/*
 *
 * nthash.hpp
 * Author: Hamid Mohamadi
 * Genome Sciences Centre,
 * British Columbia Cancer Agency
 */

#ifndef NT_HASH_H
#define NT_HASH_H

#include <stdint.h>
#include <vector>

// offset for the complement base in the random seeds table
const uint8_t cpOff = 0x07;

// shift for gerenerating multiple hash values
const int multiShift = 27;

// seed for gerenerating multiple hash values
static const uint64_t multiSeed = 0x90b45d39fb6da1fa;

// 64-bit random seeds corresponding to bases and their complements
static const uint64_t seedA = 0x3c8bfbb395c60474;
static const uint64_t seedC = 0x3193c18562a02b4c;
static const uint64_t seedG = 0x20323ed082572324;
static const uint64_t seedT = 0x295549f54be24456;
static const uint64_t seedN = 0x0000000000000000;

static const uint64_t seedTab[256] = {
    seedN, seedT, seedN, seedG, seedA, seedA, seedN, seedC, // 0..7
    seedN, seedN, seedN, seedN, seedN, seedN, seedN, seedN, // 8..15
    seedN, seedN, seedN, seedN, seedN, seedN, seedN, seedN, // 16..23
    seedN, seedN, seedN, seedN, seedN, seedN, seedN, seedN, // 24..31
    seedN, seedN, seedN, seedN, seedN, seedN, seedN, seedN, // 32..39
    seedN, seedN, seedN, seedN, seedN, seedN, seedN, seedN, // 40..47
    seedN, seedN, seedN, seedN, seedN, seedN, seedN, seedN, // 48..55
    seedN, seedN, seedN, seedN, seedN, seedN, seedN, seedN, // 56..63
    seedN, seedA, seedN, seedC, seedN, seedN, seedN, seedG, // 64..71
    seedN, seedN, seedN, seedN, seedN, seedN, seedN, seedN, // 72..79
    seedN, seedN, seedN, seedN, seedT, seedT, seedN, seedN, // 80..87
    seedN, seedN, seedN, seedN, seedN, seedN, seedN, seedN, // 88..95
    seedN, seedA, seedN, seedC, seedN, seedN, seedN, seedG, // 96..103
    seedN, seedN, seedN, seedN, seedN, seedN, seedN, seedN, // 104..111
    seedN, seedN, seedN, seedN, seedT, seedT, seedN, seedN, // 112..119
    seedN, seedN, seedN, seedN, seedN, seedN, seedN, seedN, // 120..127
    seedN, seedN, seedN, seedN, seedN, seedN, seedN, seedN, // 128..135
    seedN, seedN, seedN, seedN, seedN, seedN, seedN, seedN, // 136..143
    seedN, seedN, seedN, seedN, seedN, seedN, seedN, seedN, // 144..151
    seedN, seedN, seedN, seedN, seedN, seedN, seedN, seedN, // 152..159
    seedN, seedN, seedN, seedN, seedN, seedN, seedN, seedN, // 160..167
    seedN, seedN, seedN, seedN, seedN, seedN, seedN, seedN, // 168..175
    seedN, seedN, seedN, seedN, seedN, seedN, seedN, seedN, // 176..183
    seedN, seedN, seedN, seedN, seedN, seedN, seedN, seedN, // 184..191
    seedN, seedN, seedN, seedN, seedN, seedN, seedN, seedN, // 192..199
    seedN, seedN, seedN, seedN, seedN, seedN, seedN, seedN, // 200..207
    seedN, seedN, seedN, seedN, seedN, seedN, seedN, seedN, // 208..215
    seedN, seedN, seedN, seedN, seedN, seedN, seedN, seedN, // 216..223
    seedN, seedN, seedN, seedN, seedN, seedN, seedN, seedN, // 224..231
    seedN, seedN, seedN, seedN, seedN, seedN, seedN, seedN, // 232..239
    seedN, seedN, seedN, seedN, seedN, seedN, seedN, seedN, // 240..247
    seedN, seedN, seedN, seedN, seedN, seedN, seedN, seedN  // 248..255
};

static const uint64_t A33r[33] = {
    0x195c60474,0x12b8c08e9,0x571811d3,0xae3023a6,0x15c60474c,0xb8c08e99,0x171811d32,0xe3023a65,0x1c60474ca,0x18c08e995,0x11811d32b,0x3023a657,0x60474cae,0xc08e995c,0x1811d32b8,0x1023a6571,0x474cae3,0x8e995c6,0x11d32b8c,0x23a65718,0x474cae30,0x8e995c60,0x11d32b8c0,0x3a657181,0x74cae302,0xe995c604,0x1d32b8c08,0x1a6571811,0x14cae3023,0x995c6047,0x132b8c08e,0x6571811d,0xcae3023a
};

static const uint64_t A31l[31] = {
    0x3c8bfbb200000000,0x7917f76400000000,0xf22feec800000000,0xe45fdd9200000000,0xc8bfbb2600000000,0x917f764e00000000,0x22feec9e00000000,0x45fdd93c00000000,0x8bfbb27800000000,0x17f764f200000000,0x2feec9e400000000,0x5fdd93c800000000,0xbfbb279000000000,0x7f764f2200000000,0xfeec9e4400000000,0xfdd93c8a00000000,0xfbb2791600000000,0xf764f22e00000000,0xeec9e45e00000000,0xdd93c8be00000000,0xbb27917e00000000,0x764f22fe00000000,0xec9e45fc00000000,0xd93c8bfa00000000,0xb27917f600000000,0x64f22fee00000000,0xc9e45fdc00000000,0x93c8bfba00000000,0x27917f7600000000,0x4f22feec00000000,0x9e45fdd800000000
};

static const uint64_t C33r[33] = {
    0x162a02b4c,0xc5405699,0x18a80ad32,0x115015a65,0x2a02b4cb,0x54056996,0xa80ad32c,0x15015a658,0xa02b4cb1,0x140569962,0x80ad32c5,0x1015a658a,0x2b4cb15,0x569962a,0xad32c54,0x15a658a8,0x2b4cb150,0x569962a0,0xad32c540,0x15a658a80,0xb4cb1501,0x169962a02,0xd32c5405,0x1a658a80a,0x14cb15015,0x9962a02b,0x132c54056,0x658a80ad,0xcb15015a,0x1962a02b4,0x12c540569,0x58a80ad3,0xb15015a6
};

static const uint64_t C31l[31] = {
    0x3193c18400000000,0x6327830800000000,0xc64f061000000000,0x8c9e0c2200000000,0x193c184600000000,0x3278308c00000000,0x64f0611800000000,0xc9e0c23000000000,0x93c1846200000000,0x278308c600000000,0x4f06118c00000000,0x9e0c231800000000,0x3c18463200000000,0x78308c6400000000,0xf06118c800000000,0xe0c2319200000000,0xc184632600000000,0x8308c64e00000000,0x6118c9e00000000,0xc23193c00000000,0x1846327800000000,0x308c64f000000000,0x6118c9e000000000,0xc23193c000000000,0x8463278200000000,0x8c64f0600000000,0x118c9e0c00000000,0x23193c1800000000,0x4632783000000000,0x8c64f06000000000,0x18c9e0c200000000
};


static const uint64_t G33r[33] = {
    0x82572324,0x104ae4648,0x95c8c91,0x12b91922,0x25723244,0x4ae46488,0x95c8c910,0x12b919220,0x57232441,0xae464882,0x15c8c9104,0xb9192209,0x172324412,0xe4648825,0x1c8c9104a,0x191922095,0x12324412b,0x46488257,0x8c9104ae,0x11922095c,0x324412b9,0x64882572,0xc9104ae4,0x1922095c8,0x124412b91,0x48825723,0x9104ae46,0x122095c8c,0x4412b919,0x88257232,0x1104ae464,0x2095c8c9,0x412b9192
};

static const uint64_t G31l[31] = {
    0x20323ed000000000,0x40647da000000000,0x80c8fb4000000000,0x191f68200000000,0x323ed0400000000,0x647da0800000000,0xc8fb41000000000,0x191f682000000000,0x323ed04000000000,0x647da08000000000,0xc8fb410000000000,0x91f6820200000000,0x23ed040600000000,0x47da080c00000000,0x8fb4101800000000,0x1f68203200000000,0x3ed0406400000000,0x7da080c800000000,0xfb41019000000000,0xf682032200000000,0xed04064600000000,0xda080c8e00000000,0xb410191e00000000,0x6820323e00000000,0xd040647c00000000,0xa080c8fa00000000,0x410191f600000000,0x820323ec00000000,0x40647da00000000,0x80c8fb400000000,0x10191f6800000000
};

static const uint64_t T33r[33] = {
    0x14be24456,0x97c488ad,0x12f89115a,0x5f1222b5,0xbe24456a,0x17c488ad4,0xf89115a9,0x1f1222b52,0x1e24456a5,0x1c488ad4b,0x189115a97,0x11222b52f,0x24456a5f,0x488ad4be,0x9115a97c,0x1222b52f8,0x4456a5f1,0x88ad4be2,0x1115a97c4,0x22b52f89,0x456a5f12,0x8ad4be24,0x115a97c48,0x2b52f891,0x56a5f122,0xad4be244,0x15a97c488,0xb52f8911,0x16a5f1222,0xd4be2445,0x1a97c488a,0x152f89115,0xa5f1222b
};

static const uint64_t T31l[31] = {
    0x295549f400000000,0x52aa93e800000000,0xa55527d000000000,0x4aaa4fa200000000,0x95549f4400000000,0x2aa93e8a00000000,0x55527d1400000000,0xaaa4fa2800000000,0x5549f45200000000,0xaa93e8a400000000,0x5527d14a00000000,0xaa4fa29400000000,0x549f452a00000000,0xa93e8a5400000000,0x527d14aa00000000,0xa4fa295400000000,0x49f452aa00000000,0x93e8a55400000000,0x27d14aaa00000000,0x4fa2955400000000,0x9f452aa800000000,0x3e8a555200000000,0x7d14aaa400000000,0xfa29554800000000,0xf452aa9200000000,0xe8a5552600000000,0xd14aaa4e00000000,0xa295549e00000000,0x452aa93e00000000,0x8a55527c00000000,0x14aaa4fa00000000
};

static const uint64_t N33r[33] = {
    seedN,seedN,seedN,seedN,seedN,seedN,seedN,seedN,
    seedN,seedN,seedN,seedN,seedN,seedN,seedN,seedN,
    seedN,seedN,seedN,seedN,seedN,seedN,seedN,seedN,
    seedN,seedN,seedN,seedN,seedN,seedN,seedN,seedN,
    seedN
};

static const uint64_t N31l[31] = {
    seedN,seedN,seedN,seedN,seedN,seedN,seedN,seedN,
    seedN,seedN,seedN,seedN,seedN,seedN,seedN,seedN,
    seedN,seedN,seedN,seedN,seedN,seedN,seedN,seedN,
    seedN,seedN,seedN,seedN,seedN,seedN,seedN
};


static const uint64_t *msTab33r[256] = {
    N33r, T33r, N33r, G33r, A33r, A33r, N33r, C33r, // 0..7
    N33r, N33r, N33r, N33r, N33r, N33r, N33r, N33r, // 8..15
    N33r, N33r, N33r, N33r, N33r, N33r, N33r, N33r, // 16..23
    N33r, N33r, N33r, N33r, N33r, N33r, N33r, N33r, // 24..31
    N33r, N33r, N33r, N33r, N33r, N33r, N33r, N33r, // 32..39
    N33r, N33r, N33r, N33r, N33r, N33r, N33r, N33r, // 40..47
    N33r, N33r, N33r, N33r, N33r, N33r, N33r, N33r, // 48..55
    N33r, N33r, N33r, N33r, N33r, N33r, N33r, N33r, // 56..63
    N33r, A33r, N33r, C33r, N33r, N33r, N33r, G33r, // 64..71
    N33r, N33r, N33r, N33r, N33r, N33r, N33r, N33r, // 72..79
    N33r, N33r, N33r, N33r, T33r, T33r, N33r, N33r, // 80..87
    N33r, N33r, N33r, N33r, N33r, N33r, N33r, N33r, // 88..95
    N33r, A33r, N33r, C33r, N33r, N33r, N33r, G33r, // 96..103
    N33r, N33r, N33r, N33r, N33r, N33r, N33r, N33r, // 104..111
    N33r, N33r, N33r, N33r, T33r, T33r, N33r, N33r, // 112..119
    N33r, N33r, N33r, N33r, N33r, N33r, N33r, N33r, // 120..127
    N33r, N33r, N33r, N33r, N33r, N33r, N33r, N33r, // 128..135
    N33r, N33r, N33r, N33r, N33r, N33r, N33r, N33r, // 136..143
    N33r, N33r, N33r, N33r, N33r, N33r, N33r, N33r, // 144..151
    N33r, N33r, N33r, N33r, N33r, N33r, N33r, N33r, // 152..159
    N33r, N33r, N33r, N33r, N33r, N33r, N33r, N33r, // 160..167
    N33r, N33r, N33r, N33r, N33r, N33r, N33r, N33r, // 168..175
    N33r, N33r, N33r, N33r, N33r, N33r, N33r, N33r, // 176..183
    N33r, N33r, N33r, N33r, N33r, N33r, N33r, N33r, // 184..191
    N33r, N33r, N33r, N33r, N33r, N33r, N33r, N33r, // 192..199
    N33r, N33r, N33r, N33r, N33r, N33r, N33r, N33r, // 200..207
    N33r, N33r, N33r, N33r, N33r, N33r, N33r, N33r, // 208..215
    N33r, N33r, N33r, N33r, N33r, N33r, N33r, N33r, // 216..223
    N33r, N33r, N33r, N33r, N33r, N33r, N33r, N33r, // 224..231
    N33r, N33r, N33r, N33r, N33r, N33r, N33r, N33r, // 232..239
    N33r, N33r, N33r, N33r, N33r, N33r, N33r, N33r, // 240..247
    N33r, N33r, N33r, N33r, N33r, N33r, N33r, N33r  // 248..255
};

static const uint64_t *msTab31l[256] = {
    N31l, T31l, N31l, G31l, A31l, A31l, N31l, C31l, // 0..7
    N31l, N31l, N31l, N31l, N31l, N31l, N31l, N31l, // 8..15
    N31l, N31l, N31l, N31l, N31l, N31l, N31l, N31l, // 16..23
    N31l, N31l, N31l, N31l, N31l, N31l, N31l, N31l, // 24..31
    N31l, N31l, N31l, N31l, N31l, N31l, N31l, N31l, // 32..39
    N31l, N31l, N31l, N31l, N31l, N31l, N31l, N31l, // 40..47
    N31l, N31l, N31l, N31l, N31l, N31l, N31l, N31l, // 48..55
    N31l, N31l, N31l, N31l, N31l, N31l, N31l, N31l, // 56..63
    N31l, A31l, N31l, C31l, N31l, N31l, N31l, G31l, // 64..71
    N31l, N31l, N31l, N31l, N31l, N31l, N31l, N31l, // 72..79
    N31l, N31l, N31l, N31l, T31l, T31l, N31l, N31l, // 80..87
    N31l, N31l, N31l, N31l, N31l, N31l, N31l, N31l, // 88..95
    N31l, A31l, N31l, C31l, N31l, N31l, N31l, G31l, // 96..103
    N31l, N31l, N31l, N31l, N31l, N31l, N31l, N31l, // 104..111
    N31l, N31l, N31l, N31l, T31l, T31l, N31l, N31l, // 112..119
    N31l, N31l, N31l, N31l, N31l, N31l, N31l, N31l, // 120..127
    N31l, N31l, N31l, N31l, N31l, N31l, N31l, N31l, // 128..135
    N31l, N31l, N31l, N31l, N31l, N31l, N31l, N31l, // 136..143
    N31l, N31l, N31l, N31l, N31l, N31l, N31l, N31l, // 144..151
    N31l, N31l, N31l, N31l, N31l, N31l, N31l, N31l, // 152..159
    N31l, N31l, N31l, N31l, N31l, N31l, N31l, N31l, // 160..167
    N31l, N31l, N31l, N31l, N31l, N31l, N31l, N31l, // 168..175
    N31l, N31l, N31l, N31l, N31l, N31l, N31l, N31l, // 176..183
    N31l, N31l, N31l, N31l, N31l, N31l, N31l, N31l, // 184..191
    N31l, N31l, N31l, N31l, N31l, N31l, N31l, N31l, // 192..199
    N31l, N31l, N31l, N31l, N31l, N31l, N31l, N31l, // 200..207
    N31l, N31l, N31l, N31l, N31l, N31l, N31l, N31l, // 208..215
    N31l, N31l, N31l, N31l, N31l, N31l, N31l, N31l, // 216..223
    N31l, N31l, N31l, N31l, N31l, N31l, N31l, N31l, // 224..231
    N31l, N31l, N31l, N31l, N31l, N31l, N31l, N31l, // 232..239
    N31l, N31l, N31l, N31l, N31l, N31l, N31l, N31l, // 240..247
    N31l, N31l, N31l, N31l, N31l, N31l, N31l, N31l  // 248..255
};

// rotate "v" to the left 1 position
inline uint64_t rol1(const uint64_t v) {
    return (v << 1) | (v >> 63);
}

// rotate "v" to the right by 1 position
inline uint64_t ror1(const uint64_t v) {
    return (v >> 1) | (v << 63);
}

// rotate 31-left bits of "v" to the left by "s" positions
inline uint64_t rol31(const uint64_t v, unsigned s) {
    s%=31;
    return ((v << s) | (v >> (31 - s))) & 0x7FFFFFFF;
}

// rotate 33-right bits of "v" to the left by "s" positions
inline uint64_t rol33(const uint64_t v, unsigned s) {
    s%=33;
    return ((v << s) | (v >> (33 - s))) & 0x1FFFFFFFF;
}

// swap bit 0 with bit 33 in "v"
inline uint64_t swapbits033(const uint64_t v) {
    uint64_t x = (v ^ (v >> 33)) & 1;
    return v ^ (x | (x << 33));
}

// swap bit 32 with bit 63 in "v"
inline uint64_t swapbits3263(const uint64_t v) {
    uint64_t x = ((v >> 32) ^ (v >> 63)) & 1;
    return v ^ ((x << 32) | (x << 63));
}

// forward-strand hash value of the base kmer, i.e. fhval(kmer_0)
inline uint64_t NTF64(const char * kmerSeq, const unsigned k) {
    uint64_t hVal=0;
    for(unsigned i=0; i<k; i++) {
        hVal = rol1(hVal);
        hVal = swapbits033(hVal);
        hVal ^= seedTab[(unsigned char)kmerSeq[i]];
    }
    return hVal;
}

// reverse-strand hash value of the base kmer, i.e. rhval(kmer_0)
inline uint64_t NTR64(const char * kmerSeq, const unsigned k) {
    uint64_t hVal=0;
    for(unsigned i=0; i<k; i++) {
        hVal = rol1(hVal);
        hVal = swapbits033(hVal);
        hVal ^= seedTab[(unsigned char)kmerSeq[k-1-i]&cpOff];
    }
    return hVal;
}

// forward-strand ntHash for sliding k-mers
inline uint64_t NTF64(const uint64_t fhVal, const unsigned k, const unsigned char charOut, const unsigned char charIn) {
    uint64_t hVal = rol1(fhVal);
    hVal = swapbits033(hVal);
    hVal ^= seedTab[charIn];
    hVal ^= (msTab31l[charOut][k%31] | msTab33r[charOut][k%33]);
    return hVal;
}

// reverse-complement ntHash for sliding k-mers
inline uint64_t NTR64(const uint64_t rhVal, const unsigned k, const unsigned char charOut, const unsigned char charIn) {
    uint64_t hVal = rhVal ^ (msTab31l[charIn&cpOff][k%31] | msTab33r[charIn&cpOff][k%33]);
    hVal ^= seedTab[charOut&cpOff];
    hVal = ror1(hVal);
    hVal = swapbits3263(hVal);
    return hVal;
}

// canonical ntBase
inline uint64_t NTC64(const char * kmerSeq, const unsigned k) {
    uint64_t fhVal=0, rhVal=0;
    fhVal=NTF64(kmerSeq, k);
    rhVal=NTR64(kmerSeq, k);
    return (rhVal<fhVal)? rhVal : fhVal;
}

// canonical ntHash
inline uint64_t NTC64(const char * kmerSeq, const unsigned k, uint64_t& fhVal, uint64_t& rhVal) {
    fhVal = NTF64(kmerSeq, k);
    rhVal = NTR64(kmerSeq, k);
    return (rhVal<fhVal)? rhVal : fhVal;
}

// canonical ntHash for sliding k-mers
inline uint64_t NTC64(const unsigned char charOut, const unsigned char charIn, const unsigned k, uint64_t& fhVal, uint64_t& rhVal) {
    fhVal = NTF64(fhVal, k, charOut, charIn);
    rhVal = NTR64(rhVal, k, charOut, charIn);
    return (rhVal<fhVal)? rhVal : fhVal;
}

// forward-strand ntHash for sliding k-mers to the left
inline uint64_t NTF64L(const uint64_t rhVal, const unsigned k, const unsigned char charOut, const unsigned char charIn) {
    uint64_t hVal = rhVal ^ (msTab31l[charIn][k%31] | msTab33r[charIn][k%33]);
    hVal ^= seedTab[charOut];
    hVal = ror1(hVal);
    hVal = swapbits3263(hVal);
    return hVal;
}

// reverse-complement ntHash for sliding k-mers to the left
inline uint64_t NTR64L(const uint64_t fhVal, const unsigned k, const unsigned char charOut, const unsigned char charIn) {
    uint64_t hVal = rol1(fhVal);
    hVal = swapbits033(hVal);
    hVal ^= seedTab[charIn&cpOff];
    hVal ^= (msTab31l[charOut&cpOff][k%31] | msTab33r[charOut&cpOff][k%33]);
    return hVal;
}

// canonical ntHash for sliding k-mers to the left
inline uint64_t NTC64L(const unsigned char charOut, const unsigned char charIn, const unsigned k, uint64_t& fhVal, uint64_t& rhVal) {
    fhVal = NTF64L(fhVal, k, charOut, charIn);
    rhVal = NTR64L(rhVal, k, charOut, charIn);
    return (rhVal<fhVal)? rhVal : fhVal;
}

// ntBase with seeding option
inline uint64_t NTF64(const char * kmerSeq, const unsigned k, const unsigned seed) {
    uint64_t hVal=NTF64(kmerSeq, k);
    if(seed==0) return hVal;
    hVal *= seed ^ k * multiSeed;
    hVal ^= hVal >> multiShift;
    return hVal;
}

// canonical ntBase with seeding option
inline uint64_t NTC64(const char * kmerSeq, const unsigned k, const unsigned seed) {
    uint64_t hVal = NTC64(kmerSeq,k);
    if(seed==0) return hVal;
    hVal *= seed ^ k * multiSeed;
    hVal ^= hVal >> multiShift;
    return hVal;
}

// multihash ntHash, ntBase
inline void NTM64(const char * kmerSeq, const unsigned k, const unsigned m, uint64_t *hVal) {
    uint64_t bVal=0, tVal=0;
    bVal = NTF64(kmerSeq, k);
    hVal[0] = bVal;
    for(unsigned i=1; i<m; i++) {
        tVal = bVal * (i ^ k * multiSeed);
        tVal ^= tVal >> multiShift;
        hVal[i] =  tVal;
    }
}

// one extra hash for given base hash
inline uint64_t NTE64(const uint64_t hVal, const unsigned k, const unsigned i) {
    uint64_t tVal = hVal;
    tVal *= (i ^ k * multiSeed);
    tVal ^= tVal >> multiShift;
    return tVal;
}

// multihash ntHash for sliding k-mers
inline void NTM64(const unsigned char charOut, const unsigned char charIn, const unsigned k, const unsigned m, uint64_t *hVal) {
    uint64_t bVal=0, tVal=0;
    bVal = NTF64(hVal[0], k, charOut, charIn);
    hVal[0] = bVal;
    for(unsigned i=1; i<m; i++) {
        tVal = bVal * (i ^ k * multiSeed);
        tVal ^= tVal >> multiShift;
        hVal[i] =  tVal;
    }
}

// canonical multihash ntBase
inline void NTMC64(const char * kmerSeq, const unsigned k, const unsigned m, uint64_t *hVal) {
    uint64_t bVal=0, tVal=0;
    bVal = NTC64(kmerSeq, k);
    hVal[0] = bVal;
    for(unsigned i=1; i<m; i++) {
        tVal = bVal * (i ^ k * multiSeed);
        tVal ^= tVal >> multiShift;
        hVal[i] =  tVal;
    }
}

// canonical multihash ntHash
inline void NTMC64(const char * kmerSeq, const unsigned k, const unsigned m, uint64_t& fhVal, uint64_t& rhVal, uint64_t *hVal) {
    uint64_t bVal=0, tVal=0;
    bVal = NTC64(kmerSeq, k, fhVal, rhVal);
    hVal[0] = bVal;
    for(unsigned i=1; i<m; i++) {
        tVal = bVal * (i ^ k * multiSeed);
        tVal ^= tVal >> multiShift;
        hVal[i] =  tVal;
    }
}

// canonical multihash ntHash for sliding k-mers
inline void NTMC64(const unsigned char charOut, const unsigned char charIn, const unsigned k, const unsigned m, uint64_t& fhVal, uint64_t& rhVal, uint64_t *hVal) {
    uint64_t bVal=0, tVal=0;
    bVal = NTC64(charOut, charIn, k, fhVal, rhVal);
    hVal[0] = bVal;
    for(unsigned i=1; i<m; i++) {
        tVal = bVal * (i ^ k * multiSeed);
        tVal ^= tVal >> multiShift;
        hVal[i] =  tVal;
    }
}

/*
 * ignoring k-mers containing nonACGT using ntHash function
*/

// canonical ntBase
inline bool NTC64(const char *kmerSeq, const unsigned k, uint64_t& hVal, unsigned& locN) {
    hVal=0;
    locN=0;
    uint64_t fhVal=0,rhVal=0;
    for(int i=k-1; i>=0; i--) {
        if(seedTab[(unsigned char)kmerSeq[i]]==seedN) {
            locN=i;
            return false;
        }
        fhVal = rol1(fhVal);
        fhVal = swapbits033(fhVal);
        fhVal ^= seedTab[(unsigned char)kmerSeq[k-1-i]];

        rhVal = rol1(rhVal);
        rhVal = swapbits033(rhVal);
        rhVal ^= seedTab[(unsigned char)kmerSeq[i]&cpOff];
    }
    hVal = (rhVal<fhVal)? rhVal : fhVal;
    return true;
}

// canonical multihash ntBase
inline bool NTMC64(const char *kmerSeq, const unsigned k, const unsigned m, unsigned& locN, uint64_t* hVal) {
    uint64_t bVal=0, tVal=0, fhVal=0, rhVal=0;
    locN=0;
    for(int i=k-1; i>=0; i--) {
        if(seedTab[(unsigned char)kmerSeq[i]]==seedN) {
            locN=i;
            return false;
        }
        fhVal = rol1(fhVal);
        fhVal = swapbits033(fhVal);
        fhVal ^= seedTab[(unsigned char)kmerSeq[k-1-i]];

        rhVal = rol1(rhVal);
        rhVal = swapbits033(rhVal);
        rhVal ^= seedTab[(unsigned char)kmerSeq[i]&cpOff];
    }
    bVal = (rhVal<fhVal)? rhVal : fhVal;
    hVal[0] = bVal;
    for(unsigned i=1; i<m; i++) {
        tVal = bVal * (i ^ k * multiSeed);
        tVal ^= tVal >> multiShift;
        hVal[i] =  tVal;
    }
    return true;
}

// canonical ntHash
inline bool NTC64(const char *kmerSeq, const unsigned k, uint64_t& fhVal, uint64_t& rhVal, uint64_t& hVal, unsigned& locN) {
    hVal=fhVal=rhVal=0;
    locN=0;
    for(int i=k-1; i>=0; i--) {
        if(seedTab[(unsigned char)kmerSeq[i]]==seedN) {
            locN=i;
            return false;
        }
        fhVal = rol1(fhVal);
        fhVal = swapbits033(fhVal);
        fhVal ^= seedTab[(unsigned char)kmerSeq[k-1-i]];

        rhVal = rol1(rhVal);
        rhVal = swapbits033(rhVal);
        rhVal ^= seedTab[(unsigned char)kmerSeq[i]&cpOff];
    }
    hVal = (rhVal<fhVal)? rhVal : fhVal;
    return true;
}

// canonical multihash ntHash
inline bool NTMC64(const char *kmerSeq, const unsigned k, const unsigned m, uint64_t& fhVal, uint64_t& rhVal, unsigned& locN, uint64_t* hVal) {
    fhVal=rhVal=0;
    uint64_t bVal=0, tVal=0;
    locN=0;
    for(int i=k-1; i>=0; i--) {
        if(seedTab[(unsigned char)kmerSeq[i]]==seedN) {
            locN=i;
            return false;
        }
        fhVal = rol1(fhVal);
        fhVal = swapbits033(fhVal);
        fhVal ^= seedTab[(unsigned char)kmerSeq[k-1-i]];

        rhVal = rol1(rhVal);
        rhVal = swapbits033(rhVal);
        rhVal ^= seedTab[(unsigned char)kmerSeq[i]&cpOff];
    }
    bVal = (rhVal<fhVal)? rhVal : fhVal;
    hVal[0] = bVal;
    for(unsigned i=1; i<m; i++) {
        tVal = bVal * (i ^ k * multiSeed);
        tVal ^= tVal >> multiShift;
        hVal[i] =  tVal;
    }
    return true;
}

// strand-aware canonical multihash ntHash
inline bool NTMC64(const char *kmerSeq, const unsigned k, const unsigned m, uint64_t& fhVal, uint64_t& rhVal, unsigned& locN, uint64_t* hVal, bool& hStn) {
    fhVal=rhVal=0;
    uint64_t bVal=0, tVal=0;
    locN=0;
    for(int i=k-1; i>=0; i--) {
        if(seedTab[(unsigned char)kmerSeq[i]]==seedN) {
            locN=i;
            return false;
        }
        fhVal = rol1(fhVal);
        fhVal = swapbits033(fhVal);
        fhVal ^= seedTab[(unsigned char)kmerSeq[k-1-i]];
        
        rhVal = rol1(rhVal);
        rhVal = swapbits033(rhVal);
        rhVal ^= seedTab[(unsigned char)kmerSeq[i]&cpOff];
    }
    hStn = rhVal<fhVal;
    bVal = hStn? rhVal : fhVal;
    hVal[0] = bVal;
    for(unsigned i=1; i<m; i++) {
        tVal = bVal * (i ^ k * multiSeed);
        tVal ^= tVal >> multiShift;
        hVal[i] =  tVal;
    }
    return true;
}

// starnd-aware canonical multihash ntHash for sliding k-mers
inline void NTMC64(const unsigned char charOut, const unsigned char charIn, const unsigned k, const unsigned m, uint64_t& fhVal, uint64_t& rhVal, uint64_t *hVal, bool &hStn) {
    uint64_t bVal=0, tVal=0;
    bVal = NTC64(charOut, charIn, k, fhVal, rhVal);
    hStn = rhVal<fhVal;
    hVal[0] = bVal;
    for(unsigned i=1; i<m; i++) {
        tVal = bVal * (i ^ k * multiSeed);
        tVal ^= tVal >> multiShift;
        hVal[i] =  tVal;
    }
}

// masking canonical ntHash using spaced seed pattern
inline uint64_t maskHash(uint64_t &fkVal, uint64_t &rkVal, const char * seedSeq, const char * kmerSeq, const unsigned k) {
    uint64_t fsVal=fkVal, rsVal=rkVal;
    for(unsigned i=0; i<k; i++) {
        if(seedSeq[i]!='1') {
            fsVal ^= (msTab31l[(unsigned char)kmerSeq[i]][(k-1-i)%31] | msTab33r[(unsigned char)kmerSeq[i]][(k-1-i)%33]);
            rsVal ^= (msTab31l[(unsigned char)kmerSeq[i]&cpOff][i%31] | msTab33r[(unsigned char)kmerSeq[i]&cpOff][i%33]);
            
        }
    }
    return (rsVal<fsVal)? rsVal : fsVal;
}

// spaced seed ntHash for base kmer, i.e. fhval(kmer_0)
inline uint64_t NTS64(const char * kmerSeq, const std::vector<bool> &seed, const unsigned k, uint64_t &hVal) {
    hVal=0;
    uint64_t sVal = 0;
    for(unsigned i=0; i<k; i++) {
        hVal = rol1(hVal);
        hVal = swapbits033(hVal);
        sVal = hVal;
        hVal ^= seedTab[(unsigned char)kmerSeq[i]];
        if(seed[i]) sVal = hVal;
    }
    return sVal;
}

// spaced seed ntHash for sliding k-mers
inline uint64_t NTS64(const char * kmerSeq, const std::vector<bool> &seed, const unsigned char charOut, const unsigned char charIn, const unsigned k, uint64_t &hVal) {
    hVal = NTF64(hVal,k,charOut,charIn);
    uint64_t sVal = hVal;
    for(unsigned i=0; i<k; i++)
        if(!seed[i]) {
            sVal ^= (msTab31l[(unsigned char)kmerSeq[i]][k%31] | msTab33r[(unsigned char)kmerSeq[i]][k%33]);
        }
    return sVal;
}

// strand-aware multihash spaced seed ntHash
inline bool NTMS64(const char *kmerSeq, const std::vector<std::vector<unsigned> > &seedSeq, const unsigned k, const unsigned m, uint64_t& fhVal, uint64_t& rhVal, unsigned& locN, uint64_t* hVal, bool *hStn) {
    fhVal=rhVal=0;
    locN=0;
    for(int i=k-1; i>=0; i--) {
        if(seedTab[(unsigned char)kmerSeq[i]]==seedN) {
            locN=i;
            return false;
        }
        fhVal = rol1(fhVal);
        fhVal = swapbits033(fhVal);
        fhVal ^= seedTab[(unsigned char)kmerSeq[k-1-i]];
        
        rhVal = rol1(rhVal);
        rhVal = swapbits033(rhVal);
        rhVal ^= seedTab[(unsigned char)kmerSeq[i]&cpOff];
    }
    
    for(unsigned j=0; j<m; j++) {
        uint64_t fsVal=fhVal, rsVal=rhVal;
        for(std::vector<unsigned>::const_iterator i=seedSeq[j].begin(); i!=seedSeq[j].end(); ++i) {
            fsVal ^= (msTab31l[(unsigned char)kmerSeq[*i]][(k-1-*i)%31] | msTab33r[(unsigned char)kmerSeq[*i]][(k-1-*i)%33]);
            rsVal ^= (msTab31l[(unsigned char)kmerSeq[*i]&cpOff][*i%31] | msTab33r[(unsigned char)kmerSeq[*i]&cpOff][*i%33]);
        }
        hStn[j] = rsVal<fsVal;
        hVal[j] = hStn[j]? rsVal : fsVal;
    }
    return true;
}

// strand-aware multihash spaced seed ntHash for sliding k-mers
inline void NTMS64(const char *kmerSeq, const std::vector<std::vector<unsigned> > &seedSeq, const unsigned char charOut, const unsigned char charIn, const unsigned k, const unsigned m, uint64_t& fhVal, uint64_t& rhVal, uint64_t *hVal, bool *hStn) {
    fhVal = NTF64(fhVal,k,charOut,charIn);
    rhVal = NTR64(rhVal,k,charOut,charIn);
    for(unsigned j=0; j<m; j++) {
        uint64_t fsVal=fhVal, rsVal=rhVal;
        for(std::vector<unsigned>::const_iterator i=seedSeq[j].begin(); i!=seedSeq[j].end(); ++i) {
            fsVal ^= (msTab31l[(unsigned char)kmerSeq[*i]][(k-1-*i)%31] | msTab33r[(unsigned char)kmerSeq[*i]][(k-1-*i)%33]);
            rsVal ^= (msTab31l[(unsigned char)kmerSeq[*i]&cpOff][*i%31] | msTab33r[(unsigned char)kmerSeq[*i]&cpOff][*i%33]);;
        }
        hStn[j] = rsVal<fsVal;
        hVal[j] = hStn[j]? rsVal : fsVal;
    }
}

// strand-aware multihash spaced seed ntHash with multiple hashes per seed
inline bool NTMSM64(const char *kmerSeq, const std::vector<std::vector<unsigned> > &seedSeq, const unsigned k, const unsigned m, const unsigned m2,
    uint64_t& fhVal, uint64_t& rhVal, unsigned& locN, uint64_t* hVal, bool *hStn)
{
    fhVal=rhVal=0;
    locN=0;
    for(int i=k-1; i>=0; i--) {
        if(seedTab[(unsigned char)kmerSeq[i]]==seedN) {
            locN=i;
            return false;
        }
        fhVal = rol1(fhVal);
        fhVal = swapbits033(fhVal);
        fhVal ^= seedTab[(unsigned char)kmerSeq[k-1-i]];

        rhVal = rol1(rhVal);
        rhVal = swapbits033(rhVal);
        rhVal ^= seedTab[(unsigned char)kmerSeq[i]&cpOff];
    }

    for(unsigned j=0; j<m; j++) {
        uint64_t fsVal=fhVal, rsVal=rhVal;
        for(std::vector<unsigned>::const_iterator i=seedSeq[j].begin(); i!=seedSeq[j].end(); ++i) {
            fsVal ^= (msTab31l[(unsigned char)kmerSeq[*i]][(k-1-*i)%31] | msTab33r[(unsigned char)kmerSeq[*i]][(k-1-*i)%33]);
            rsVal ^= (msTab31l[(unsigned char)kmerSeq[*i]&cpOff][*i%31] | msTab33r[(unsigned char)kmerSeq[*i]&cpOff][*i%33]);
        }
        hStn[j * m2] = rsVal<fsVal;
        hVal[j * m2] = hStn[j * m2]? rsVal : fsVal;
        for(unsigned j2=1; j2<m2; j2++) {
            uint64_t tVal = hVal[j * m2] * (j2 ^ k * multiSeed);
            tVal ^= tVal >> multiShift;
            hStn[j * m2 + j2] = hStn[j * m2];
            hVal[j * m2 + j2] = tVal;
        }
    }
    return true;
}

// strand-aware multihash spaced seed ntHash for sliding k-mers with multiple hashes per seed
inline void NTMSM64(const char *kmerSeq, const std::vector<std::vector<unsigned> > &seedSeq, const unsigned char charOut, const unsigned char charIn,
const unsigned k, const unsigned m, const unsigned m2, uint64_t& fhVal, uint64_t& rhVal, uint64_t *hVal, bool *hStn)
{
    fhVal = NTF64(fhVal,k,charOut,charIn);
    rhVal = NTR64(rhVal,k,charOut,charIn);
    for(unsigned j=0; j<m; j++) {
        uint64_t fsVal=fhVal, rsVal=rhVal;
        for(std::vector<unsigned>::const_iterator i=seedSeq[j].begin(); i!=seedSeq[j].end(); ++i) {
            fsVal ^= (msTab31l[(unsigned char)kmerSeq[*i]][(k-1-*i)%31] | msTab33r[(unsigned char)kmerSeq[*i]][(k-1-*i)%33]);
            rsVal ^= (msTab31l[(unsigned char)kmerSeq[*i]&cpOff][*i%31] | msTab33r[(unsigned char)kmerSeq[*i]&cpOff][*i%33]);
        }
        hStn[j * m2] = rsVal<fsVal;
        hVal[j * m2] = hStn[j * m2]? rsVal : fsVal;
        for(unsigned j2=1; j2<m2; j2++) {
            uint64_t tVal = hVal[j * m2] * (j2 ^ k * multiSeed);
            tVal ^= tVal >> multiShift;
            hStn[j * m2 + j2] = hStn[j * m2];
            hVal[j * m2 + j2] = tVal;
        }
    }
}


#endif