Hashids 原理及实现

Hashids是一个将数字转化为长度较短、唯一且不连续的值的库。特点是：

• 对非负整数都可以生成唯一短id
• 可以设置不同的盐，具有保密性
• 递增的输入产生的输出无法预测
• 代码较短，且不依赖于第三方库

原理

进制转换：将10进制的整数转化为 62 进制（26个字母大小写+10个数字），可扩展为任意进制。

  private static String hash(long input, String alphabet) {
    String hash = "";
    final int alphabetLen = alphabet.length();

    do {
      final int index = (int) (input % alphabetLen);
      if (index >= 0 && index < alphabet.length()) {
        hash = alphabet.charAt(index) + hash;
      }
      input /= alphabetLen;
    } while (input > 0);

    return hash;
  }

Fisher-Yates Shuffle算法

其基本思想就是从原始数组中随机取一个之前没取过的数字到新的数组中，具体如下：

1. 初始化原始数组和新数组，原始数组长度为n(已知)；
2. 从还没处理的数组（假如还剩k个）中，随机产生一个[0, k)之间的数字p（假设数组从0开始）；
3. 从剩下的k个数中把第p个数取出；
4. 重复步骤2和3直到数字全部取完；
5. 从步骤3取出的数字序列便是一个打乱了的数列。

Hashids 使用了 Fisher–Yates 洗牌算法的变种。从原始数组中随机取一个之前没取过的数字与当前数组最后一个元素交换位置。

private static Random random = new Random();
private static String shuffle(String alphabet) {
    char[] shuffle = alphabet.toCharArray();
    for (int i = shuffle.length - 1; i >= 1; i--) {
        int j = random.nextInt(i+1);
        char tmp = shuffle[i];
        shuffle[i] = shuffle[j];
        shuffle[j] = tmp;
    }
    return new String(shuffle);
}

<dependency>
    <groupId>org.hashids</groupId>
    <artifactId>hashids</artifactId>
    <version>1.0.3</version>
</dependency>

代码解析

import java.util.ArrayList;

public class Hashids {
    // Hashids 能够加密的最大值
    public static final long MAX_NUMBER = 9007199254740992L;
    // 默认编解码字符串
    private static final String DEFAULT_ALPHABET = "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ1234567890";
    // 默认分隔符
    private static final String DEFAULT_SEPS = "cfhistuCFHISTU";
    // 默认盐值
    private static final String DEFAULT_SALT = "";
    // 默认最小加密后字符串的长度
    private static final int DEFAULT_MIN_HASH_LENGTH = 0;
    // 最小编解码字符串
    private static final int MIN_ALPHABET_LENGTH = 16;
    //
    private static final double SEP_DIV = 3.5;
    //
    private static final int GUARD_DIV = 12;
    // 编解码盐值
    private final String salt;
    // 编码后最小的字符长度
    private final int minHashLength;
    // 编解码字符串
    private final String alphabet;
    // 多个数字编解码的分界符
    private final String seps;
    // 补齐至 minHashLength 长度添加的字符列表
    private final String guards;
    // 以上 3 个字符串中的字符是互不包含关系

    public Hashids() {
        this(DEFAULT_SALT);
    }

    public Hashids(String salt) {
        this(salt, 0);
    }

    public Hashids(String salt, int minHashLength) {
        this(salt, minHashLength, DEFAULT_ALPHABET);
    }

    public Hashids(String salt, int minHashLength, String alphabet) {
        this.salt = salt != null ? salt : DEFAULT_SALT;
        this.minHashLength = minHashLength > 0 ? minHashLength : DEFAULT_MIN_HASH_LENGTH;

        // alphabet 中的字符需要唯一
        final StringBuilder uniqueAlphabet = new StringBuilder();
        for (int i = 0; i < alphabet.length(); i++) {
            if (uniqueAlphabet.indexOf(String.valueOf(alphabet.charAt(i))) == -1) {
                uniqueAlphabet.append(alphabet.charAt(i));
            }
        }
        alphabet = uniqueAlphabet.toString();

        // 最小长度检查
        if (alphabet.length() < MIN_ALPHABET_LENGTH) {
            throw new IllegalArgumentException(
                    "alphabet must contain at least " + MIN_ALPHABET_LENGTH + " unique characters");
        }

        // 编解码表不能含有空格
        if (alphabet.contains(" ")) {
            throw new IllegalArgumentException("alphabet cannot contains spaces");
        }

        // seps 只能包含 alphabet 中的字符
        // alphabet 不能包含 seps 中的字符
        String seps = DEFAULT_SEPS;
        for (int i = 0; i < seps.length(); i++) {
            final int j = alphabet.indexOf(seps.charAt(i));
            if (j == -1) {
                seps = seps.substring(0, i) + " " + seps.substring(i + 1);
            } else {
                alphabet = alphabet.substring(0, j) + " " + alphabet.substring(j + 1);
            }
        }
        alphabet = alphabet.replaceAll("\\s+", "");
        seps = seps.replaceAll("\\s+", "");

        // 根据 salt 打乱默认 seps 的顺序
        seps = Hashids.consistentShuffle(seps, this.salt);

        if ((seps.isEmpty()) || (((float) alphabet.length() / seps.length()) > SEP_DIV)) {
            int seps_len = (int) Math.ceil(alphabet.length() / SEP_DIV);

            if (seps_len == 1) {
                seps_len++;
            }

            if (seps_len > seps.length()) {
                final int diff = seps_len - seps.length();
                seps += alphabet.substring(0, diff);
                alphabet = alphabet.substring(diff);
            } else {
                seps = seps.substring(0, seps_len);
            }
        }

        // 根据 salt 打乱编解码表 alphabet 的顺序
        alphabet = Hashids.consistentShuffle(alphabet, this.salt);
        // use double to round up
        final int guardCount = (int) Math.ceil((double) alphabet.length() / GUARD_DIV);

        String guards;
        if (alphabet.length() < 3) {
            guards = seps.substring(0, guardCount);
            seps = seps.substring(guardCount);
        } else {
            guards = alphabet.substring(0, guardCount);
            alphabet = alphabet.substring(guardCount);
        }
        this.guards = guards;
        this.alphabet = alphabet;
        this.seps = seps;
    }

    // 将数字加密为字符串
    public String encode(long... numbers) {
        if (numbers.length == 0) {
            return "";
        }

        for (final long number : numbers) {
            // 只能加解密非负整数
            if (number < 0) {
                return "";
            }
            if (number > MAX_NUMBER) {
                throw new IllegalArgumentException("number can not be greater than " + MAX_NUMBER + "L");
            }
        }
        return this._encode(numbers);
    }

    // 将字符串解密为数字
    public long[] decode(String hash) {
        if (hash.isEmpty()) {
            return new long[0];
        }

        String validChars = this.alphabet + this.guards + this.seps;
        for (int i = 0; i < hash.length(); i++) {
            if (validChars.indexOf(hash.charAt(i)) == -1) {
                return new long[0];
            }
        }

        return this._decode(hash, this.alphabet);
    }

    private String _encode(long... numbers) {
        long numberHashInt = 0;
        for (int i = 0; i < numbers.length; i++) {
            numberHashInt += (numbers[i] % (i + 100));
        }
        String alphabet = this.alphabet;
        // 第一位不参与编解码
        final char ret = alphabet.charAt((int) (numberHashInt % alphabet.length()));

        long num;
        long sepsIndex, guardIndex;
        String buffer;
        final StringBuilder ret_strB = new StringBuilder(this.minHashLength);
        ret_strB.append(ret);
        char guard;

        for (int i = 0; i < numbers.length; i++) {
            num = numbers[i];
            buffer = ret + this.salt + alphabet;

            alphabet = Hashids.consistentShuffle(alphabet, buffer.substring(0, alphabet.length()));
            final String last = Hashids.hash(num, alphabet);

            ret_strB.append(last);

            // 不是最后一个待加密数字，增加分隔符
            // 按照一定规则在 seps 选择一位分隔符，如何选择无所谓，因为 seps 不存在于 alphabet
            if (i + 1 < numbers.length) {
                if (last.length() > 0) {
                    num %= (last.charAt(0) + i);
                    sepsIndex = (int) (num % this.seps.length());
                } else {
                    sepsIndex = 0;
                }
                ret_strB.append(this.seps.charAt((int) sepsIndex));
            }
        }

        // 小于最小长度 minHashLength 则进行扩充
        String ret_str = ret_strB.toString();
        // 扩充两位
        if (ret_str.length() < this.minHashLength) {
            guardIndex = (numberHashInt + (ret_str.charAt(0))) % this.guards.length();
            guard = this.guards.charAt((int) guardIndex);

            ret_str = guard + ret_str;

            if (ret_str.length() < this.minHashLength) {
                guardIndex = (numberHashInt + (ret_str.charAt(2))) % this.guards.length();
                guard = this.guards.charAt((int) guardIndex);

                ret_str += guard;
            }
        }

        final int halfLen = alphabet.length() / 2;
        while (ret_str.length() < this.minHashLength) {
            alphabet = Hashids.consistentShuffle(alphabet, alphabet);
            // 使用编解码字符串填充，可见上述填充的两位前后都是冗余字段
            ret_str = alphabet.substring(halfLen) + ret_str + alphabet.substring(0, halfLen);
            final int excess = ret_str.length() - this.minHashLength;
            // 长度长于 minHashLength 限制
            if (excess > 0) {
                final int start_pos = excess / 2;
                ret_str = ret_str.substring(start_pos, start_pos + this.minHashLength);
            }
        }

        return ret_str;
    }

    private long[] _decode(String hash, String alphabet) {
        final ArrayList<Long> ret = new ArrayList<Long>();

        int i = 0;
        final String regexp = "[" + this.guards + "]";
        String hashBreakdown = hash.replaceAll(regexp, " ");
        String[] hashArray = hashBreakdown.split(" ");

        // 见下面代码说明
        if (hashArray.length == 3 || hashArray.length == 2) {
            i = 1;
        }

        if (hashArray.length > 0) {
            hashBreakdown = hashArray[i];
            if (!hashBreakdown.isEmpty()) {
                // 首位无用字段
                final char lottery = hashBreakdown.charAt(0);

                hashBreakdown = hashBreakdown.substring(1);
                hashBreakdown = hashBreakdown.replaceAll("[" + this.seps + "]", " ");
                hashArray = hashBreakdown.split(" ");

                String subHash, buffer;
                for (final String aHashArray : hashArray) {
                    subHash = aHashArray;
                    buffer = lottery + this.salt + alphabet;
                    alphabet = Hashids.consistentShuffle(alphabet, buffer.substring(0, alphabet.length()));
                    ret.add(Hashids.unhash(subHash, alphabet));
                }
            }
        }

        // transform from List<Long> to long[]
        long[] arr = new long[ret.size()];
        for (int k = 0; k < arr.length; k++) {
            arr[k] = ret.get(k);
        }

        // 加密校验
        if (!this.encode(arr).equals(hash)) {
            arr = new long[0];
        }

        return arr;
    }

    private static String consistentShuffle(String alphabet, String salt) {
        if (salt.length() <= 0) {
            return alphabet;
        }

        int asc_val, j;
        final char[] tmpArr = alphabet.toCharArray();
        for (int i = tmpArr.length - 1, v = 0, p = 0; i > 0; i--, v++) {
            v %= salt.length();
            asc_val = salt.charAt(v);
            p += asc_val;
            j = (asc_val + v + p) % i;
            final char tmp = tmpArr[j];
            tmpArr[j] = tmpArr[i];
            tmpArr[i] = tmp;
        }

        return new String(tmpArr);
    }

    /**
     * 将整数转换为 alphabet.length() 进制
     *
     * @param input
     * @param alphabet
     * @return
     */
    private static String hash(long input, String alphabet) {
        String hash = "";
        final int alphabetLen = alphabet.length();

        do {
            final int index = (int) (input % alphabetLen);
            if (index >= 0 && index < alphabet.length()) {
                hash = alphabet.charAt(index) + hash;
            }
            input /= alphabetLen;
        } while (input > 0);

        return hash;
    }

    /**
     * 将 alphabet.length() 进制转换为整数
     *
     * @param input
     * @param alphabet
     * @return
     */
    private static Long unhash(String input, String alphabet) {
        long number = 0, pos;

        for (int i = 0; i < input.length(); i++) {
            pos = alphabet.indexOf(input.charAt(i));
            number = number * alphabet.length() + pos;
        }

        return number;
    }
}

几种编码状态

    @Test
    public void test5() {
        String str = "456a123b456";
        String s = str.replaceAll("[abc]", " ");
        System.out.println(s);
        System.out.println(Arrays.toString(s.split(" ")));
    }
// 456 123 456
// [456, 123, 456]

https://hashids.org/java/
https://blog.csdn.net/qq_26399665/article/details/79831490