### 背景介绍 ###

之前这边有写过关于布隆过滤器过滤器的安装使用，但是这次使用使用jrebloom来实现，  
jrebloom目前只提供一个maven依赖供我们使用。  
![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L2ZhamluZ19mZWl5dWU_size_16_color_FFFFFF_t_70]

考虑到后续代码维护，bug的修复。使用redisson来做布隆过滤器是更加安全做法。  
![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L2ZhamluZ19mZWl5dWU_size_16_color_FFFFFF_t_70 1]  
下面是关于布隆过滤器的使用：  
[布隆过滤器的原理、redis布隆过滤器的安装和使用][redis]

### 代码实现 ###

下面使用Redisson对布隆过滤器进行的封装。  
1、引入maven依赖

<dependency>
                <groupId>org.redisson</groupId>
                <artifactId>redisson</artifactId>
                <version>3.11.2</version>
            </dependency>

2、关于redis的配置

redis配置：

#客户端超时时间单位是毫秒 默认是2000
    redis.timeout=10000
    #最大空闲数
    redis.maxIdle=300
    #连接池的最大数据库连接数。设为0表示无限制,如果是jedis 2.4以后用redis.maxTotal
    #redis.maxActive=600
    #控制一个pool可分配多少个jedis实例,用来替换上面的redis.maxActive,如果是jedis 2.4以后用该属性
    redis.maxTotal=2000
    #最大建立连接等待时间。如果超过此时间将接到异常。设为-1表示无限制。
    redis.maxWaitMillis=1000
    
    redis.nodes=192.168.25.128:7000,192.168.25.128:7001,192.168.25.128:7002,192.168.25.128:7003,192.168.25.128:7004,192.168.25.128:7005,192.168.25.128:7006,192.168.25.128:7007

3、配置文件的书写

@Configuration
    @PropertySource("classpath:conf/redis.properties")
    public class RedisConfig {
    
    
        @Value("${redis.nodes}")
        private String clusterNodes;
    
    
        @Bean
        public RedissonClient getBloomFilter(){
            Config config = new Config();
            ClusterServersConfig clusterServersConfig = config.useClusterServers();
            String[] cNodes = clusterNodes.split(",");
            //分割出集群节点
            for (String node : cNodes) {
                clusterServersConfig.addNodeAddress("redis://" + node);
    
            }
            return Redisson.create(config);
        }
    
    
    }

4、service的的代码

public interface RedissonBloomFilterService {
        /**
         * 初始化一个布隆过滤器
         *
         * @param key        key
         * @param expireDate 过期时间长度
         * @param timeUnit   过期时间单位
         * @return 是否创建成功
         */
        boolean initNewFilter(String key, Long expireDate, TimeUnit timeUnit, long expectedInsertions, double falseProbability);
    
        /**
         * 添加value到今天布隆过滤器
         * @param key  key
         * @param value  需要添加到过滤器的值
         * @return  是否添加成功
         */
        boolean add(final String key, final String value);
    
        /**
         * 添加value到今天布隆过滤器
         * @param key  key
         * @param list  需要添加到过滤器的值
         * @return  是否添加成功
         */
        boolean add(final String key, final List<String> list);
    
        /**
         * 判断对应key是否有value值
         * @param key  key
         * @param value 对应key的value值
         * @return 应key是否有value值
         */
        boolean containsValue(final String key, final String value);
    
        /**
         * 查看是否存在key值
         * @param key  key
         * @return 是否有key值
         */
        boolean containKey(final String key);
    
    }

实现类：

@Slf4j
    @Service
    public class RedissonBloomServiceImpl implements RedissonBloomFilterService {
    
        public static String prefix = "redission_bf_";
    
    
        private static DateTimeFormatter formatter = DateTimeFormatter.ofPattern("yyyyMMdd");
        private static final int TIME_FORMAT_LENGTH = 8;
    
        @Resource
        private RedissonClient redissonClient;
        /**
         * 1、第一层的key为传入key值
         * 2、第二次key为  prefix + key+20200718
         */
        private Map<String, ConcurrentHashMap<String, RBloomFilter>> bloomFilterMap = new ConcurrentHashMap<>();
    
    @Override
        public synchronized boolean initNewFilter(String key, Long expireDate, TimeUnit timeUnit, long expectedInsertions, double falseProbability) {
    
            String timeFormatter = formatter.format(LocalDateTime.now());
            RBloomFilter<String> bloomFilter = redissonClient.getBloomFilter(prefix + timeFormatter + key);
            //初始化布隆过滤器
            boolean isSuccess = bloomFilter.tryInit(expectedInsertions, falseProbability);
    
            if (isSuccess) {
                //设置过期时间
                bloomFilter.expire(expireDate, timeUnit);
    
                //设置过期监听器
                bloomFilter.addListener(new ExpiredObjectListener() {
                    @Override
                    public void onExpired(String key) {
                        log.info("onExpired callback running key is {}", key);
                        if (key.startsWith(prefix)) {
                            String firstKey = key.substring(prefix.length() + TIME_FORMAT_LENGTH);
                            ConcurrentHashMap<String, RBloomFilter> map = bloomFilterMap.get(firstKey);
                            if (Objects.isNull(map)) {
                                log.warn("listener callback key is empty key is {}", key);
                                return;
                            }
                            map.remove(key);
                        }
                    }
                });
    
            } 
            
            ConcurrentHashMap<String, RBloomFilter> map = bloomFilterMap.computeIfAbsent(key, (k) -> {
                return new ConcurrentHashMap<String, RBloomFilter>(4);
            });
            map.put(prefix + timeFormatter + key, bloomFilter);
            isSuccess = true;
    
            return isSuccess;
        }
    
        @Override
        public boolean add(String key, String value) {
            String timeFormatter = formatter.format(LocalDateTime.now());
            ConcurrentHashMap<String, RBloomFilter> map = bloomFilterMap.get(key);
            if (Objects.isNull(map)) {
                log.error("add name key one value is null, name is {},value is {}", key, value);
                return false;
            }
            RBloomFilter rBloomFilter = map.get(prefix + timeFormatter + key);
            if (Objects.isNull(rBloomFilter)) {
                log.error("add name key one value is null, name is {},redis key is {},value is {}", key, prefix + timeFormatter + key, value);
                return false;
            }
            return rBloomFilter.add(value);
        }
    
        @Override
        public boolean add(String key, List<String> list) {
            String timeFormatter = formatter.format(LocalDateTime.now());
            ConcurrentHashMap<String, RBloomFilter> map = bloomFilterMap.get(key);
            if (Objects.isNull(map)) {
                log.error("add name key list value is null, name is {}", key);
                return false;
            }
            RBloomFilter rBloomFilter = map.get(prefix + timeFormatter + key);
            if (Objects.isNull(rBloomFilter)) {
                log.error("add name key list value is null, name is {},redis key is {}", key,prefix + timeFormatter + key);
                return false;
            }
            list.forEach(value -> rBloomFilter.add(value));
            return true;
        }
    
        @Override
        public boolean containsValue(String key, String value) {
            ConcurrentHashMap<String, RBloomFilter> map = bloomFilterMap.get(key);
            if (Objects.isNull(map)) {
                log.error("containsValue key is exist, key is {}", key);
                return false;
            }
            for (Map.Entry<String, RBloomFilter> entry : map.entrySet()) {
                if (entry.getValue().contains(value)) {
                    return true;
                }
            }
            return false;
        }
    
    
        @Override
        public boolean containKey(String key) {
            return bloomFilterMap.containsKey(key);
        }
    
    }

测试类：（这里的测试，有关于已经存在值的判断）

@Test
        public void testRedissonBloom() throws InterruptedException {
            //创建过滤器
            redissonBloomFilterService.initNewFilter("bloomkey1", 60L, SECONDS, 50, 0.1);
            redissonBloomFilterService.initNewFilter("bloomkey2", 60L, SECONDS, 50, 0.1);
    
            for (int i = 0; i < 50; i++) {
                redissonBloomFilterService.add("bloomkey1", "" + i);
                if (i > 30) {
                    redissonBloomFilterService.add("bloomkey2", "" + i);
                }
            }
            int j1 = 0;
            int j2 = 0;
            for (int i = 0; i < 100; i++) {
                boolean exists = redissonBloomFilterService.containsValue("bloomkey1", "" + i);
                if (exists) {
                    System.out.println("bloomkey1有误判" + (++j1) + "i是" + i);
                }
                boolean exists2 = redissonBloomFilterService.containsValue("bloomkey2", "" + i);
                if (exists2) {
                    System.out.println("bloomkey2" +
                            "有误判" + (++j2) + "i是" + i);
                }
            }
    
            CountDownLatch beginCount = new CountDownLatch(1);
            beginCount.await();
    
        }

[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L2ZhamluZ19mZWl5dWU_size_16_color_FFFFFF_t_70]: /images/20230209/acfe0efc79e649d59d957fdaf6629e9b.png
[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L2ZhamluZ19mZWl5dWU_size_16_color_FFFFFF_t_70 1]: /images/20230209/6dc055f7a8f44c459e7fbcc17d7fb7d1.png
[redis]: https://blog.csdn.net/fajing_feiyue/article/details/105469422