示例#1
0
文件: tmap_map1.c 项目: Brainiarc7/TS
void
tmap_map1_print_max_diff(tmap_map_opt_t *opt, int32_t stage)
{
  int32_t i, k, l;

  // initialize
  for(i=0;i<=TMAP_MAP_OPT_MAX_DIFF_READ_LENGTH;i++) {
      opt->max_diff_table[i] = 0;
  }

  if(opt->max_diff < 0) {
      if(0 < stage) tmap_progress_print("calculating maximum differences in map1 for stage %d", stage);
      else tmap_progress_print("calculating maximum differences in map1");

      for(i = 17, k = 0;i <= TMAP_MAP_OPT_MAX_DIFF_READ_LENGTH;i++) {
          l = tmap_map1_cal_maxdiff(i, opt->max_err_rate, opt->max_diff_fnr);
          if(l != k ) {
              tmap_progress_print("%dbp reads will have at most %d differences", i, l);
          }
          opt->max_diff_table[i] = l;
          k = l;
      }
  }
  else {
      for(i=0;i <= TMAP_MAP_OPT_MAX_DIFF_READ_LENGTH;i++) {
          opt->max_diff_table[i] = opt->max_diff;
      }
  }
}
示例#2
0
static void 
tmap_index_core(tmap_index_opt_t *opt)
{
  uint64_t ref_len = 0;

  // pack the reference sequence
  ref_len = tmap_refseq_fasta2pac(opt->fn_fasta, TMAP_FILE_NO_COMPRESSION, 0);
      
  if(TMAP_INDEX_TOO_BIG_GENOME <= ref_len) { // too big (2^32 - 1)!
      tmap_error("Reference sequence too large", Exit, OutOfRange);
  }

  // check returned genome size
  if(opt->is_large < 0) {
      if(TMAP_INDEX_LARGE_GENOME <= ref_len) { 
          opt->is_large = 1;
          tmap_progress_print("defaulting to \"bwtsw\" BWT construction algorithm");
      }
      else {
          opt->is_large = 0;
          tmap_progress_print("defaulting to \"is\" BWT construction algorithm");
      }
  }

  // create the bwt 
  tmap_bwt_pac2bwt(opt->fn_fasta, opt->is_large, opt->occ_interval, opt->hash_width, opt->check_hash);

  // create the suffix array
  tmap_sa_bwt2sa(opt->fn_fasta, opt->sa_interval);

  // pack the reference sequence
  ref_len = tmap_refseq_fasta2pac(opt->fn_fasta, TMAP_FILE_NO_COMPRESSION, 1);
}
示例#3
0
文件: tmap_shm.c 项目: Brainiarc7/TS
static int32_t
tmap_shmget(key_t key, size_t size, int32_t shmflg, int32_t create)
{
  int32_t shmid, i;

  if(0 == create) {
      // try a number of times before failing
      for(i=0,shmid=-1;shmid<0 && i<TMAP_SHMGET_RETRIES-1;i++) {
          if(0 <= (shmid = shmget(key, size, shmflg))) {
              return shmid;
          }
          tmap_progress_print("could not get shared memory, %d more %s", 
                              TMAP_SHMGET_RETRIES-i-1,
                              (1 != TMAP_SHMGET_RETRIES-i-1) ? "retries" : "retry");
          tmap_progress_print("retrying in %d seconds", TMAP_SHMGET_SLEEP);
          // sleep and retry
          sleep(TMAP_SHMGET_SLEEP);
      }
  }
  if((shmid = shmget(key, size, shmflg)) < 0) {
      tmap_error(NULL, Exit, SharedMemoryGet);
  }

  return shmid;
}
示例#4
0
tmap_index_t*
tmap_index_init(const char *fn_fasta, key_t shm_key, int32_t mm)
{
  tmap_index_t *index = NULL;

  index = tmap_calloc(1, sizeof(tmap_index_t), "index");

  index->shm_key = shm_key;
  index->mm      = mm;

  // get the reference information
  // primary 65380; sa_intv: 32
  // seq_len = 97004
  //n_sa = 3032, sa 67973 .. 18446744073709551615

  if (1 == index->mm) {
      tmap_progress_print("Retrieving reference data from memory map");
      index->refseq = tmap_refseq_mm_read(fn_fasta);
      index->bwt = tmap_bwt_mm_read(fn_fasta);
      index->sa = tmap_sa_mm_read(fn_fasta);
      tmap_progress_print2("Reference data retrieved from memory map");
  } else if(0 == index->shm_key) {
      tmap_progress_print("reading in reference data");
      index->refseq = tmap_refseq_read(fn_fasta);
      index->bwt = tmap_bwt_read(fn_fasta);
      index->sa = tmap_sa_read(fn_fasta);
      tmap_progress_print2("reference data read in");
  }
  else {
      tmap_progress_print("retrieving reference data from shared memory");
      index->shm = tmap_shm_init(index->shm_key, 0, 0);
      if(NULL == (index->refseq = tmap_refseq_shm_unpack(tmap_shm_get_buffer(index->shm, TMAP_SHM_LISTING_REFSEQ)))) {
          tmap_error("the packed reference sequence was not found in shared memory", Exit, SharedMemoryListing);
      }
      if(NULL == (index->bwt = tmap_bwt_shm_unpack(tmap_shm_get_buffer(index->shm, TMAP_SHM_LISTING_BWT)))) {
          tmap_error("the BWT string was not found in shared memory", Exit, SharedMemoryListing);
      }
      if(NULL == (index->sa = tmap_sa_shm_unpack(tmap_shm_get_buffer(index->shm, TMAP_SHM_LISTING_SA)))) {
          tmap_error("the SA was not found in shared memory", Exit, SharedMemoryListing);
      }
      tmap_progress_print2("reference data retrieved from shared memory");
  }

  if((index->refseq->len << 1) != index->bwt->seq_len) {
      tmap_error("refseq and bwt lengths do not match", Exit, OutOfRange);
  }
  if((index->refseq->len << 1) != index->sa->seq_len) {
      tmap_error("refseq and sa lengths do not match", Exit, OutOfRange);
  }
  
  return index;
}
示例#5
0
void
tmap_refseq_pac2revpac(const char *fn_fasta)
{
  uint32_t i, j, c;
  tmap_refseq_t *refseq=NULL, *refseq_rev=NULL;

  tmap_progress_print("reversing the packed reference FASTA");

  refseq = tmap_refseq_read(fn_fasta, 0);

  // shallow copy
  refseq_rev = tmap_calloc(1, sizeof(tmap_refseq_t), "refseq_rev");
  (*refseq_rev) = (*refseq);

  // update sequence
  refseq_rev->seq = NULL;
  refseq_rev->seq = tmap_calloc(tmap_refseq_seq_memory(refseq->len), sizeof(uint8_t), "refseq_rev->seq");
  for(i=0;i<refseq->len;i++) {
      c = tmap_refseq_seq_i(refseq, i);
      j = refseq->len - i - 1;
      tmap_refseq_seq_store_i(refseq_rev, j, c);
  }

  // write
  tmap_refseq_write(refseq_rev, fn_fasta, 1);

  // free
  free(refseq_rev->seq);
  free(refseq_rev);
  tmap_refseq_destroy(refseq);

  tmap_progress_print2("reversed the packed reference FASTA");
}
示例#6
0
文件: tmap_index.c 项目: Lingrui/TS
tmap_index_t*
tmap_index_init(const char *fn_fasta, key_t shm_key)
{
  tmap_index_t *index = NULL;

  index = tmap_calloc(1, sizeof(tmap_index_t), "index");

  index->shm_key = shm_key;

  // get the reference information
  if(0 == index->shm_key) {
      tmap_progress_print("reading in reference data");
      index->refseq = tmap_refseq_read(fn_fasta);
      index->bwt = tmap_bwt_read(fn_fasta);
      index->sa = tmap_sa_read(fn_fasta);
      tmap_progress_print2("reference data read in");
  }
  else {
      tmap_progress_print("retrieving reference data from shared memory");
      index->shm = tmap_shm_init(index->shm_key, 0, 0);
      if(NULL == (index->refseq = tmap_refseq_shm_unpack(tmap_shm_get_buffer(index->shm, TMAP_SHM_LISTING_REFSEQ)))) {
          tmap_error("the packed reference sequence was not found in shared memory", Exit, SharedMemoryListing);
      }
      if(NULL == (index->bwt = tmap_bwt_shm_unpack(tmap_shm_get_buffer(index->shm, TMAP_SHM_LISTING_BWT)))) {
          tmap_error("the BWT string was not found in shared memory", Exit, SharedMemoryListing);
      }
      if(NULL == (index->sa = tmap_sa_shm_unpack(tmap_shm_get_buffer(index->shm, TMAP_SHM_LISTING_SA)))) {
          tmap_error("the SA was not found in shared memory", Exit, SharedMemoryListing);
      }
      tmap_progress_print2("reference data retrieved from shared memory");
  }

  if((index->refseq->len << 1) != index->bwt->seq_len) {
      tmap_error("refseq and bwt lengths do not match", Exit, OutOfRange);
  }
  if((index->refseq->len << 1) != index->sa->seq_len) {
      tmap_error("refseq and sa lengths do not match", Exit, OutOfRange);
  }
  
  return index;
}
示例#7
0
void
tmap_sa_bwt2sa(const char *fn_fasta, uint32_t intv)
{
    int64_t isa, s; // S(isa) = sa
    uint64_t i;
    tmap_bwt_t *bwt = NULL;
    tmap_sa_t *sa = NULL;

    tmap_progress_print("constructing the SA from the BWT string");

    bwt = tmap_bwt_read(fn_fasta);

    sa = tmap_calloc(1, sizeof(tmap_sa_t), "sa");

    sa->primary = bwt->primary;
    sa->sa_intv = intv;
    sa->seq_len = bwt->seq_len;
    sa->n_sa = (bwt->seq_len + intv) / intv;

    // calculate SA value
    sa->sa = tmap_calloc(sa->n_sa, sizeof(tmap_bwt_int_t), "sa->sa");
    isa = 0;
    s = bwt->seq_len;
    for(i = 0; i < bwt->seq_len; ++i) {
        if(isa % intv == 0) sa->sa[isa/intv] = s;
        --s;
        isa = tmap_bwt_invPsi(bwt, isa);
    }
    if(isa % intv == 0) sa->sa[isa/intv] = s;
    sa->sa[0] = (tmap_bwt_int_t)-1; // before this line, bwt->sa[0] = bwt->seq_len

    tmap_sa_write(fn_fasta, sa);

    tmap_bwt_destroy(bwt);
    tmap_sa_destroy(sa);
    sa=NULL;
    bwt=NULL;

    tmap_progress_print2("constructed the SA from the BWT string");
}
示例#8
0
文件: tmap_shm.c 项目: Brainiarc7/TS
tmap_shm_t *
tmap_shm_init(key_t key, size_t size, int32_t create)
{
  tmap_shm_t *shm = NULL;
  int32_t i, shmflg = 0;
  struct shmid_ds buf;

  shm = tmap_calloc(1, sizeof(tmap_shm_t), "shm");
  shm->key = key;
  shm->size = size;

  if(1 == create) {
      shm->size += sizeof(uint32_t); // add for synchronization
      shm->size += sizeof(uint32_t); // add for on/off bits for listing what is in memory
      shm->size += 32*sizeof(size_t); // add for the byte size of each listing
      shmflg = IPC_CREAT | IPC_EXCL | 0666;
      shm->creator = 1;
  }
  else {
      shmflg = 0666;
      shm->creator = 0;
  }

  // get the shared memory id
  shm->shmid = tmap_shmget(shm->key, shm->size, shmflg, create);

  // attach the shared memory
  shm->ptr = tmap_shmat(shm->shmid, NULL, 0);
  shm->buf = ((uint8_t*)shm->ptr);
  shm->buf += sizeof(uint32_t); // synchronization 
  shm->buf += sizeof(uint32_t) + 32*sizeof(size_t); // listings

  tmap_shmctl(shm->shmid, IPC_STAT, &buf);
  if(1 == create) {
      // check that the current process created the shared memory
      if(buf.shm_cpid != getpid() || TMAP_SHM_READY == tmap_shm_get_state(shm)) {
          tmap_error("shared memory was not created by the current process", Exit, OutOfRange);
      }
      tmap_shm_set_not_ready(shm);
      if(buf.shm_segsz != shm->size) {
          tmap_error("shared memory size does not match the expected size", Exit, OutOfRange);
      }
  }
  else {
      // try a number of times before failing
      for(i=0;i<TMAP_SHMGET_RETRIES;i++) {
          if(TMAP_SHM_READY == tmap_shm_get_state(shm)) {
              break;
          }
          tmap_progress_print("shared memory not ready, %d more retries", TMAP_SHMGET_RETRIES-i-1);
          tmap_progress_print("retrying in %d seconds", TMAP_SHMGET_SLEEP);
          // sleep and retry
          sleep(TMAP_SHMGET_SLEEP);
      }
      if(TMAP_SHMGET_RETRIES == i) {
          tmap_error("shared memory did not become available", Exit, SharedMemoryGet);
      }
      
      // set the size
      shm->size = buf.shm_segsz;
  }

  return shm;
}
示例#9
0
uint64_t
tmap_refseq_fasta2pac(const char *fn_fasta, int32_t compression)
{
  tmap_file_t *fp_pac = NULL, *fp_anno = NULL;
  tmap_seq_io_t *seqio = NULL;
  tmap_seq_t *seq = NULL;
  tmap_refseq_t *refseq = NULL;
  char *fn_pac = NULL, *fn_anno = NULL;
  uint8_t buffer[TMAP_REFSEQ_BUFFER_SIZE];
  int32_t i, j, l, buffer_length;
  uint32_t num_IUPAC_found= 0, amb_bases_mem = 0;
  uint8_t x = 0;
  uint64_t ref_len;

  tmap_progress_print("packing the reference FASTA");

  refseq = tmap_calloc(1, sizeof(tmap_refseq_t), "refseq");

  refseq->version_id = TMAP_VERSION_ID; 
  refseq->package_version = tmap_string_clone2(PACKAGE_VERSION);
  refseq->seq = buffer; // IMPORTANT: must nullify later
  refseq->annos = NULL;
  refseq->num_annos = 0;
  refseq->len = 0;
  refseq->is_rev = 0;
  refseq->is_shm = 0;
  memset(buffer, 0, TMAP_REFSEQ_BUFFER_SIZE);
  buffer_length = 0;

  // input files
  seqio = tmap_seq_io_init(fn_fasta, TMAP_SEQ_TYPE_FQ, 0, compression);
  seq = tmap_seq_init(TMAP_SEQ_TYPE_FQ);

  // output files
  fn_pac = tmap_get_file_name(fn_fasta, TMAP_PAC_FILE);
  fp_pac = tmap_file_fopen(fn_pac, "wb", TMAP_PAC_COMPRESSION);

  // read in sequences
  while(0 <= (l = tmap_seq_io_read(seqio, seq))) {
      tmap_anno_t *anno = NULL;
      tmap_progress_print2("packing contig [%s:1-%d]", seq->data.fq->name->s, l);

      refseq->num_annos++;
      refseq->annos = tmap_realloc(refseq->annos, sizeof(tmap_anno_t)*refseq->num_annos, "refseq->annos");
      anno = &refseq->annos[refseq->num_annos-1];
      
      anno->name = tmap_string_clone(seq->data.fq->name); 
      anno->len = l;
      anno->offset = (1 == refseq->num_annos) ? 0 : refseq->annos[refseq->num_annos-2].offset + refseq->annos[refseq->num_annos-2].len;
      anno->amb_positions_start = NULL;
      anno->amb_positions_end = NULL;
      anno->amb_bases = NULL;
      anno->num_amb = 0;
      amb_bases_mem = 0;

      // fill the buffer
      for(i=0;i<l;i++) {
          uint8_t c = tmap_nt_char_to_int[(int)seq->data.fq->seq->s[i]];
          // handle IUPAC codes 
          if(4 <= c) {
              int32_t k;
              // warn users about IUPAC codes
              if(0 == num_IUPAC_found) { 
                  tmap_error("IUPAC codes were found and will be converted to non-matching DNA bases", Warn, OutOfRange);
                  for(j=4;j<15;j++) {
                      c = tmap_iupac_char_to_bit_string[(int)tmap_iupac_int_to_char[j]];
                      // get the lexicographically smallest base not compatible with this code
                      for(k=0;k<4;k++) {
                          if(!(c & (0x1 << k))) {
                              break;
                          }
                      } 
                      tmap_progress_print2("IUPAC code %c will be converted to %c", tmap_iupac_int_to_char[j], "ACGTN"[k & 3]);
                  }
              }
              num_IUPAC_found++;
              
              // change it to a mismatched base than the IUPAC code
              c = tmap_iupac_char_to_bit_string[(int)seq->data.fq->seq->s[i]];

              // store IUPAC bases
              if(amb_bases_mem <= anno->num_amb) { // allocate more memory if necessary
                  amb_bases_mem = anno->num_amb + 1;
                  tmap_roundup32(amb_bases_mem);
                  anno->amb_positions_start = tmap_realloc(anno->amb_positions_start, sizeof(uint32_t) * amb_bases_mem, "anno->amb_positions_start");
                  anno->amb_positions_end = tmap_realloc(anno->amb_positions_end, sizeof(uint32_t) * amb_bases_mem, "anno->amb_positions_end");
                  anno->amb_bases = tmap_realloc(anno->amb_bases, sizeof(uint8_t) * amb_bases_mem, "anno->amb_bases");
              }
              // encode stretches of the same base
              if(0 < anno->num_amb
                 && anno->amb_positions_end[anno->num_amb-1] == i
                 && anno->amb_bases[anno->num_amb-1] == tmap_iupac_char_to_int[(int)seq->data.fq->seq->s[i]]) {
                 anno->amb_positions_end[anno->num_amb-1]++; // expand the range 
              }
              else {
                  // new ambiguous base and range
                  anno->num_amb++;
                  anno->amb_positions_start[anno->num_amb-1] = i+1; // one-based
                  anno->amb_positions_end[anno->num_amb-1] = i+1; // one-based
                  anno->amb_bases[anno->num_amb-1] = tmap_iupac_char_to_int[(int)seq->data.fq->seq->s[i]];
              }
              
              // get the lexicographically smallest base not compatible with
              // this code
              for(j=0;j<4;j++) {
                  if(!(c & (0x1 << j))) {
                      break;
                  }
              } 
              c = j & 3; // Note: Ns will go to As
          }
          if(3 < c) {
              tmap_error("bug encountered", Exit, OutOfRange);
          }
          if(buffer_length == (TMAP_REFSEQ_BUFFER_SIZE << 2)) { // 2-bit
              if(tmap_refseq_seq_memory(buffer_length) != tmap_file_fwrite(buffer, sizeof(uint8_t), tmap_refseq_seq_memory(buffer_length), fp_pac)) {
                  tmap_error(fn_pac, Exit, WriteFileError);
              }
              memset(buffer, 0, TMAP_REFSEQ_BUFFER_SIZE);
              buffer_length = 0;
          }
          tmap_refseq_seq_store_i(refseq, buffer_length, c);
          buffer_length++;
      }
      refseq->len += l;
      // re-size the amibiguous bases
      if(anno->num_amb < amb_bases_mem) {
          amb_bases_mem = anno->num_amb;
          anno->amb_positions_start = tmap_realloc(anno->amb_positions_start, sizeof(uint32_t) * amb_bases_mem, "anno->amb_positions_start");
          anno->amb_positions_end = tmap_realloc(anno->amb_positions_end, sizeof(uint32_t) * amb_bases_mem, "anno->amb_positions_end");
          anno->amb_bases = tmap_realloc(anno->amb_bases, sizeof(uint8_t) * amb_bases_mem, "anno->amb_bases");
      }
  }
  // write out the buffer
  if(tmap_refseq_seq_memory(buffer_length) != tmap_file_fwrite(buffer, sizeof(uint8_t), tmap_refseq_seq_memory(buffer_length), fp_pac)) {
      tmap_error(fn_pac, Exit, WriteFileError);
  }
  if(refseq->len % 4 == 0) { // add an extra byte if we completely filled all bits
      if(1 != tmap_file_fwrite(&x, sizeof(uint8_t), 1, fp_pac)) {
          tmap_error(fn_pac, Exit, WriteFileError);
      }
  }
  // store number of unused bits at the last byte
  x = refseq->len % 4;
  if(1 != tmap_file_fwrite(&x, sizeof(uint8_t), 1, fp_pac)) {
      tmap_error(fn_pac, Exit, WriteFileError);
  }
  refseq->seq = NULL; // IMPORTANT: nullify this
  ref_len = refseq->len; // save for return
      
  tmap_progress_print2("total genome length [%u]", refseq->len);
  if(0 < num_IUPAC_found) {
      if(1 == num_IUPAC_found) {
          tmap_progress_print("%u IUPAC base was found and converted to a DNA base", num_IUPAC_found);
      }
      else {
          tmap_progress_print("%u IUPAC bases were found and converted to DNA bases", num_IUPAC_found);
      }
  }

  // write annotation file
  fn_anno = tmap_get_file_name(fn_fasta, TMAP_ANNO_FILE);
  fp_anno = tmap_file_fopen(fn_anno, "wb", TMAP_ANNO_COMPRESSION);
  tmap_refseq_write_anno(fp_anno, refseq); 

  // close files
  tmap_file_fclose(fp_pac);
  tmap_file_fclose(fp_anno);

  // check sequence name uniqueness
  for(i=0;i<refseq->num_annos;i++) {
      for(j=i+1;j<refseq->num_annos;j++) {
          if(0 == strcmp(refseq->annos[i].name->s, refseq->annos[j].name->s)) {
              tmap_file_fprintf(tmap_file_stderr, "Contigs have the same name: #%d [%s] and #%d [%s]\n",
                                i+1, refseq->annos[i].name->s, 
                                j+1, refseq->annos[j].name->s); 
              tmap_error("Contig names must be unique", Exit, OutOfRange);
          }
      }
  }

  tmap_refseq_destroy(refseq); 
  tmap_seq_io_destroy(seqio);
  tmap_seq_destroy(seq);
  free(fn_pac);
  free(fn_anno);

  tmap_progress_print2("packed the reference FASTA");

  tmap_refseq_pac2revpac(fn_fasta);

  return ref_len;
}