<!-- https://mvnrepository.com/artifact/com.google.guava/guava -->
<dependency>
    <groupId>com.google.guava</groupId>
    <artifactId>guava</artifactId>
    <version>25.1-jre</version>
</dependency>

 1 /**
 2    * Creates a {@link BloomFilter BloomFilter<T>} with the expected number of
 3    * insertions and expected false positive probability.
 4    *
 5    * <p>Note that overflowing a {@code BloomFilter} with significantly more elements
 6    * than specified, will result in its saturation, and a sharp deterioration of its
 7    * false positive probability.
 8    *
 9    * <p>The constructed {@code BloomFilter<T>} will be serializable if the provided
10    * {@code Funnel<T>} is.
11    *
12    * <p>It is recommended that the funnel be implemented as a Java enum. This has the
13    * benefit of ensuring proper serialization and deserialization, which is important
14    * since {@link #equals} also relies on object identity of funnels.
15    *
16    * @param funnel the funnel of T\'s that the constructed {@code BloomFilter<T>} will use
17    * @param expectedInsertions the number of expected insertions to the constructed
18    *     {@code BloomFilter<T>}; must be positive
19    * @param fpp the desired false positive probability (must be positive and less than 1.0)
20    * @return a {@code BloomFilter}
21    */
22   public static <T> BloomFilter<T> create(
23       Funnel<T> funnel, int expectedInsertions /* n */, double fpp) {
24     checkNotNull(funnel);
25     checkArgument(expectedInsertions >= 0, "Expected insertions (%s) must be >= 0",
26         expectedInsertions);
27     checkArgument(fpp > 0.0, "False positive probability (%s) must be > 0.0", fpp);
28     checkArgument(fpp < 1.0, "False positive probability (%s) must be < 1.0", fpp);
29     if (expectedInsertions == 0) {
30       expectedInsertions = 1;
31     }
32     /*
33      * TODO(user): Put a warning in the javadoc about tiny fpp values,
34      * since the resulting size is proportional to -log(p), but there is not
35      * much of a point after all, e.g. optimalM(1000, 0.0000000000000001) = 76680
36      * which is less than 10kb. Who cares!
37      */
38     long numBits = optimalNumOfBits(expectedInsertions, fpp);
39     int numHashFunctions = optimalNumOfHashFunctions(expectedInsertions, numBits);
40     try {
41       return new BloomFilter<T>(new BitArray(numBits), numHashFunctions, funnel,
42           BloomFilterStrategies.MURMUR128_MITZ_32);
43     } catch (IllegalArgumentException e) {
44       throw new IllegalArgumentException("Could not create BloomFilter of " + numBits + " bits", e);
45     }
46   }

1 /**
2    * Returns {@code true} if the element <i>might</i> have been put in this Bloom filter,
3    * {@code false} if this is <i>definitely</i> not the case.
4    */
5   public boolean mightContain(T object) {
6     return strategy.mightContain(object, funnel, numHashFunctions, bits);
7   }

 1 public static void main(String... args){
 2         /**
 3          * 创建一个插入对象为一亿，误报率为0.01%的布隆过滤器
 4          */
 5         BloomFilter<CharSequence> bloomFilter = BloomFilter.create(Funnels.stringFunnel(Charset.forName("utf-8")), 100000000, 0.0001);
 6         bloomFilter.put("121");
 7         bloomFilter.put("122");
 8         bloomFilter.put("123");
 9         System.out.println(bloomFilter.mightContain("121"));
10     }