com.google.common.hash.Hashing Maven / Gradle / Ivy
/*
* Copyright (C) 2011 The Guava Authors
*
* Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except
* in compliance with the License. You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software distributed under the License
* is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express
* or implied. See the License for the specific language governing permissions and limitations under
* the License.
*/
package com.google.common.hash;
import static com.google.common.base.Preconditions.checkArgument;
import com.google.common.annotations.Beta;
import com.google.common.annotations.VisibleForTesting;
import com.google.common.base.Supplier;
import java.nio.ByteBuffer;
import java.security.MessageDigest;
import java.util.Iterator;
import java.util.zip.Adler32;
import java.util.zip.CRC32;
import java.util.zip.Checksum;
/**
* Static methods to obtain {@link HashFunction} instances, and other static hashing-related
* utilities.
*
* @author Kevin Bourrillion
* @author Dimitris Andreou
* @author Kurt Alfred Kluever
* @since 11.0
*/
@Beta
public final class Hashing {
private Hashing() {}
/**
* Used to randomize {@link #goodFastHash} instances, so that programs which persist anything
* dependent on hashcodes of those, will fail sooner than later.
*/
private static final int GOOD_FAST_HASH_SEED = (int) System.currentTimeMillis();
// Used by goodFastHash when minimumBits == 32.
private static final HashFunction GOOD_FAST_HASH_FUNCTION_32 = murmur3_32(GOOD_FAST_HASH_SEED);
// Used by goodFastHash when 32 < minimumBits <= 128.
private static final HashFunction GOOD_FAST_HASH_FUNCTION_128 = murmur3_128(GOOD_FAST_HASH_SEED);
/**
* Returns a general-purpose, non-cryptographic-strength, streaming hash function that
* produces hash codes of length at least {@code minimumBits}. Users without specific
* compatibility requirements and who do not persist the hash codes are encouraged to
* choose this hash function.
*
* Repeated calls to {@link #goodFastHash} with the same {@code minimumBits} value will
* return {@link HashFunction} instances with identical behavior (but not necessarily the
* same instance) for the duration of the current virtual machine.
*
*
Warning: the implementation is unspecified and is subject to change.
*
* @throws IllegalArgumentException if {@code minimumBits} is not positive
*/
public static HashFunction goodFastHash(int minimumBits) {
int bits = checkPositiveAndMakeMultipleOf32(minimumBits);
if (bits == 32) {
return GOOD_FAST_HASH_FUNCTION_32;
}
if (bits <= 128) {
return GOOD_FAST_HASH_FUNCTION_128;
}
// Otherwise, join together some 128-bit murmur3s
int hashFunctionsNeeded = (bits + 127) / 128;
HashFunction[] hashFunctions = new HashFunction[hashFunctionsNeeded];
hashFunctions[0] = GOOD_FAST_HASH_FUNCTION_128;
int seed = GOOD_FAST_HASH_SEED;
for (int i = 1; i < hashFunctionsNeeded; i++) {
seed += 1500450271; // a prime; shouldn't matter
hashFunctions[i] = murmur3_128(seed);
}
return new ConcatenatedHashFunction(hashFunctions);
}
/**
* Returns a hash function implementing the
*
* 32-bit murmur3 algorithm, x86 variant (little-endian variant),
* using the given seed value.
*
*
The exact C++ equivalent is the MurmurHash3_x86_32 function (Murmur3A).
*/
public static HashFunction murmur3_32(int seed) {
return new Murmur3_32HashFunction(seed);
}
/**
* Returns a hash function implementing the
*
* 32-bit murmur3 algorithm, x86 variant (little-endian variant),
* using a seed value of zero.
*
*
The exact C++ equivalent is the MurmurHash3_x86_32 function (Murmur3A).
*/
public static HashFunction murmur3_32() {
return MURMUR3_32;
}
private static final HashFunction MURMUR3_32 = new Murmur3_32HashFunction(0);
/**
* Returns a hash function implementing the
*
* 128-bit murmur3 algorithm, x64 variant (little-endian variant),
* using the given seed value.
*
*
The exact C++ equivalent is the MurmurHash3_x64_128 function (Murmur3F).
*/
public static HashFunction murmur3_128(int seed) {
return new Murmur3_128HashFunction(seed);
}
/**
* Returns a hash function implementing the
*
* 128-bit murmur3 algorithm, x64 variant (little-endian variant),
* using a seed value of zero.
*
*
The exact C++ equivalent is the MurmurHash3_x64_128 function (Murmur3F).
*/
public static HashFunction murmur3_128() {
return MURMUR3_128;
}
private static final HashFunction MURMUR3_128 = new Murmur3_128HashFunction(0);
/**
* Returns a hash function implementing the MD5 hash algorithm (128 hash bits) by delegating to
* the MD5 {@link MessageDigest}.
*/
public static HashFunction md5() {
return MD5;
}
private static final HashFunction MD5 = new MessageDigestHashFunction("MD5", "Hashing.md5()");
/**
* Returns a hash function implementing the SHA-1 algorithm (160 hash bits) by delegating to the
* SHA-1 {@link MessageDigest}.
*/
public static HashFunction sha1() {
return SHA_1;
}
private static final HashFunction SHA_1 =
new MessageDigestHashFunction("SHA-1", "Hashing.sha1()");
/**
* Returns a hash function implementing the SHA-256 algorithm (256 hash bits) by delegating to
* the SHA-256 {@link MessageDigest}.
*/
public static HashFunction sha256() {
return SHA_256;
}
private static final HashFunction SHA_256 =
new MessageDigestHashFunction("SHA-256", "Hashing.sha256()");
/**
* Returns a hash function implementing the SHA-512 algorithm (512 hash bits) by delegating to the
* SHA-512 {@link MessageDigest}.
*/
public static HashFunction sha512() {
return SHA_512;
}
private static final HashFunction SHA_512 =
new MessageDigestHashFunction("SHA-512", "Hashing.sha512()");
/**
* Returns a hash function implementing the CRC-32 checksum algorithm (32 hash bits) by delegating
* to the {@link CRC32} {@link Checksum}.
*
*
To get the {@code long} value equivalent to {@link Checksum#getValue()} for a
* {@code HashCode} produced by this function, use {@link HashCode#padToLong()}.
*
* @since 14.0
*/
public static HashFunction crc32() {
return CRC_32;
}
private static final HashFunction CRC_32 =
checksumHashFunction(ChecksumType.CRC_32, "Hashing.crc32()");
/**
* Returns a hash function implementing the Adler-32 checksum algorithm (32 hash bits) by
* delegating to the {@link Adler32} {@link Checksum}.
*
*
To get the {@code long} value equivalent to {@link Checksum#getValue()} for a
* {@code HashCode} produced by this function, use {@link HashCode#padToLong()}.
*
* @since 14.0
*/
public static HashFunction adler32() {
return ADLER_32;
}
private static final HashFunction ADLER_32 =
checksumHashFunction(ChecksumType.ADLER_32, "Hashing.adler32()");
private static HashFunction checksumHashFunction(ChecksumType type, String toString) {
return new ChecksumHashFunction(type, type.bits, toString);
}
enum ChecksumType implements Supplier {
CRC_32(32) {
@Override
public Checksum get() {
return new CRC32();
}
},
ADLER_32(32) {
@Override
public Checksum get() {
return new Adler32();
}
};
private final int bits;
ChecksumType(int bits) {
this.bits = bits;
}
@Override
public abstract Checksum get();
}
// Lazy initialization holder class idiom.
/**
* If {@code hashCode} has enough bits, returns {@code hashCode.asLong()}, otherwise
* returns a {@code long} value with {@code hashCode.asInt()} as the least-significant
* four bytes and {@code 0x00} as each of the most-significant four bytes.
*
* @deprecated Use {@code HashCode.padToLong()} instead. This method is scheduled to be
* removed in Guava 15.0.
*/
@Deprecated
public static long padToLong(HashCode hashCode) {
return hashCode.padToLong();
}
/**
* Assigns to {@code hashCode} a "bucket" in the range {@code [0, buckets)}, in a uniform
* manner that minimizes the need for remapping as {@code buckets} grows. That is,
* {@code consistentHash(h, n)} equals:
*
*
* - {@code n - 1}, with approximate probability {@code 1/n}
*
- {@code consistentHash(h, n - 1)}, otherwise (probability {@code 1 - 1/n})
*
*
* See the wikipedia
* article on consistent hashing for more information.
*
* If you might want to have weights for the buckets in the future, take a look at
* {@code weightedConsistentHash}.
*/
public static int consistentHash(HashCode hashCode, int buckets) {
return consistentHash(hashCode.padToLong(), buckets);
}
/**
* Assigns to {@code input} a "bucket" in the range {@code [0, buckets)}, in a uniform
* manner that minimizes the need for remapping as {@code buckets} grows. That is,
* {@code consistentHash(h, n)} equals:
*
*
* - {@code n - 1}, with approximate probability {@code 1/n}
*
- {@code consistentHash(h, n - 1)}, otherwise (probability {@code 1 - 1/n})
*
*
* See the wikipedia
* article on consistent hashing for more information.
*
* If you might want to have weights for the buckets in the future, take a look at
* {@code weightedConsistentHash}.
*/
public static int consistentHash(long input, int buckets) {
checkArgument(buckets > 0, "buckets must be positive: %s", buckets);
LinearCongruentialGenerator generator = new LinearCongruentialGenerator(input);
int candidate = 0;
int next;
// Jump from bucket to bucket until we go out of range
while (true) {
next = (int) ((candidate + 1) / generator.nextDouble());
if (next >= 0 && next < buckets) {
candidate = next;
} else {
return candidate;
}
}
}
/**
* Returns a hash code, having the same bit length as each of the input hash codes,
* that combines the information of these hash codes in an ordered fashion. That
* is, whenever two equal hash codes are produced by two calls to this method, it
* is as likely as possible that each was computed from the same
* input hash codes in the same order.
*
* @throws IllegalArgumentException if {@code hashCodes} is empty, or the hash codes
* do not all have the same bit length
*/
public static HashCode combineOrdered(Iterable hashCodes) {
Iterator iterator = hashCodes.iterator();
checkArgument(iterator.hasNext(), "Must be at least 1 hash code to combine.");
int bits = iterator.next().bits();
byte[] resultBytes = new byte[bits / 8];
for (HashCode hashCode : hashCodes) {
byte[] nextBytes = hashCode.asBytes();
checkArgument(nextBytes.length == resultBytes.length,
"All hashcodes must have the same bit length.");
for (int i = 0; i < nextBytes.length; i++) {
resultBytes[i] = (byte) (resultBytes[i] * 37 ^ nextBytes[i]);
}
}
return HashCodes.fromBytesNoCopy(resultBytes);
}
/**
* Returns a hash code, having the same bit length as each of the input hash codes,
* that combines the information of these hash codes in an unordered fashion. That
* is, whenever two equal hash codes are produced by two calls to this method, it
* is as likely as possible that each was computed from the same
* input hash codes in some order.
*
* @throws IllegalArgumentException if {@code hashCodes} is empty, or the hash codes
* do not all have the same bit length
*/
public static HashCode combineUnordered(Iterable hashCodes) {
Iterator iterator = hashCodes.iterator();
checkArgument(iterator.hasNext(), "Must be at least 1 hash code to combine.");
byte[] resultBytes = new byte[iterator.next().bits() / 8];
for (HashCode hashCode : hashCodes) {
byte[] nextBytes = hashCode.asBytes();
checkArgument(nextBytes.length == resultBytes.length,
"All hashcodes must have the same bit length.");
for (int i = 0; i < nextBytes.length; i++) {
resultBytes[i] += nextBytes[i];
}
}
return HashCodes.fromBytesNoCopy(resultBytes);
}
/**
* Checks that the passed argument is positive, and ceils it to a multiple of 32.
*/
static int checkPositiveAndMakeMultipleOf32(int bits) {
checkArgument(bits > 0, "Number of bits must be positive");
return (bits + 31) & ~31;
}
// TODO(kevinb): Maybe expose this class via a static Hashing method?
@VisibleForTesting
static final class ConcatenatedHashFunction extends AbstractCompositeHashFunction {
private final int bits;
ConcatenatedHashFunction(HashFunction... functions) {
super(functions);
int bitSum = 0;
for (HashFunction function : functions) {
bitSum += function.bits();
}
this.bits = bitSum;
}
@Override
HashCode makeHash(Hasher[] hashers) {
// TODO(user): Get rid of the ByteBuffer here?
byte[] bytes = new byte[bits / 8];
ByteBuffer buffer = ByteBuffer.wrap(bytes);
for (Hasher hasher : hashers) {
buffer.put(hasher.hash().asBytes());
}
return HashCodes.fromBytesNoCopy(bytes);
}
@Override
public int bits() {
return bits;
}
}
/**
* Linear CongruentialGenerator to use for consistent hashing.
* See http://en.wikipedia.org/wiki/Linear_congruential_generator
*/
private static final class LinearCongruentialGenerator {
private long state;
public LinearCongruentialGenerator(long seed) {
this.state = seed;
}
public double nextDouble() {
state = 2862933555777941757L * state + 1;
return ((double) ((int) (state >>> 33) + 1)) / (0x1.0p31);
}
}
}