bwa_perf/bwt.cpp

#include <stdint.h>
#include <stdlib.h>
#include <stdio.h>
#include <string.h>
#include <string>
#include <iostream>

#include "util.h"
#include "bwt.h"

using namespace std;

const static char BASE[4] = {'A', 'C', 'G', 'T'};

// 打印32位整型数据中包含的pre-bwt：bwt
static void print_base_uint32(uint32_t p)
{
    const char BASE[4] = {'A', 'C', 'G', 'T'};
    for (int i = 30; i > 0; i -= 2)
    {
        int b1 = p >> i & 3;
        cout << BASE[b1] << endl;
    }
}

void dump_bwt(const char *fn, const bwt_t *bwt)
{
    FILE *fp;
    fp = xopen(fn, "wb");
    err_fwrite(&bwt->primary, sizeof(bwtint_t), 1, fp);
    err_fwrite(bwt->L2 + 1, sizeof(bwtint_t), 4, fp);
    err_fwrite(bwt->bwt, 4, bwt->bwt_size, fp);
    err_fflush(fp);
    err_fclose(fp);
}

// 计算一个字节构成的T,G,C,A序列，对应的每个碱基的个数(按T,G,C,A顺序存储在32位整数中，每个占8位)
void bwt_gen_cnt_table(bwt_t *bwt)
{
    int i, j;
    for (i = 0; i != 256; ++i)
    {
        uint32_t x = 0;
        for (j = 0; j != 4; ++j)
            x |= (((i & 3) == j) + ((i >> 2 & 3) == j) + ((i >> 4 & 3) == j) + (i >> 6 == j)) << (j << 3);
        bwt->cnt_table[i] = x;
    }
}

// 解析两bit的bwt碱基序列，这个只有bwt str，可以包含也可不包含occ check point
bwt_t *restore_bwt(const char *fn)
{
    bwt_t *bwt;
    bwt = (bwt_t *)calloc(1, sizeof(bwt_t));
    FILE *fp = fopen(fn, "rb");

    fseek(fp, 0, SEEK_END);
    bwt->bwt_size = (ftell(fp) - sizeof(bwtint_t) * 5) >> 2; // 以32位word为单位计算的size
    bwt->bwt = (uint32_t *)calloc(bwt->bwt_size, 4);
    fseek(fp, 0, SEEK_SET);
    fread(&bwt->primary, sizeof(bwtint_t), 1, fp);
    fread(bwt->L2 + 1, sizeof(bwtint_t), 4, fp);
    fread_fix(fp, bwt->bwt_size << 2, bwt->bwt);
    bwt->seq_len = bwt->L2[4];

    // char *buf = (char *)calloc(bwt->seq_len + 1, 1);
    // for (bwtint_t i = 0; i < bwt->seq_len; ++i)
    // {
    // 	buf[i] = BASE[bwt->bwt[i >> 4] >> ((15 - (i & 15)) << 1) & 3];
    // 	cout << buf[i];
    // }
    // cout << endl;

    fclose(fp);
    bwt_gen_cnt_table(bwt); // 字节所能表示的各种碱基组合中，各个碱基的累积数量
    return bwt;
}

// 根据原始的字符串bwt创建interval-bwt，用uint32_t来表示occ
void create_interval_occ_bwt(bwt_t *bwt)
{
    bwtint_t i, k, n_occ;
    uint32_t *buf;
    uint32_t c[4];

    n_occ = (bwt->seq_len + OCC_INTERVAL - 1) / OCC_INTERVAL + 1;
    bwt->bwt_size += n_occ * 4;  // the new size
    buf = (uint32_t *)calloc(bwt->bwt_size, 4); // will be the new bwt
    c[0] = c[1] = c[2] = c[3] = 0;
    // 计算occ，生成naive bwt
    for (i = k = 0; i < bwt->seq_len; ++i)
    {
        if (i % 1000000 == 0)
        {
            cout << "line: " << i << '\t' << ((uint64_t)c[0] + c[1] + c[2] + c[3]) << '\t';
            cout << c[0] << ' ' << c[1] << ' ' << c[2] << ' ' << c[3] << endl;
        }
        if (i % OCC_INTERVAL == 0)
        {
            memcpy(buf + k, c, sizeof(uint32_t) * 4);
            k += 4; // in fact: sizeof(bwtint_t)=4*(sizeof(bwtint_t)/4) 每个c包含多少个32位
        }
        if (i % 16 == 0)
            buf[k++] = bwt->bwt[i / 16]; // 16 == sizeof(uint32_t)/2, 2个bit表示一个碱基
        ++c[bwt_B00(bwt, i)];
    }
    // the last element
    // cout << c[0] << '\t' << c[1] << '\t' << c[2] << '\t' << c[3] << endl;
    memcpy(buf + k, c, sizeof(uint32_t) * 4);
    xassert(k + 4 == bwt->bwt_size, "inconsistent bwt_size");
    // update bwt
    free(bwt->bwt);
    bwt->bwt = buf;
}

// 对64位整型数据y，计算碱基c的累积个数
static inline int __occ_aux(uint64_t y, int c)
{
    // reduce nucleotide counting to bits counting
    y = ((c & 2) ? y : ~y) >> 1 & ((c & 1) ? y : ~y) & 0x5555555555555555ull;
    // count the number of 1s in y
    y = (y & 0x3333333333333333ull) + (y >> 2 & 0x3333333333333333ull);
    return ((y + (y >> 4)) & 0xf0f0f0f0f0f0f0full) * 0x101010101010101ull >> 56;
}

// k行(包含)之前，碱基c的累计总数, interval大于等于32才能正确计算
bwtint_t bwt_occ(const bwt_t *bwt, bwtint_t k, uint8_t c)
{
    bwtint_t n;
    uint32_t *p, *end;

    if (k == bwt->seq_len)
        return bwt->L2[c + 1] - bwt->L2[c];
    if (k == (bwtint_t)(-1))
        return 0;
    k -= (k >= bwt->primary); // because $ is not in bwt

    // retrieve Occ at k/OCC_INTERVAL
    n = ((bwtint_t *)(p = bwt_occ_intv(bwt, k)))[c];
    // cout << "bwt_occ - 1: " << (int)c << '\t' << k << '\t' << n << endl;
    p += sizeof(bwtint_t); // jump to the start of the first BWT cell

    // calculate Occ up to the last k/32
    end = p + (((k >> 5) - ((k & ~OCC_INTV_MASK) >> 5)) << 1);
    for (; p < end; p += 2)
        n += __occ_aux((uint64_t)p[0] << 32 | p[1], c);

    // cout << "bwt_occ - 2: " << (int)c << '\t' << k << '\t' << n << endl;
    //  calculate Occ
    n += __occ_aux(((uint64_t)p[0] << 32 | p[1]) & ~((1ull << ((~k & 31) << 1)) - 1), c);
    if (c == 0)
        n -= ~k & 31; // corrected for the masked bits
    // cout << "bwt_occ - 3: " << (int)c << '\t' << k << '\t' << n << endl;
    return n;
}

// 统计k行（bwt mtx行）之前4种碱基累积数量，这里的k是bwt矩阵里的行，比bwt字符串多1
void bwt_occ4(const bwt_t *bwt, bwtint_t k, bwtint_t cnt[4])
{
    bwtint_t x = 0;
    uint32_t *p, tmp, *end;
    // bwtint_t bwt_k_base_line = k >> OCC_INTV_SHIFT << OCC_INTV_SHIFT;
    if (k == (bwtint_t)(-1))
    {
        memset(cnt, 0, 4 * sizeof(bwtint_t));
        return;
    }
    k -= (k >= bwt->primary); // because $ is not in bwt
    p = bwt_occ_intv(bwt, k);
    //cout << "k: " << k << "; occ: " << p[0] << ' ' << p[1] << ' ' << p[2] << ' ' << p[3] << endl;
    cnt[0] = p[0];
    cnt[1] = p[1];
    cnt[2] = p[2];
    cnt[3] = p[3];
    p += 4;                              // sizeof(bwtint_t) = 4*(sizeof(bwtint_t)/sizeof(uint32_t))
    end = p + ((k >> 4) - ((k & ~OCC_INTV_MASK) >> 4)); // this is the end point of the following loop
    //end = p + 1;
    // cout << "k - kbase: " << k - bwt_k_base_line << endl;
    for (; p < end; ++p)
    {
        x += __occ_aux4(bwt, *p);
        //if (k == 3965453116 || k == 3965453672 || k == 2668688087 || k == 2668688550)
        //    print_base_uint32(*p);
    }
    tmp = *p & ~((1U << ((~k & 15) << 1)) - 1);
    //if (k == 3965453116 || k == 3965453672 || k == 2668688087 || k == 2668688550)
    //    print_base_uint32(tmp);
    x += __occ_aux4(bwt, tmp) - (~k & 15); // 这里多算了A，要减去
    cnt[0] += x & 0xff;
    cnt[1] += x >> 8 & 0xff;
    cnt[2] += x >> 16 & 0xff;
    cnt[3] += x >> 24;
    //if (k == 3965453116 || k == 3965453672 || k == 2668688087 || k == 2668688550)
    //    cout << "final-occ: " << cnt[0] << ' ' << cnt[1] << ' ' << cnt[2] << ' ' << cnt[3] << endl;
}

// an analogy to bwt_occ4() but more efficient, requiring k <= l
void bwt_2occ4(const bwt_t *bwt, bwtint_t k, bwtint_t l, bwtint_t cntk[4], bwtint_t cntl[4])
{
    // double tm_t = realtime();
    // bwt_occ4(bwt, k, cntk);
    // bwt_occ4(bwt, l, cntl);
    // return;
    bwtint_t _k, _l;
    _k = k - (k >= bwt->primary);
    _l = l - (l >= bwt->primary);
    if (_l >> OCC_INTV_SHIFT != _k >> OCC_INTV_SHIFT || k == (bwtint_t)(-1) || l == (bwtint_t)(-1))
    {
        bwt_occ4(bwt, k, cntk);
        bwt_occ4(bwt, l, cntl);
    }
    else
    {
        bwtint_t x, y;
        uint32_t *p, tmp, *endk, *endl;
        k -= (k >= bwt->primary); // because $ is not in bwt
        l -= (l >= bwt->primary);
        p = bwt_occ_intv(bwt, k);
        cntk[0] = p[0];
        cntk[1] = p[1];
        cntk[2] = p[2];
        cntk[3] = p[3];
        p += 4; // sizeof(bwtint_t) = 4*(sizeof(bwtint_t)/sizeof(uint32_t))
        // prepare cntk[]
        endk = p + ((k >> 4) - ((k & ~OCC_INTV_MASK) >> 4));
        endl = p + ((l >> 4) - ((l & ~OCC_INTV_MASK) >> 4));
        for (x = 0; p < endk; ++p)
            x += __occ_aux4(bwt, *p);
        y = x;
        tmp = *p & ~((1U << ((~k & 15) << 1)) - 1);
        x += __occ_aux4(bwt, tmp) - (~k & 15);
        // calculate cntl[] and finalize cntk[]
        for (; p < endl; ++p)
            y += __occ_aux4(bwt, *p);
        tmp = *p & ~((1U << ((~l & 15) << 1)) - 1);
        y += __occ_aux4(bwt, tmp) - (~l & 15);
        memcpy(cntl, cntk, 4 * sizeof(bwtint_t));
        cntk[0] += x & 0xff;
        cntk[1] += x >> 8 & 0xff;
        cntk[2] += x >> 16 & 0xff;
        cntk[3] += x >> 24;
        cntl[0] += y & 0xff;
        cntl[1] += y >> 8 & 0xff;
        cntl[2] += y >> 16 & 0xff;
        cntl[3] += y >> 24;
    }
    // t5 += realtime() - tm_t;
    // f1 += 1;
}

void bwt_extend(const bwt_t *bwt, const bwtintv_t *ik, bwtintv_t ok[4], int is_back)
{
    bwtint_t tk[4], tl[4];
    int i;
    bwt_2occ4(bwt, ik->x[!is_back] - 1, ik->x[!is_back] - 1 + ik->x[2], tk, tl); // tk表示在k行之前所有各个碱基累积出现次数，tl表示在l行之前的累积
    // cout << "bwt-1-d: " << ik->x[!is_back] - 1 << '\t' << tk[0] << '\t' << tk[1] << '\t' << tk[2] << '\t' << tk[3] << endl;
    // cout << "bwt-1-d: " << ik->x[!is_back] - 1 + ik->x[2] << '\t' << tl[0] << '\t' << tl[1] << '\t' << tl[2] << '\t' << tl[3] << endl;
    // 这里是反向扩展
    for (i = 0; i != 4; ++i)
    {
        ok[i].x[!is_back] = bwt->L2[i] + 1 + tk[i]; // 起始行位置，互补链
        ok[i].x[2] = tl[i] - tk[i];                 // 间隔
    }
    // 因为计算的是互补碱基，所以3对应着0,2对应1，下边是正向扩展
    ok[3].x[is_back] = ik->x[is_back] + (ik->x[!is_back] <= bwt->primary && ik->x[!is_back] + ik->x[2] - 1 >= bwt->primary);
    ok[2].x[is_back] = ok[3].x[is_back] + ok[3].x[2];
    ok[1].x[is_back] = ok[2].x[is_back] + ok[2].x[2];
    ok[0].x[is_back] = ok[1].x[is_back] + ok[1].x[2];
}

// 利用bwt搜索seed，完整搜索，只需要单向搜索
void bwt_search(bwt_t *bwt, const string &q)
{
    bwtintv_t ik, ok[4];
    int i, c, x = 0;

    bwt_set_intv(bwt, bval(q[x]), ik);
    ik.info = x + 1;
    // cout << ik.x[0] << '\t' << ik.x[1] << '\t' << ik.x[2] << endl;

    for (i = x + 1; i < (int)q.size(); ++i)
    {
        if (bval(q[i]) < 4)
        {
            c = cbval(q[i]);
            bwt_extend(bwt, &ik, ok, 0);
            ik = ok[c];
            ik.info = i + 1;
            // cout << "bwt-1: " << ik.x[0] << '\t' << ik.x[1] << '\t' << ik.x[2] << endl;
        }
    }
}

// 扩展两次
void bwt_extend2(const bwt_t *bwt, bwtintv_t *ik, bwtintv_t ok[4], int is_back, int c1, int c2)
{
    bwtint_t tk[4], tl[4];
    // oc1, oc2;
    // int int1 = 0, int2 = 0;
    int i;
    bwt_2occ4(bwt, ik->x[!is_back] - 1, ik->x[!is_back] - 1 + ik->x[2], tk, tl); // tk表示在k行之前所有各个碱基累积出现次数，tl表示在l行之前的累积

    // 这里是反向扩展
    for (i = 0; i != 4; ++i)
    {
        ok[i].x[!is_back] = bwt->L2[i] + 1 + tk[i]; // 起始行位置，互补链
        ok[i].x[2] = tl[i] - tk[i];                 // 间隔
    }

    // 因为计算的是互补碱基，所以3对应着0,2对应1，下边是正向扩展
    ok[3].x[is_back] = ik->x[is_back] + (ik->x[!is_back] <= bwt->primary && ik->x[!is_back] + ik->x[2] - 1 >= bwt->primary);
    ok[2].x[is_back] = ok[3].x[is_back] + ok[3].x[2];
    ok[1].x[is_back] = ok[2].x[is_back] + ok[2].x[2];
    ok[0].x[is_back] = ok[1].x[is_back] + ok[1].x[2];
    //cout << (ik->x[!is_back] <= bwt->primary && ik->x[!is_back] + ik->x[2] - 1 >= bwt->primary) << endl;
    *ik = ok[c1];

    //oc1 = oc2 = 0;
    //for (i = 3; i > c1; --i)
    //{
    //    int1 += tl[i] - tk[i];
    //    oc1 += tk[i];
    //    oc2 += tl[i];
    //}
    //cout << "bwt-2 interval-1: " << int1 << ' ' << oc1 << ' ' << oc2 << endl;

    bwt_2occ4(bwt, ik->x[!is_back] - 1, ik->x[!is_back] - 1 + ik->x[2], tk, tl);
    for (i = 0; i != 4; ++i)
    {
        ok[i].x[!is_back] = bwt->L2[i] + 1 + tk[i]; // 起始行位置，互补链
        ok[i].x[2] = tl[i] - tk[i];                 // 间隔
    }
    // 因为计算的是互补碱基，所以3对应着0,2对应1，下边是正向扩展
    ok[3].x[is_back] = ik->x[is_back] + (ik->x[!is_back] <= bwt->primary && ik->x[!is_back] + ik->x[2] - 1 >= bwt->primary);
    ok[2].x[is_back] = ok[3].x[is_back] + ok[3].x[2];
    ok[1].x[is_back] = ok[2].x[is_back] + ok[2].x[2];
    ok[0].x[is_back] = ok[1].x[is_back] + ok[1].x[2];

    // oc1 = oc2 = 0;
    // for (i = 3; i > c2; --i)
    // {
    //     int2 += tl[i] - tk[i];
    //     oc1 += tk[i];
    //     oc2 += tl[i];
    // }
    // cout << "bwt-2 interval-2: " << int2 << ' ' << oc1 << ' ' << oc2 << endl;
}

// 每次扩展两步
bwtintv_t bwt_search2(bwt_t *bwt, const string &q)
{
    bwtintv_t ik, ok[4];
    int i, c1, c2, x = 0;

    bwt_set_intv(bwt, bval(q[x]), ik);
    ik.info = x + 1;
    //cout << "bwt-2: " << ik.x[0] << '\t' << ik.x[1] << '\t' << ik.x[2] << endl;

    for (i = x + 1; i + 1 < (int)q.size(); i += 2)
    {
        if (bval(q[i]) < 4 && bval(q[i + 1]) < 4)
        {

            c1 = cbval(q[i]);
            c2 = cbval(q[i + 1]);

           // double tm_t = realtime();
            //cout << BASE[c2] << BASE[c1] << endl;
            bwt_extend2(bwt, &ik, ok, 0, c1, c2);
            // t7 += realtime() - tm_t;

            ik = ok[c2];
            ik.info = i + 1;
            //cout << "bwt-2: " << ik.x[0] << '\t' << ik.x[1] << '\t' << ik.x[2] << endl;
        }
        else
        {
            break;
        }
    }
    if (i < (int)q.size() && bval(q[i]) < 4)
    { // 最后一次扩展
        c1 = cbval(q[i]);
        // double tm_t = realtime();
        bwt_extend(bwt, &ik, ok, 0);
        // t10 += realtime() - tm_t;
        ik = ok[c1];
        ik.info = i + 1;
        //cout << "bwt-2: " << ik.x[0] << '\t' << ik.x[1] << '\t' << ik.x[2] << endl;
    }
    // cout << ik.x[0] << '\t' << ik.x[1] << '\t' << ik.x[2] << endl;
    return ik;
}