int
main(int argc, char **argv)
{
ESL_GETOPTS *go = NULL;
ESL_ALPHABET *nt_abc = esl_alphabet_Create(eslDNA);
ESL_ALPHABET *aa_abc = esl_alphabet_Create(eslAMINO);
ESL_GENCODE *gcode = NULL;
ESL_GENCODE_WORKSTATE *wrk = NULL;
char *seqfile = NULL;
int informat = eslSQFILE_UNKNOWN;
ESL_SQFILE *sqfp = NULL;
int status;
/*****************************************************************
* command line parsing
*****************************************************************/
go = esl_getopts_Create(options);
if (esl_opt_ProcessCmdline(go, argc, argv) != eslOK ||
esl_opt_VerifyConfig(go) != eslOK)
{
printf("Failed to parse command line: %s\n", go->errbuf);
esl_usage(stdout, argv[0], usage);
printf("\nTo see more help on available options, do %s -h\n\n", argv[0]);
exit(1);
}
if (esl_opt_GetBoolean(go, "-h") )
{
esl_banner(stdout, argv[0], banner);
esl_usage (stdout, argv[0], usage);
puts("\n where options are:");
esl_opt_DisplayHelp(stdout, go, /*docgroup=*/0, /*indent=*/2, /*textwidth=*/80);
puts("\nAvailable NCBI genetic code tables (for -c <id>):");
esl_gencode_DumpAltCodeTable(stdout);
exit(0);
}
if (esl_opt_ArgNumber(go) != 1)
{
printf("Incorrect number of command line arguments.\n");
esl_usage(stdout, argv[0], usage);
printf("\nTo see more help on available options, do %s -h\n\n", argv[0]);
exit(1);
}
seqfile = esl_opt_GetArg(go, 1);
if (esl_opt_IsOn(go, "--informat") &&
(informat = esl_sqio_EncodeFormat(esl_opt_GetString(go, "--informat"))) == eslMSAFILE_UNKNOWN)
esl_fatal("%s is not a valid input sequence file format for --informat", esl_opt_GetString(go, "--informat"));
status = esl_sqfile_OpenDigital(nt_abc, seqfile, informat, /*env=*/NULL, &sqfp);
if (status == eslENOTFOUND) esl_fatal("Failed to find (or open) sequence file %s", seqfile);
else if (status == eslEFORMAT) esl_fatal("Failed to recognize format of sequence file %s", seqfile);
else if (status != eslOK) esl_fatal("Failure in opening sequence file %s; code %d", seqfile, status);
/* A limitation. The esl_sqio_ReadWindow() interface needs to use SSI positioning
* to read reverse complement, and that doesn't work on nonrewindable streams.
*/
if ( esl_opt_GetBoolean(go, "-W") && ! esl_sqfile_IsRewindable(sqfp) && ! esl_opt_GetBoolean(go, "--watson"))
esl_fatal("esl-translate can't read reverse complement from a nonrewindable stream (stdin pipe, .gz file, etc).");
/* Set up the genetic code. Default = NCBI 1, the standard code; allow ORFs to start at any aa
*/
gcode = esl_gencode_Create(nt_abc, aa_abc);
esl_gencode_Set(gcode, esl_opt_GetInteger(go, "-c")); // default = 1, the standard genetic code
if (esl_opt_GetBoolean(go, "-m")) esl_gencode_SetInitiatorOnlyAUG(gcode);
else if (! esl_opt_GetBoolean(go, "-M")) esl_gencode_SetInitiatorAny(gcode); // note this is the default, if neither -m or -M are set
/* Set up the workstate structure, which contains both stateful
* info about our position in <sqfp> and the DNA <sq>, as well as
* one-time config info from options
*/
wrk = esl_gencode_WorkstateCreate(go, gcode);
/* The two styles of main processing loop:
*/
if (esl_opt_GetBoolean(go, "-W")) do_by_windows(gcode, wrk, sqfp);
else do_by_sequences(gcode, wrk, sqfp);
esl_gencode_WorkstateDestroy(wrk);
esl_sqfile_Close(sqfp);
esl_gencode_Destroy(gcode);
esl_alphabet_Destroy(aa_abc);
esl_alphabet_Destroy(nt_abc);
esl_getopts_Destroy(go);
return 0;
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = NULL;
char *seqfile = NULL;
ESL_SQFILE *sqfp = NULL;
int infmt = eslSQFILE_UNKNOWN;
int alphatype = eslUNKNOWN;
ESL_ALPHABET *abc = NULL;
ESL_SQ *sq = NULL;
int64_t nseq = 0;
int64_t nres = 0;
int64_t small = 0;
int64_t large = 0;
double *monoc = NULL; /* monoresidue composition per sequence */
double *monoc_all = NULL; /* monoresidue composition over all seqs */
int do_comp = FALSE;
int status = eslOK;
int wstatus;
int i;
int do_stall; /* used to stall when debugging */
/* Parse command line */
go = esl_getopts_Create(options);
if (esl_opt_ProcessCmdline(go, argc, argv) != eslOK) cmdline_failure(argv[0], "Failed to parse command line: %s\n", go->errbuf);
if (esl_opt_VerifyConfig(go) != eslOK) cmdline_failure(argv[0], "Error in app configuration: %s\n", go->errbuf);
if (esl_opt_GetBoolean(go, "-h") ) cmdline_help(argv[0], go);
if (esl_opt_ArgNumber(go) != 1) cmdline_failure(argv[0], "Incorrect number of command line arguments.\n");
seqfile = esl_opt_GetArg(go, 1);
do_comp = esl_opt_GetBoolean(go, "-c");
if (esl_opt_GetString(go, "--informat") != NULL) {
infmt = esl_sqio_EncodeFormat(esl_opt_GetString(go, "--informat"));
if (infmt == eslSQFILE_UNKNOWN) esl_fatal("%s is not a valid input sequence file format for --informat");
}
do_stall = esl_opt_GetBoolean(go, "--stall"); /* a stall point for attaching gdb */
while (do_stall);
/* open input file */
status = esl_sqfile_Open(seqfile, infmt, NULL, &sqfp);
if (status == eslENOTFOUND) esl_fatal("No such file %s", seqfile);
else if (status == eslEFORMAT) esl_fatal("Format of seqfile %s unrecognized.", seqfile);
else if (status != eslOK) esl_fatal("Open failed, code %d.", status);
if (esl_opt_GetBoolean(go, "--rna")) alphatype = eslRNA;
else if (esl_opt_GetBoolean(go, "--dna")) alphatype = eslDNA;
else if (esl_opt_GetBoolean(go, "--amino")) alphatype = eslAMINO;
else {
status = esl_sqfile_GuessAlphabet(sqfp, &alphatype);
if (status == eslEAMBIGUOUS) esl_fatal("Couldn't guess alphabet from first sequence in %s", seqfile);
else if (status == eslEFORMAT) esl_fatal("Parse failed (sequence file %s):\n%s\n",
sqfp->filename, esl_sqfile_GetErrorBuf(sqfp));
else if (status == eslENODATA) esl_fatal("Sequence file %s contains no data?", seqfile);
else if (status != eslOK) esl_fatal("Failed to guess alphabet (error code %d)\n", status);
}
abc = esl_alphabet_Create(alphatype);
sq = esl_sq_CreateDigital(abc);
esl_sqfile_SetDigital(sqfp, abc);
if (do_comp) {
ESL_ALLOC(monoc, (abc->Kp) * sizeof(double));
ESL_ALLOC(monoc_all, (abc->Kp) * sizeof(double));
esl_vec_DSet(monoc_all, abc->Kp, 0.0);
esl_vec_DSet(monoc, abc->Kp, 0.0);
}
while ((wstatus = esl_sqio_ReadWindow(sqfp, 0, 4096, sq)) != eslEOF)
{
if (wstatus == eslOK)
{
if (do_comp)
for (i = 1; i <= sq->n; i++)
monoc[sq->dsq[i]]++;
}
else if (wstatus == eslEOD)
{
if (nseq == 0) { small = large = sq->L; }
else {
small = ESL_MIN(small, sq->L);
large = ESL_MAX(large, sq->L);
}
if (esl_opt_GetBoolean(go, "-a")) {
printf("= %-20s %8" PRId64 " %s\n", sq->name, sq->L, (sq->desc != NULL) ? sq->desc : "");
}
nres += sq->L;
nseq++;
esl_sq_Reuse(sq);
if (do_comp) {
esl_vec_DAdd(monoc_all, monoc, abc->Kp);
esl_vec_DSet(monoc, abc->Kp, 0.0);
}
}
else if (wstatus == eslEFORMAT) esl_fatal("Parse failed (sequence file %s):\n%s\n",
sqfp->filename, esl_sqfile_GetErrorBuf(sqfp));
else esl_fatal("Unexpected error %d reading sequence file %s",
wstatus, sqfp->filename);
}
printf("Format: %s\n", esl_sqio_DecodeFormat(sqfp->format));
printf("Alphabet type: %s\n", esl_abc_DecodeType(abc->type));
printf("Number of sequences: %" PRId64 "\n", nseq);
printf("Total # residues: %" PRId64 "\n", nres);
printf("Smallest: %" PRId64 "\n", small);
printf("Largest: %" PRId64 "\n", large);
printf("Average length: %.1f\n", (float) nres / (float) nseq);
if (do_comp) {
show_overall_composition(abc, monoc_all, nres);
free(monoc);
free(monoc_all);
}
esl_alphabet_Destroy(abc);
esl_sq_Destroy(sq);
esl_sqfile_Close(sqfp);
esl_getopts_Destroy(go);
return 0;
ERROR:
return status;
}
/* serial_master()
* The serial version of hmmsearch.
* For each query HMM in <hmmdb> search the database for hits.
*
* A master can only return if it's successful. All errors are handled
* immediately and fatally with p7_Fail().
*/
static int
serial_master(ESL_GETOPTS *go, struct cfg_s *cfg)
{
FILE *ofp = stdout; /* output file for results (default stdout) */
FILE *tblfp = NULL; /* output stream for tabular per-seq (--tblout) */
FILE *dfamtblfp = NULL; /* output stream for tabular Dfam format (--dfamtblout) */
FILE *aliscoresfp = NULL; /* output stream for alignment scores (--aliscoresout) */
// P7_HMM *hmm = NULL; /* one HMM query */
// P7_SCOREDATA *scoredata = NULL;
int seqfmt = eslSQFILE_UNKNOWN; /* format of seqfile */
ESL_SQFILE *sqfp = NULL; /* open seqfile */
P7_HMMFILE *hfp = NULL; /* open HMM database file */
ESL_ALPHABET *abc = NULL; /* sequence alphabet */
P7_OPROFILE *om = NULL; /* target profile */
ESL_STOPWATCH *w = NULL; /* timing */
ESL_SQ *qsq = NULL; /* query sequence */
int nquery = 0;
int textw;
int status = eslOK;
int hstatus = eslOK;
int sstatus = eslOK;
int i;
int ncpus = 0;
int infocnt = 0;
WORKER_INFO *info = NULL;
#ifdef HMMER_THREADS
P7_OM_BLOCK *block = NULL;
ESL_THREADS *threadObj= NULL;
ESL_WORK_QUEUE *queue = NULL;
#endif
char errbuf[eslERRBUFSIZE];
double window_beta = -1.0 ;
int window_length = -1;
if (esl_opt_IsUsed(go, "--w_beta")) { if ( ( window_beta = esl_opt_GetReal(go, "--w_beta") ) < 0 || window_beta > 1 ) esl_fatal("Invalid window-length beta value\n"); }
if (esl_opt_IsUsed(go, "--w_length")) { if (( window_length = esl_opt_GetInteger(go, "--w_length")) < 4 ) esl_fatal("Invalid window length value\n"); }
w = esl_stopwatch_Create();
if (esl_opt_GetBoolean(go, "--notextw")) textw = 0;
else textw = esl_opt_GetInteger(go, "--textw");
/* If caller declared an input format, decode it */
if (esl_opt_IsOn(go, "--qformat")) {
seqfmt = esl_sqio_EncodeFormat(esl_opt_GetString(go, "--qformat"));
if (seqfmt == eslSQFILE_UNKNOWN) p7_Fail("%s is not a recognized input sequence file format\n", esl_opt_GetString(go, "--qformat"));
}
/* validate options if running as a daemon */
// if (esl_opt_IsOn(go, "--daemon")) {
/* running as a daemon, the input format must be type daemon */
// if (seqfmt != eslSQFILE_UNKNOWN && seqfmt != eslSQFILE_DAEMON)
// esl_fatal("Input format %s not supported. Must be daemon\n", esl_opt_GetString(go, "--qformat"));
// seqfmt = eslSQFILE_DAEMON;
// if (strcmp(cfg->seqfile, "-") != 0) esl_fatal("Query sequence file must be '-'\n");
// }
/* Open the target profile database to get the sequence alphabet */
status = p7_hmmfile_OpenE(cfg->hmmfile, p7_HMMDBENV, &hfp, errbuf);
if (status == eslENOTFOUND) p7_Fail("File existence/permissions problem in trying to open HMM file %s.\n%s\n", cfg->hmmfile, errbuf);
else if (status == eslEFORMAT) p7_Fail("File format problem, trying to open HMM file %s.\n%s\n", cfg->hmmfile, errbuf);
else if (status != eslOK) p7_Fail("Unexpected error %d in opening HMM file %s.\n%s\n", status, cfg->hmmfile, errbuf);
if (! hfp->is_pressed) p7_Fail("Failed to open binary auxfiles for %s: use hmmpress first\n", hfp->fname);
hstatus = p7_oprofile_ReadMSV(hfp, &abc, &om);
if (hstatus == eslEFORMAT) p7_Fail("bad format, binary auxfiles, %s:\n%s", cfg->hmmfile, hfp->errbuf);
else if (hstatus == eslEINCOMPAT) p7_Fail("HMM file %s contains different alphabets", cfg->hmmfile);
else if (hstatus != eslOK) p7_Fail("Unexpected error in reading HMMs from %s", cfg->hmmfile);
p7_oprofile_Destroy(om);
p7_hmmfile_Close(hfp);
/* Open the query sequence database */
status = esl_sqfile_OpenDigital(abc, cfg->seqfile, seqfmt, NULL, &sqfp);
if (status == eslENOTFOUND) p7_Fail("Failed to open sequence file %s for reading\n", cfg->seqfile);
else if (status == eslEFORMAT) p7_Fail("Sequence file %s is empty or misformatted\n", cfg->seqfile);
else if (status == eslEINVAL) p7_Fail("Can't autodetect format of a stdin or .gz seqfile");
else if (status != eslOK) p7_Fail("Unexpected error %d opening sequence file %s\n", status, cfg->seqfile);
if (sqfp->format > 100) // breaking the law! That range is reserved for msa, for aligned formats
p7_Fail("%s contains a multiple sequence alignment; expect unaligned sequences, like FASTA\n", cfg->seqfile);
qsq = esl_sq_CreateDigital(abc);
/* Open the results output files */
if (esl_opt_IsOn(go, "-o")) { if ((ofp = fopen(esl_opt_GetString(go, "-o"), "w")) == NULL) esl_fatal("Failed to open output file %s for writing\n", esl_opt_GetString(go, "-o")); }
if (esl_opt_IsOn(go, "--tblout")) { if ((tblfp = fopen(esl_opt_GetString(go, "--tblout"), "w")) == NULL) esl_fatal("Failed to open tabular per-seq output file %s for writing\n", esl_opt_GetString(go, "--tblfp")); }
if (esl_opt_IsOn(go, "--dfamtblout")) { if ((dfamtblfp = fopen(esl_opt_GetString(go, "--dfamtblout"),"w")) == NULL) esl_fatal("Failed to open tabular dfam output file %s for writing\n", esl_opt_GetString(go, "--dfamtblout")); }
if (esl_opt_IsOn(go, "--aliscoresout")) { if ((aliscoresfp = fopen(esl_opt_GetString(go, "--aliscoresout"),"w")) == NULL) esl_fatal("Failed to open alignment scores output file %s for writing\n", esl_opt_GetString(go, "--aliscoresout")); }
output_header(ofp, go, cfg->hmmfile, cfg->seqfile);
#ifdef HMMER_THREADS
/* initialize thread data */
if (esl_opt_IsOn(go, "--cpu")) ncpus = esl_opt_GetInteger(go, "--cpu");
else esl_threads_CPUCount(&ncpus);
if (ncpus > 0)
{
threadObj = esl_threads_Create(&pipeline_thread);
queue = esl_workqueue_Create(ncpus * 2);
}
#endif
infocnt = (ncpus == 0) ? 1 : ncpus;
ESL_ALLOC(info, sizeof(*info) * infocnt);
for (i = 0; i < infocnt; ++i)
{
info[i].bg = p7_bg_Create(abc);
#ifdef HMMER_THREADS
info[i].queue = queue;
#endif
}
#ifdef HMMER_THREADS
for (i = 0; i < ncpus * 2; ++i)
{
block = p7_oprofile_CreateBlock(BLOCK_SIZE);
if (block == NULL) esl_fatal("Failed to allocate sequence block");
status = esl_workqueue_Init(queue, block);
if (status != eslOK) esl_fatal("Failed to add block to work queue");
}
#endif
/* Outside loop: over each query sequence in <seqfile>. */
while ((sstatus = esl_sqio_Read(sqfp, qsq)) == eslOK)
{
if (sstatus == eslEMEM) p7_Fail("Memory allocation error reading sequence file\n", status);
if (sstatus == eslEINCONCEIVABLE) p7_Fail("Unexpected error %d reading sequence file\n", status);
// if (qsq->L > NHMMER_MAX_RESIDUE_COUNT) p7_Fail("Input sequence %s in file %s exceeds maximum length of %d bases.\n", qsq->name, cfg->seqfile, NHMMER_MAX_RESIDUE_COUNT);
nquery++;
esl_stopwatch_Start(w);
/* Open the target profile database */
status = p7_hmmfile_OpenE(cfg->hmmfile, p7_HMMDBENV, &hfp, NULL);
if (status != eslOK) p7_Fail("Unexpected error %d in opening hmm file %s.\n", status, cfg->hmmfile);
#ifdef HMMER_THREADS
/* if we are threaded, create a lock to prevent multiple readers */
if (ncpus > 0)
{
status = p7_hmmfile_CreateLock(hfp);
if (status != eslOK) p7_Fail("Unexpected error %d creating lock\n", status);
}
#endif
if (fprintf(ofp, "Query: %s [L=%ld]\n", qsq->name, (long) qsq->n) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (qsq->acc[0] != 0 && fprintf(ofp, "Accession: %s\n", qsq->acc) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (qsq->desc[0] != 0 && fprintf(ofp, "Description: %s\n", qsq->desc) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
for (i = 0; i < infocnt; ++i)
{
/* Create processing pipeline and hit list */
info[i].th = p7_tophits_Create();
info[i].pli = p7_pipeline_Create(go, 100, 100, TRUE, p7_SCAN_MODELS); /* M_hint = 100, L_hint = 100 are just dummies for now */
info[i].pli->hfp = hfp; /* for two-stage input, pipeline needs <hfp> */
p7_pli_NewSeq(info[i].pli, qsq);
info[i].qsq = qsq;
if ( esl_opt_IsUsed(go, "--toponly") )
info[i].pli->strand = p7_STRAND_TOPONLY;
else if ( esl_opt_IsUsed(go, "--bottomonly") )
info[i].pli->strand = p7_STRAND_BOTTOMONLY;
else
info[i].pli->strand = p7_STRAND_BOTH;
#ifdef HMMER_THREADS
if (ncpus > 0) esl_threads_AddThread(threadObj, &info[i]);
#endif
}
#ifdef HMMER_THREADS
if (ncpus > 0) hstatus = thread_loop(threadObj, queue, hfp);
else hstatus = serial_loop(info, hfp);
#else
hstatus = serial_loop(info, hfp);
#endif
switch(hstatus)
{
case eslEFORMAT: p7_Fail("bad file format in HMM file %s", cfg->hmmfile); break;
case eslEINCOMPAT: p7_Fail("HMM file %s contains different alphabets", cfg->hmmfile); break;
case eslEOF:
case eslOK: /* do nothing */
break;
default: p7_Fail("Unexpected error in reading HMMs from %s", cfg->hmmfile);
}
/* merge the results of the search results */
for (i = 1; i < infocnt; ++i)
{
p7_tophits_Merge(info[0].th, info[i].th);
p7_pipeline_Merge(info[0].pli, info[i].pli);
p7_pipeline_Destroy(info[i].pli);
p7_tophits_Destroy(info[i].th);
}
/* modify e-value to account for number of models */
for (i = 0; i < info->th->N ; i++)
{
info->th->unsrt[i].lnP += log((float)info->pli->nmodels);
info->th->unsrt[i].dcl[0].lnP = info->th->unsrt[i].lnP;
info->th->unsrt[i].sortkey = -1.0 * info->th->unsrt[i].lnP;
}
/* it's possible to have duplicates based on how viterbi ranges can overlap */
p7_tophits_SortByModelnameAndAlipos(info->th);
p7_tophits_RemoveDuplicates(info->th, info->pli->use_bit_cutoffs);
/* Print results */
p7_tophits_SortBySortkey(info->th);
p7_tophits_Threshold(info->th, info->pli);
//tally up total number of hits and target coverage
info->pli->n_output = info->pli->pos_output = 0;
for (i = 0; i < info->th->N; i++) {
if ( (info->th->hit[i]->flags & p7_IS_REPORTED) || info->th->hit[i]->flags & p7_IS_INCLUDED) {
info->pli->n_output++;
info->pli->pos_output += abs(info->th->hit[i]->dcl[0].jali - info->th->hit[i]->dcl[0].iali) + 1;
}
}
p7_tophits_Targets(ofp, info->th, info->pli, textw); if (fprintf(ofp, "\n\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
p7_tophits_Domains(ofp, info->th, info->pli, textw); if (fprintf(ofp, "\n\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (tblfp) p7_tophits_TabularTargets(tblfp, qsq->name, qsq->acc, info->th, info->pli, (nquery == 1));
if (dfamtblfp) p7_tophits_TabularXfam(dfamtblfp, qsq->name, NULL, info->th, info->pli);
if (aliscoresfp) p7_tophits_AliScores(aliscoresfp, qsq->name, info->th );
esl_stopwatch_Stop(w);
info->pli->nseqs = 1;
p7_pli_Statistics(ofp, info->pli, w);
if (fprintf(ofp, "//\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
fflush(ofp);
p7_hmmfile_Close(hfp);
p7_pipeline_Destroy(info->pli);
p7_tophits_Destroy(info->th);
esl_sq_Reuse(qsq);
}
if (sstatus == eslEFORMAT) esl_fatal("Parse failed (sequence file %s):\n%s\n",
sqfp->filename, esl_sqfile_GetErrorBuf(sqfp));
else if (sstatus != eslEOF) esl_fatal("Unexpected error %d reading sequence file %s",
sstatus, sqfp->filename);
/* Terminate outputs - any last words?
*/
if (tblfp) p7_tophits_TabularTail(tblfp, "hmmscan", p7_SCAN_MODELS, cfg->seqfile, cfg->hmmfile, go);
if (ofp) { if (fprintf(ofp, "[ok]\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed"); }
/* Cleanup - prepare for successful exit
*/
for (i = 0; i < infocnt; ++i)
p7_bg_Destroy(info[i].bg);
#ifdef HMMER_THREADS
if (ncpus > 0)
{
esl_workqueue_Reset(queue);
while (esl_workqueue_Remove(queue, (void **) &block) == eslOK)
p7_oprofile_DestroyBlock(block);
esl_workqueue_Destroy(queue);
esl_threads_Destroy(threadObj);
}
#endif
free(info);
esl_sq_Destroy(qsq);
esl_stopwatch_Destroy(w);
esl_alphabet_Destroy(abc);
esl_sqfile_Close(sqfp);
if (ofp != stdout) fclose(ofp);
if (tblfp) fclose(tblfp);
if (dfamtblfp) fclose(dfamtblfp);
if (aliscoresfp) fclose(aliscoresfp);
return eslOK;
ERROR:
if (ofp != stdout) fclose(ofp);
if (tblfp) fclose(tblfp);
if (dfamtblfp) fclose(dfamtblfp);
if (aliscoresfp) fclose(aliscoresfp);
return status;
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = NULL; /* application configuration */
char *seqfile = NULL; /* sequence file name */
char *maskfile = NULL; /* mask coordinate file name */
int infmt = eslSQFILE_UNKNOWN; /* format code for seqfile */
int outfmt = eslSQFILE_FASTA; /* format code for output seqs */
ESL_SQFILE *sqfp = NULL; /* open sequence file */
ESL_FILEPARSER *maskefp = NULL; /* open mask coord file */
FILE *ofp = NULL; /* output stream for masked seqs */
char *source = NULL; /* name of current seq to mask */
char *p1, *p2; /* pointers used in parsing */
int64_t start, end; /* start, end coord for masking */
int64_t i, j, pos; /* coords in a sequence */
int64_t overmask; /* # of extra residues to mask */
ESL_SQ *sq = esl_sq_Create(); /* current sequence */
int do_fetching;
int do_lowercase;
int maskchar;
int status; /* easel return code */
/****************************************************************************
* Parse command line
****************************************************************************/
go = esl_getopts_Create(options);
if (esl_opt_ProcessCmdline(go, argc, argv) != eslOK) cmdline_failure(argv[0], "Failed to parse command line: %s\n", go->errbuf);
if (esl_opt_VerifyConfig(go) != eslOK) cmdline_failure(argv[0], "Error in configuration: %s\n", go->errbuf);
if (esl_opt_GetBoolean(go, "-h") ) cmdline_help (argv[0], go);
if (esl_opt_ArgNumber(go) != 2) cmdline_failure(argv[0], "Incorrect number of command line arguments.\n");
do_fetching = esl_opt_GetBoolean(go, "-R");
do_lowercase = esl_opt_GetBoolean(go, "-l");
overmask = (esl_opt_IsOn(go, "-x") ? esl_opt_GetInteger(go, "-x") : 0);
maskchar = (esl_opt_IsOn(go, "-m") ? esl_opt_GetChar(go, "-m") : 'X');
seqfile = esl_opt_GetArg(go, 1);
maskfile = esl_opt_GetArg(go, 2);
/* Open the <seqfile>: text mode, not digital */
if (esl_opt_GetString(go, "--informat") != NULL) {
infmt = esl_sqio_EncodeFormat(esl_opt_GetString(go, "--informat"));
if (infmt == eslSQFILE_UNKNOWN) cmdline_failure(argv[0], "%s is not a valid input sequence file format for --informat");
}
status = esl_sqfile_Open(seqfile, infmt, NULL, &sqfp);
if (status == eslENOTFOUND) cmdline_failure(argv[0], "Sequence file %s not found.\n", seqfile);
else if (status == eslEFORMAT) cmdline_failure(argv[0], "Format of file %s unrecognized.\n", seqfile);
else if (status == eslEINVAL) cmdline_failure(argv[0], "Can't autodetect stdin or .gz.\n");
else if (status != eslOK) cmdline_failure(argv[0], "Open failed, code %d.\n", status);
if(do_fetching)
{
status = esl_sqfile_OpenSSI(sqfp, NULL);
if (status == eslEFORMAT) cmdline_failure(argv[0], "SSI index is in incorrect format\n");
else if (status == eslERANGE) cmdline_failure(argv[0], "SSI index is in 64-bit format and we can't read it\n");
else if (status != eslOK) cmdline_failure(argv[0], "Failed to open SSI index\n");
}
/* Open the <maskfile> */
if (esl_fileparser_Open(maskfile, NULL, &maskefp) != eslOK)
cmdline_failure(argv[0], "Failed to open mask coordinate file %s\n", maskfile);
esl_fileparser_SetCommentChar(maskefp, '#');
/* Open the output file, if any */
if (esl_opt_GetString(go, "-o") != NULL)
{
if ((ofp = fopen(esl_opt_GetString(go, "-o"), "w")) == NULL)
cmdline_failure(argv[0], "Failed to open output file %s\n", esl_opt_GetString(go, "-o"));
}
else ofp = stdout;
/****************************************************************************
* Main loop over lines in <maskfile>
****************************************************************************/
/* Read one data line at a time from the <maskfile>;
* parse into data fields <seqname> <start> <end>
*/
while (esl_fileparser_NextLine(maskefp) == eslOK)
{
/* First field is sequence name */
if (esl_fileparser_GetTokenOnLine(maskefp, &source, NULL) != eslOK)
esl_fatal("Failed to read source seq name on line %d of file %s\n", maskefp->linenumber, maskfile);
/* Get the sequence */
if (do_fetching)
{ /* If the <seqfile> is SSI indexed, try to reposition it and read <source> seq by random access */
status = esl_sqio_Fetch(sqfp, source, sq);
if (status == eslENOTFOUND) esl_fatal("seq %s not found in SSI index for file %s\n", source, sqfp->filename);
else if (status == eslEINVAL) esl_fatal("No SSI index or can't reposition in file %s\n", sqfp->filename);
else if (status == eslEFORMAT) esl_fatal("Parse failed:\n%s\n", esl_sqfile_GetErrorBuf(sqfp));
else if (status != eslOK) esl_fatal("Unexpected failure in fetching %s from file %s\n", source, sqfp->filename);
}
else
{ /* else, assume we're reading sequentially; <sqfile> and <maskfile> have seqs in same order */
status = esl_sqio_Read(sqfp, sq);
if (status == eslEOF) esl_fatal("File %s ended prematurely; didn't find %s\n", sqfp->filename, source);
else if (status == eslEFORMAT) esl_fatal("Parse failed:\n%s\n", esl_sqfile_GetErrorBuf(sqfp));
else if (status != eslOK) esl_fatal("Unexpected error reading sequence file %s\n", sqfp->filename);
if ((strcmp(sq->name, source) != 0) && (strcmp(sq->acc, source) != 0))
esl_fatal("Sequences in <sqfile> and <maskfile> aren't in same order; try -R");
}
/* If we're masking by lowercase, first make sure everything's uppercase */
if (do_lowercase)
for (pos = 0; pos < sq->n; pos++)
if (isalpha(sq->seq[pos]))
sq->seq[pos] = toupper(sq->seq[pos]);
/* Next two fields are <start>, <end> for the masking */
/* possible future extension: wrap loop around this, enable multiple masked regions */
if (esl_fileparser_GetTokenOnLine(maskefp, &p1, NULL) != eslOK)
esl_fatal("Failed to read start coord on line %d of file %s\n", maskefp->linenumber, maskfile);
start = strtoll(p1, &p2, 0) - 1;
if (esl_fileparser_GetTokenOnLine(maskefp, &p2, NULL) != eslOK)
esl_fatal("Failed to read end coord on line %d of file %s\n", maskefp->linenumber, maskfile);
end = strtoll(p2, &p1, 0) - 1;
/* Do the masking */
if (esl_opt_GetBoolean(go, "-r")) /* Reverse masking */
{ /* leave start..end unmasked; mask prefix 0..start-1, end+1..L-1 */
i = 0;
j = ESL_MIN(sq->n-1, start - 1 + overmask);
for (pos = i; pos <= j; pos++)
if (isalpha(sq->seq[pos]))
sq->seq[pos] = (do_lowercase ? tolower(sq->seq[pos]) : maskchar);
i = ESL_MAX(0, end + 1 - overmask);
j = sq->n-1;
for (pos = i; pos <= j; pos++)
if (isalpha(sq->seq[pos]))
sq->seq[pos] = (do_lowercase ? tolower(sq->seq[pos]) : maskchar);
}
else
{ /* normal: mask start..end */
i = ESL_MAX(0, start - overmask);
j = ESL_MIN(sq->n-1, end + overmask);
for (pos = i; pos <= j; pos++)
if (isalpha(sq->seq[pos]))
sq->seq[pos] = (do_lowercase ? tolower(sq->seq[pos]) : maskchar);
}
esl_sqio_Write(ofp, sq, outfmt, FALSE);
esl_sq_Reuse(sq);
}
esl_sq_Destroy(sq);
esl_fileparser_Close(maskefp);
esl_sqfile_Close(sqfp);
esl_getopts_Destroy(go);
if (ofp != stdout) fclose(ofp);
return 0;
}
/* seq_shuffling()
* SRE, Tue Jan 22 08:35:51 2008 [Market Street Cafe, Leesburg]
*
* Shuffling of input sequences.
*
* Fixed-length (L>0) vs. full-length (L=0) modes handled differently.
* In fixed-length mode:
* <shuff->seq> only needs to be allocated once, for L
* <targ> is an allocated copy of a random subseq of length L
* sequences < L residues long can't be shuffled
* In full-length mode:
* <shuff->seq> is grown to length <sq->n> for each input seq
* <targ> just points to <sq->seq>
*/
static int
seq_shuffling(ESL_GETOPTS *go, ESL_RANDOMNESS *r, FILE *ofp, int outfmt)
{
char *seqfile = esl_opt_GetArg(go, 1);
int infmt = eslSQFILE_UNKNOWN;
ESL_SQFILE *sqfp = NULL;
ESL_SQ *sq = esl_sq_Create();
ESL_SQ *shuff = esl_sq_Create();
char *targ = NULL;
int N = esl_opt_GetInteger(go, "-N");
int L = esl_opt_GetInteger(go, "-L"); /* L>0 means select random fixed-len subseqs */
int kmers = 0;
int i;
int status;
if (esl_opt_GetString(go, "--informat") != NULL) {
infmt = esl_sqio_EncodeFormat(esl_opt_GetString(go, "--informat"));
if (infmt == eslSQFILE_UNKNOWN) esl_fatal("%s is not a valid input sequence file format for --informat");
}
if (esl_opt_IsOn(go, "-k")) kmers = esl_opt_GetInteger(go, "-k");
status = esl_sqfile_Open(seqfile, infmt, NULL, &sqfp);
if (status == eslENOTFOUND) esl_fatal("No such file %s", seqfile);
else if (status == eslEFORMAT) esl_fatal("Format of seqfile %s unrecognized.", seqfile);
else if (status == eslEINVAL) esl_fatal("Can't autodetect stdin or .gz.");
else if (status != eslOK) esl_fatal("Open failed, code %d.", status);
if (L>0) {
esl_sq_GrowTo(shuff, L);
shuff->n = L;
ESL_ALLOC(targ, sizeof(char) * (L+1));
}
while ((status = esl_sqio_Read(sqfp, sq)) == eslOK)
{
if (L == 0) { /* shuffling entire sequence */
esl_sq_GrowTo(shuff, sq->n); /* make sure shuff can hold sq */
shuff->n = sq->n;
targ = sq->seq;
} else {
if (sq->n < L) continue; /* reject seqs < L long */
}
for (i = 0; i < N; i++)
{
if (L > 0) { /* fixed-len mode: copy a random subseq */
int pos = esl_rnd_Roll(r, sq->n - L + 1);
strncpy(targ, sq->seq + pos, L);
targ[L] = '\0';
}
/* Do the requested kind of shuffling */
if (esl_opt_GetBoolean(go, "-m")) esl_rsq_CShuffle (r, targ, shuff->seq); /* monoresidue shuffling */
else if (esl_opt_GetBoolean(go, "-d")) esl_rsq_CShuffleDP (r, targ, shuff->seq); /* diresidue shuffling */
else if (esl_opt_IsOn (go, "-k")) esl_rsq_CShuffleKmers(r, targ, kmers, shuff->seq); /* diresidue shuffling */
else if (esl_opt_GetBoolean(go, "-0")) esl_rsq_CMarkov0 (r, targ, shuff->seq); /* 0th order Markov */
else if (esl_opt_GetBoolean(go, "-1")) esl_rsq_CMarkov1 (r, targ, shuff->seq); /* 1st order Markov */
else if (esl_opt_GetBoolean(go, "-r")) esl_rsq_CReverse ( targ, shuff->seq); /* reverse */
else if (esl_opt_IsOn (go, "-w")) { /* regionally shuffle */
int W= esl_opt_GetInteger(go, "-w"); esl_rsq_CShuffleWindows(r, targ, W, shuff->seq);
}
/* Set the name of the shuffled sequence */
if (N > 1) esl_sq_FormatName(shuff, "%s-shuffled-%d", sq->name, i);
else esl_sq_FormatName(shuff, "%s-shuffled", sq->name);
/* Output the resulting sequence */
esl_sqio_Write(ofp, shuff, outfmt, FALSE);
/* don't need to reuse the shuffled sequence: we will use exactly the same memory */
}
esl_sq_Reuse(sq);
}
if (status == eslEFORMAT) esl_fatal("Parse failed (sequence file %s):\n%s\n",
sqfp->filename, esl_sqfile_GetErrorBuf(sqfp));
else if (status != eslEOF) esl_fatal("Unexpected error %d reading sequence file %s",
status, sqfp->filename);
if (L>0) free(targ);
esl_sq_Destroy(shuff);
esl_sq_Destroy(sq);
esl_sqfile_Close(sqfp);
return eslOK;
ERROR:
if (targ != NULL) free(targ);
esl_sq_Destroy(shuff);
esl_sq_Destroy(sq);
esl_sqfile_Close(sqfp);
return status;
}
/* serial_master()
* The serial version of hmmsearch.
* For each query HMM in <hmmfile> search the database for hits.
*
* A master can only return if it's successful. All errors are handled
* immediately and fatally with p7_Fail(). We also use the
* ESL_EXCEPTION and ERROR: mechanisms, but only because we know we're
* using a fatal exception handler.
*/
static int
serial_master(ESL_GETOPTS *go, struct cfg_s *cfg)
{
FILE *ofp = stdout; /* results output file (-o) */
P7_HMMFILE *hfp = NULL; /* open input HMM file */
ESL_SQFILE *dbfp = NULL; /* open input sequence file */
P7_HMM *hmm = NULL; /* one HMM query */
ESL_ALPHABET *abc = NULL; /* digital alphabet */
int dbfmt = eslSQFILE_UNKNOWN; /* format code for sequence database file */
ESL_STOPWATCH *w;
int textw = 0;
int nquery = 0;
int status = eslOK;
int hstatus = eslOK;
int sstatus = eslOK;
int i;
int ncpus = 0;
int infocnt = 0;
WORKER_INFO *info = NULL;
char errbuf[eslERRBUFSIZE];
w = esl_stopwatch_Create();
if (esl_opt_GetBoolean(go, "--notextw")) textw = 0;
else textw = esl_opt_GetInteger(go, "--textw");
if (esl_opt_IsOn(go, "--tformat")) {
dbfmt = esl_sqio_EncodeFormat(esl_opt_GetString(go, "--tformat"));
if (dbfmt == eslSQFILE_UNKNOWN) p7_Fail("%s is not a recognized sequence database file format\n", esl_opt_GetString(go, "--tformat"));
}
/* Open the target sequence database */
status = esl_sqfile_Open(cfg->dbfile, dbfmt, p7_SEQDBENV, &dbfp);
if (status == eslENOTFOUND) p7_Fail("Failed to open sequence file %s for reading\n", cfg->dbfile);
else if (status == eslEFORMAT) p7_Fail("Sequence file %s is empty or misformatted\n", cfg->dbfile);
else if (status == eslEINVAL) p7_Fail("Can't autodetect format of a stdin or .gz seqfile");
else if (status != eslOK) p7_Fail("Unexpected error %d opening sequence file %s\n", status, cfg->dbfile);
if (esl_opt_IsUsed(go, "--restrictdb_stkey") || esl_opt_IsUsed(go, "--restrictdb_n")) {
if (esl_opt_IsUsed(go, "--ssifile"))
esl_sqfile_OpenSSI(dbfp, esl_opt_GetString(go, "--ssifile"));
else
esl_sqfile_OpenSSI(dbfp, NULL);
}
/* Open the query profile HMM file */
status = p7_hmmfile_OpenE(cfg->hmmfile, NULL, &hfp, errbuf);
if (status == eslENOTFOUND) p7_Fail("File existence/permissions problem in trying to open HMM file %s.\n%s\n", cfg->hmmfile, errbuf);
else if (status == eslEFORMAT) p7_Fail("File format problem in trying to open HMM file %s.\n%s\n", cfg->hmmfile, errbuf);
else if (status != eslOK) p7_Fail("Unexpected error %d in opening HMM file %s.\n%s\n", status, cfg->hmmfile, errbuf);
/* Open the results output files */
if (esl_opt_IsOn(go, "-o")) { if ((ofp = fopen(esl_opt_GetString(go, "-o"), "w+")) == NULL) p7_Fail("Failed to open output file %s for writing\n", esl_opt_GetString(go, "-o")); }
infocnt = 1;
ESL_ALLOC(info, sizeof(*info) * infocnt);
/* <abc> is not known 'til first HMM is read. */
hstatus = p7_hmmfile_Read(hfp, &abc, &hmm);
if (hstatus == eslOK)
{
/* One-time initializations after alphabet <abc> becomes known */
// output_header(ofp, go, cfg->hmmfile, cfg->dbfile);
// dbfp->abc = abc; //ReadBlock requires knowledge of the alphabet to decide how best to read blocks
// for (i = 0; i < infocnt; ++i)
// {
// info[i].bg = p7_bg_Create(abc);
// }
}
/* Outer loop: over each query HMM in <hmmfile>. */
// while (hstatus == eslOK)
// {
P7_PROFILE *gm = NULL;
P7_OPROFILE *om = NULL; /* optimized query profile */
nquery++;
esl_stopwatch_Start(w);
/* seqfile may need to be rewound (multiquery mode) */
if (nquery > 1)
{
if (! esl_sqfile_IsRewindable(dbfp))
esl_fatal("Target sequence file %s isn't rewindable; can't search it with multiple queries", cfg->dbfile);
if (! esl_opt_IsUsed(go, "--restrictdb_stkey") )
esl_sqfile_Position(dbfp, 0); //only re-set current position to 0 if we're not planning to set it in a moment
}
if ( cfg->firstseq_key != NULL ) { //it's tempting to want to do this once and capture the offset position for future passes, but ncbi files make this non-trivial, so this keeps it general
sstatus = esl_sqfile_PositionByKey(dbfp, cfg->firstseq_key);
if (sstatus != eslOK)
p7_Fail("Failure setting restrictdb_stkey to %d\n", cfg->firstseq_key);
}
// if (fprintf(ofp, "Query: %s [M=%d]\n", hmm->name, hmm->M) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
// if (hmm->acc) { if (fprintf(ofp, "Accession: %s\n", hmm->acc) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed"); }
// if (hmm->desc) { if (fprintf(ofp, "Description: %s\n", hmm->desc) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed"); }
/* Convert to an optimized model */
gm = p7_profile_Create (hmm->M, abc);
om = p7_oprofile_Create(hmm->M, abc);
// p7_ProfileConfig(hmm, info->bg, gm, 100, p7_LOCAL); /* 100 is a dummy length for now; and MSVFilter requires local mode */
p7_oprofile_Convert(gm, om); /* <om> is now p7_LOCAL, multihit */
for (i = 0; i < infocnt; ++i)
{
/* Create processing pipeline and hit list */
info[i].th = p7_tophits_Create();
info[i].om = p7_oprofile_Clone(om);
info[i].pli = p7_pipeline_Create(go, om->M, 100, FALSE, p7_SEARCH_SEQS); /* L_hint = 100 is just a dummy for now */
P7_PIPELINE *pli = info[i].pli;
pli->nmodels++;
pli->nnodes += info[i].om->M;
// if (pli->Z_setby == p7_ZSETBY_NTARGETS && pli->mode == p7_SCAN_MODELS) pli->Z = pli->nmodels;
// if (pli->do_biasfilter) p7_bg_SetFilter(info[i].bg, info[i].om->M, info[i].om->compo);
// if (pli->mode == p7_SEARCH_SEQS)
// status = p7_pli_NewModelThresholds(pli, info[i].om);
pli->W = info[i].om->max_length;
}
sstatus = serial_loop(info, dbfp, cfg->n_targetseq, ofp);
switch(sstatus)
{
case eslEFORMAT:
esl_fatal("Parse failed (sequence file %s):\n%s\n",
dbfp->filename, esl_sqfile_GetErrorBuf(dbfp));
break;
case eslEOF:
/* do nothing */
break;
default:
esl_fatal("Unexpected error %d reading sequence file %s", sstatus, dbfp->filename);
}
/* merge the results of the search results */
for (i = 1; i < infocnt; ++i)
{
p7_tophits_Merge(info[0].th, info[i].th);
p7_pipeline_Merge(info[0].pli, info[i].pli);
p7_pipeline_Destroy(info[i].pli);
p7_tophits_Destroy(info[i].th);
p7_oprofile_Destroy(info[i].om);
}
/* Print the results. */
p7_tophits_SortBySortkey(info->th);
p7_tophits_Threshold(info->th, info->pli);
// p7_tophits_Targets(ofp, info->th, info->pli, textw); if (fprintf(ofp, "\n\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
// p7_tophits_Domains(ofp, info->th, info->pli, textw); if (fprintf(ofp, "\n\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
esl_stopwatch_Stop(w);
// p7_pli_Statistics(ofp, info->pli, w);
// if (fprintf(ofp, "//\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
p7_pipeline_Destroy(info->pli);
p7_tophits_Destroy(info->th);
p7_oprofile_Destroy(info->om);
p7_oprofile_Destroy(om);
p7_profile_Destroy(gm);
p7_hmm_Destroy(hmm);
// hstatus = p7_hmmfile_Read(hfp, &abc, &hmm);
// } /* end outer loop over query HMMs */
switch(hstatus) {
case eslEOD: p7_Fail("read failed, HMM file %s may be truncated?", cfg->hmmfile); break;
case eslEFORMAT: p7_Fail("bad file format in HMM file %s", cfg->hmmfile); break;
case eslEINCOMPAT: p7_Fail("HMM file %s contains different alphabets", cfg->hmmfile); break;
case eslEOF:
case eslOK: /* do nothing. EOF is what we want. */ break;
default: p7_Fail("Unexpected error (%d) in reading HMMs from %s", hstatus, cfg->hmmfile);
}
/* Terminate outputs... any last words?
*/
/* Cleanup - prepare for exit
*/
// for (i = 0; i < infocnt; ++i)
// p7_bg_Destroy(info[i].bg);
free(info);
p7_hmmfile_Close(hfp);
esl_sqfile_Close(dbfp);
esl_alphabet_Destroy(abc);
esl_stopwatch_Destroy(w);
if (ofp != stdout) fclose(ofp);
printf("44HHHH \n");
return eslOK;
ERROR:
return eslFAIL;
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = NULL; /* application configuration */
char *hmmfile = NULL; /* HMM file name */
char *seqfile = NULL; /* sequence file name */
char *mapfile = NULL; /* optional mapped MSA file name */
int infmt = eslSQFILE_UNKNOWN;
int outfmt = eslMSAFILE_STOCKHOLM;
P7_HMMFILE *hfp = NULL; /* open HMM file */
ESL_SQFILE *sqfp = NULL; /* open sequence file */
char *outfile = NULL; /* output filename */
FILE *ofp = stdout; /* output stream */
ESL_SQ **sq = NULL; /* array of sequences */
void *p = NULL; /* tmp ptr for reallocation */
int nseq = 0; /* # of sequences in <seqfile> */
int mapseq = 0; /* # of sequences in mapped MSA */
int totseq = 0; /* # of seqs in all sources */
ESL_ALPHABET *abc = NULL; /* alphabet (set from the HMM file)*/
P7_HMM *hmm = NULL;
P7_TRACE **tr = NULL; /* array of tracebacks */
ESL_MSA *msa = NULL; /* resulting multiple alignment */
int msaopts = 0; /* flags to p7_tracealign_Seqs() */
int idx; /* counter over seqs, traces */
int status; /* easel/hmmer return code */
char errbuf[eslERRBUFSIZE];
/* Parse the command line
*/
go = esl_getopts_Create(options);
if (esl_opt_ProcessCmdline(go, argc, argv) != eslOK) cmdline_failure(argv[0], "Failed to parse command line: %s\n", go->errbuf);
if (esl_opt_VerifyConfig(go) != eslOK) cmdline_failure(argv[0], "Error in configuration: %s\n", go->errbuf);
if (esl_opt_GetBoolean(go, "-h") ) cmdline_help (argv[0], go);
if (esl_opt_ArgNumber(go) != 2) cmdline_failure(argv[0], "Incorrect number of command line arguments.\n");
hmmfile = esl_opt_GetArg(go, 1);
seqfile = esl_opt_GetArg(go, 2);
if (strcmp(hmmfile, "-") == 0 && strcmp(seqfile, "-") == 0)
cmdline_failure(argv[0], "Either <hmmfile> or <seqfile> may be '-' (to read from stdin), but not both.\n");
msaopts |= p7_ALL_CONSENSUS_COLS; /* default as of 3.1 */
if (esl_opt_GetBoolean(go, "--trim")) msaopts |= p7_TRIM;
/* If caller declared an input format, decode it
*/
if (esl_opt_IsOn(go, "--informat")) {
infmt = esl_sqio_EncodeFormat(esl_opt_GetString(go, "--informat"));
if (infmt == eslSQFILE_UNKNOWN) cmdline_failure(argv[0], "%s is not a recognized input sequence file format\n", esl_opt_GetString(go, "--informat"));
}
/* Determine output alignment file format */
outfmt = eslx_msafile_EncodeFormat(esl_opt_GetString(go, "--outformat"));
if (outfmt == eslMSAFILE_UNKNOWN) cmdline_failure(argv[0], "%s is not a recognized output MSA file format\n", esl_opt_GetString(go, "--outformat"));
/* Open output stream */
if ( (outfile = esl_opt_GetString(go, "-o")) != NULL)
{
if ((ofp = fopen(outfile, "w")) == NULL)
cmdline_failure(argv[0], "failed to open -o output file %s for writing\n", outfile);
}
/* If caller forced an alphabet on us, create the one the caller wants
*/
if (esl_opt_GetBoolean(go, "--amino")) abc = esl_alphabet_Create(eslAMINO);
else if (esl_opt_GetBoolean(go, "--dna")) abc = esl_alphabet_Create(eslDNA);
else if (esl_opt_GetBoolean(go, "--rna")) abc = esl_alphabet_Create(eslRNA);
/* Read one HMM, and make sure there's only one.
*/
status = p7_hmmfile_OpenE(hmmfile, NULL, &hfp, errbuf);
if (status == eslENOTFOUND) p7_Fail("File existence/permissions problem in trying to open HMM file %s.\n%s\n", hmmfile, errbuf);
else if (status == eslEFORMAT) p7_Fail("File format problem in trying to open HMM file %s.\n%s\n", hmmfile, errbuf);
else if (status != eslOK) p7_Fail("Unexpected error %d in opening HMM file %s.\n%s\n", status, hmmfile, errbuf);
status = p7_hmmfile_Read(hfp, &abc, &hmm);
if (status == eslEFORMAT) p7_Fail("Bad file format in HMM file %s:\n%s\n", hfp->fname, hfp->errbuf);
else if (status == eslEINCOMPAT) p7_Fail("HMM in %s is not in the expected %s alphabet\n", hfp->fname, esl_abc_DecodeType(abc->type));
else if (status == eslEOF) p7_Fail("Empty HMM file %s? No HMM data found.\n", hfp->fname);
else if (status != eslOK) p7_Fail("Unexpected error in reading HMMs from %s\n", hfp->fname);
status = p7_hmmfile_Read(hfp, &abc, NULL);
if (status != eslEOF) p7_Fail("HMM file %s does not contain just one HMM\n", hfp->fname);
p7_hmmfile_Close(hfp);
/* We're going to build up two arrays: sequences and traces.
* If --mapali option is chosen, the first set of sequences/traces is from the provided alignment
*/
if ( (mapfile = esl_opt_GetString(go, "--mapali")) != NULL)
{
map_alignment(mapfile, hmm, &sq, &tr, &mapseq);
}
totseq = mapseq;
/* Read digital sequences into an array (possibly concat'ed onto mapped seqs)
*/
status = esl_sqfile_OpenDigital(abc, seqfile, infmt, NULL, &sqfp);
if (status == eslENOTFOUND) p7_Fail("Failed to open sequence file %s for reading\n", seqfile);
else if (status == eslEFORMAT) p7_Fail("Sequence file %s is empty or misformatted\n", seqfile);
else if (status != eslOK) p7_Fail("Unexpected error %d opening sequence file %s\n", status, seqfile);
ESL_RALLOC(sq, p, sizeof(ESL_SQ *) * (totseq + 1));
sq[totseq] = esl_sq_CreateDigital(abc);
nseq = 0;
while ((status = esl_sqio_Read(sqfp, sq[totseq+nseq])) == eslOK)
{
nseq++;
ESL_RALLOC(sq, p, sizeof(ESL_SQ *) * (totseq+nseq+1));
sq[totseq+nseq] = esl_sq_CreateDigital(abc);
}
if (status == eslEFORMAT) esl_fatal("Parse failed (sequence file %s):\n%s\n",
sqfp->filename, esl_sqfile_GetErrorBuf(sqfp));
else if (status != eslEOF) esl_fatal("Unexpected error %d reading sequence file %s", status, sqfp->filename);
esl_sqfile_Close(sqfp);
totseq += nseq;
/* Remaining initializations, including trace array allocation
*/
ESL_RALLOC(tr, p, sizeof(P7_TRACE *) * totseq);
for (idx = mapseq; idx < totseq; idx++)
tr[idx] = p7_trace_CreateWithPP();
p7_tracealign_computeTraces(hmm, sq, mapseq, totseq - mapseq, tr);
p7_tracealign_Seqs(sq, tr, totseq, hmm->M, msaopts, hmm, &msa);
eslx_msafile_Write(ofp, msa, outfmt);
for (idx = 0; idx <= totseq; idx++) esl_sq_Destroy(sq[idx]); /* including sq[nseq] because we overallocated */
for (idx = 0; idx < totseq; idx++) p7_trace_Destroy(tr[idx]);
free(sq);
free(tr);
esl_msa_Destroy(msa);
p7_hmm_Destroy(hmm);
if (ofp != stdout) fclose(ofp);
esl_alphabet_Destroy(abc);
esl_getopts_Destroy(go);
return eslOK;
ERROR:
return status;
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = NULL; /* application configuration */
char *seqfile = NULL; /* sequence file name */
int infmt = eslSQFILE_UNKNOWN; /* format code for seqfile */
ESL_SQFILE *sqfp = NULL; /* open sequence file */
FILE *ofp = NULL; /* output stream for sequences */
int status; /* easel return code */
/***********************************************
* Parse command line
***********************************************/
go = esl_getopts_Create(options);
if (esl_opt_ProcessCmdline(go, argc, argv) != eslOK) cmdline_failure(argv[0], "Failed to parse command line: %s\n", go->errbuf);
if (esl_opt_VerifyConfig(go) != eslOK) cmdline_failure(argv[0], "Error in configuration: %s\n", go->errbuf);
if (esl_opt_GetBoolean(go, "-h") ) cmdline_help (argv[0], go);
if (esl_opt_ArgNumber(go) < 1) cmdline_failure(argv[0], "Incorrect number of command line arguments.\n");
/* Open the sequence file */
seqfile = esl_opt_GetArg(go, 1);
if (esl_opt_GetString(go, "--informat") != NULL) {
infmt = esl_sqio_EncodeFormat(esl_opt_GetString(go, "--informat"));
if (infmt == eslSQFILE_UNKNOWN) esl_fatal("%s is not a valid input sequence file format for --informat", esl_opt_GetString(go, "--informat"));
}
status = esl_sqfile_Open(seqfile, infmt, NULL, &sqfp);
if (status == eslENOTFOUND) cmdline_failure(argv[0], "Sequence file %s not found.\n", seqfile);
else if (status == eslEFORMAT) cmdline_failure(argv[0], "Format of file %s unrecognized.\n", seqfile);
else if (status == eslEINVAL) cmdline_failure(argv[0], "Can't autodetect stdin or .gz.\n");
else if (status != eslOK) cmdline_failure(argv[0], "Open failed, code %d.\n", status);
/* Open the output file, if any */
if (esl_opt_GetBoolean(go, "-O"))
{
if ((ofp = fopen(esl_opt_GetArg(go, 2), "w")) == NULL)
cmdline_failure(argv[0], "Failed to open output file %s\n", esl_opt_GetArg(go, 2));
}
else if (esl_opt_GetString(go, "-o") != NULL)
{
if ((ofp = fopen(esl_opt_GetString(go, "-o"), "w")) == NULL)
cmdline_failure(argv[0], "Failed to open output file %s\n", esl_opt_GetString(go, "-o"));
}
else ofp = stdout;
/* Indexing mode */
if (esl_opt_GetBoolean(go, "--index"))
{
if (esl_opt_ArgNumber(go) != 1) cmdline_failure(argv[0], "Incorrect number of command line arguments.\n");
if (sqfp->data.ascii.do_gzip) cmdline_failure(argv[0], "Can't index a .gz compressed file");
if (sqfp->data.ascii.do_stdin) cmdline_failure(argv[0], "Can't index a standard input pipe");
create_ssi_index(go, sqfp);
}
/* List retrieval mode */
else if (esl_opt_GetBoolean(go, "-f"))
{
if (esl_opt_ArgNumber(go) != 2) cmdline_failure(argv[0], "Incorrect number of command line arguments.\n");
/* Open the SSI index for retrieval */
if (! sqfp->data.ascii.do_gzip && ! sqfp->data.ascii.do_stdin && ! esl_sqio_IsAlignment(sqfp->format))
{
status = esl_sqfile_OpenSSI(sqfp, NULL);
if (status == eslEFORMAT) cmdline_failure(argv[0], "SSI index is in incorrect format\n");
else if (status == eslERANGE) cmdline_failure(argv[0], "SSI index is in 64-bit format and we can't read it\n");
else if (status != eslOK) cmdline_failure(argv[0], "Failed to open SSI index\n");
}
if (esl_opt_GetBoolean(go, "-C")) multifetch_subseq(go, ofp, esl_opt_GetArg(go, 2), sqfp);
else multifetch (go, ofp, esl_opt_GetArg(go, 2), sqfp);
}
/* Single sequence retrieval mode */
else
{
if (esl_opt_ArgNumber(go) != 2) cmdline_failure(argv[0], "Incorrect number of command line arguments.\n");
char *key = esl_opt_GetArg(go, 2);
char *cstring = esl_opt_GetString(go, "-c");
char *newname = esl_opt_GetString(go, "-n");
/* Open the SSI index for retrieval */
if (! sqfp->data.ascii.do_gzip && ! sqfp->data.ascii.do_stdin && ! esl_sqio_IsAlignment(sqfp->format))
{
status = esl_sqfile_OpenSSI(sqfp, NULL);
if (status == eslEFORMAT) cmdline_failure(argv[0], "SSI index is in incorrect format\n");
else if (status == eslERANGE) cmdline_failure(argv[0], "SSI index is in 64-bit format and we can't read it\n");
else if (status != eslOK) cmdline_failure(argv[0], "Failed to open SSI index\n");
}
/* -c: subsequence retrieval; else full sequence retrieval */
if (cstring != NULL)
{
uint32_t start, end;
status = esl_regexp_ParseCoordString(cstring, &start, &end);
if (status == eslESYNTAX) esl_fatal("-c takes arg of subseq coords <from>..<to>; %s not recognized", cstring);
if (status == eslFAIL) esl_fatal("Failed to find <from> or <to> coord in %s", cstring);
onefetch_subseq(go, ofp, sqfp, newname, key, start, end);
if (ofp != stdout) printf("\n\nRetrieved subsequence %s/%d-%d.\n", key, start, end);
}
else
{
onefetch(go, ofp, esl_opt_GetArg(go, 2), sqfp);
if (ofp != stdout) printf("\n\nRetrieved sequence %s.\n", esl_opt_GetArg(go, 2));
}
}
esl_sqfile_Close(sqfp);
esl_getopts_Destroy(go);
return 0;
}
/* Function: main()
* Synopsis: break input sequence set into chunks, for each one building the
* Burrows-Wheeler transform and corresponding FM-index. Maintain requisite
* meta data.
* Notes: Currently depends on the divsufsort-lite code of Yuta Mori, though this
* could easily be replaced.
*/
int
main(int argc, char **argv)
{
char tmp_filename[16] = "fmtmpXXXXXX";
FILE *fptmp = NULL;
FILE *fp = NULL;
uint8_t *T = NULL;
uint8_t *BWT = NULL;
int *SA = NULL; //what I write will be 32-bit ints, but I need to keep this as int so it'll work with libdivsufsort
uint32_t *SAsamp = NULL;
uint32_t *occCnts_sb = NULL; // same indexing as above
uint32_t *cnts_sb = NULL;
uint16_t *occCnts_b = NULL; // this is logically a 2D array, but will be indexed as occ_cnts[alph_size*index + char] (instead of occ_cnts[index][char])
uint16_t *cnts_b = NULL;
FM_METADATA *meta = NULL;
clock_t t1, t2;
struct tms ts1, ts2;
long i,j,c;
int status = eslOK;
int chars_per_byte;
int num_freq_cnts_sb ;
int num_freq_cnts_b ;
int num_SA_samples ;
int infmt = eslSQFILE_UNKNOWN;
int alphatype = eslUNKNOWN;
int alphaguess =eslUNKNOWN;
ESL_ALPHABET *abc = NULL;
ESL_SQ *sq = NULL;
ESL_SQFILE *sqfp = NULL;
ESL_SQ *tmpsq = NULL;
ESL_SQ_BLOCK *block = NULL;
char *fname_in = NULL;
char *fname_out= NULL;
int block_size = 50000000;
int sq_cnt = 0;
int use_tmpsq = 0;
uint64_t block_length;
uint64_t total_char_count = 0;
int max_block_size;
int numblocks = 0;
uint32_t numseqs = 0;
int allocedseqs = 1000;
uint32_t seq_offset = 0;
uint32_t ambig_offset = 0;
uint32_t overlap = 0;
uint16_t seq_cnt;
uint16_t ambig_cnt;
uint32_t prev_numseqs = 0;
int compressed_bytes;
uint32_t term_loc;
ESL_GETOPTS *go = NULL; /* command line processing */
uint8_t ambig_repl = 0;
int in_ambig_run = 0;
FM_AMBIGLIST ambig_list;
ESL_ALLOC (meta, sizeof(FM_METADATA));
if (meta == NULL)
esl_fatal("unable to allocate memory to store FM meta data\n");
ESL_ALLOC (meta->ambig_list, sizeof(FM_AMBIGLIST));
if (meta->ambig_list == NULL)
esl_fatal("unable to allocate memory to store FM ambiguity data\n");
fm_initAmbiguityList(meta->ambig_list);
meta->alph_type = fm_DNA;
meta->freq_SA = 8;
meta->freq_cnt_b = 256;
meta->freq_cnt_sb = pow(2,16); //65536 - that's the # values in a short
meta->seq_count = 0;
ESL_ALLOC (meta->seq_data, allocedseqs * sizeof(FM_SEQDATA));
if (meta->seq_data == NULL )
esl_fatal("unable to allocate memory to store FM sequence data\n");
process_commandline(argc, argv, &go, &fname_in, &fname_out);
if (esl_opt_IsOn(go, "--bin_length")) meta->freq_cnt_b = esl_opt_GetInteger(go, "--bin_length");
if ( meta->freq_cnt_b < 32 || meta->freq_cnt_b >4096 || (meta->freq_cnt_b & (meta->freq_cnt_b - 1)) ) // test power of 2
esl_fatal("bin_length must be a power of 2, at least 128, and at most 4096\n");
if (esl_opt_IsOn(go, "--sa_freq")) meta->freq_SA = esl_opt_GetInteger(go, "--sa_freq");
if ( (meta->freq_SA & (meta->freq_SA - 1)) ) // test power of 2
esl_fatal ("SA_freq must be a power of 2\n");
if (esl_opt_IsOn(go, "--block_size")) block_size = 1000000 * esl_opt_GetInteger(go, "--block_size");
if ( block_size <=0 )
esl_fatal ("block_size must be a positive number\n");
//start timer
t1 = times(&ts1);
output_header(stdout, go, fname_in, fname_out);
if (esl_opt_GetString(go, "--informat") != NULL) {
infmt = esl_sqio_EncodeFormat(esl_opt_GetString(go, "--informat"));
if (infmt == eslSQFILE_UNKNOWN) esl_fatal("%s is not a valid input sequence file format for --informat");
}
status = esl_sqfile_Open(fname_in, infmt, NULL, &sqfp);
if (status == eslENOTFOUND) esl_fatal("No such file %s", fname_in);
else if (status == eslEFORMAT) esl_fatal("Format of seqfile %s unrecognized.", fname_in);
else if (status != eslOK) esl_fatal("Open failed, code %d.", status);
meta->fwd_only = 0;
if (esl_opt_IsUsed(go, "--alph")) {
meta->alph = esl_opt_GetString(go, "--alph") ;
if ( esl_strcmp(meta->alph, "dna")==0 || esl_strcmp(meta->alph, "rna")==0) {
meta->alph_type = fm_DNA;
alphatype = eslDNA;
} else if (esl_strcmp(meta->alph, "dna_full")==0 || esl_strcmp(meta->alph, "rna_full")==0) {
meta->alph_type = fm_DNA_full;
alphatype = eslDNA;
} else if (esl_strcmp(meta->alph, "amino")==0) {
meta->alph_type = fm_AMINO;
alphatype = eslAMINO;
meta->fwd_only = 1;
} else {
esl_fatal("Unknown alphabet type. Try 'dna', 'dna_full', or 'amino'\n%s", "");
}
} else {
esl_sqfile_GuessAlphabet(sqfp, &alphaguess);
if (alphaguess == eslDNA || alphaguess == eslRNA) {
meta->alph_type = fm_DNA;
alphatype = eslDNA;
} else if (alphaguess == eslAMINO) {
meta->alph_type = fm_AMINO;
alphatype = eslAMINO;
meta->fwd_only = 1;
} else {
esl_fatal("Unknown alphabet type. Try 'dna', 'dna_full', or 'amino'\n%s", "");
}
}
if (esl_opt_IsOn(go, "--fwd_only") )
meta->fwd_only = 1;
meta->alph = NULL;
//getInverseAlphabet
fm_alphabetCreate(meta, &(meta->charBits));
chars_per_byte = 8/meta->charBits;
//shift inv_alph up one, to make space for '$' at 0
for (i=0; i<256; i++)
if ( meta->inv_alph[i] >= 0)
meta->inv_alph[i]++;
abc = esl_alphabet_Create(alphatype);
sq = esl_sq_CreateDigital(abc);
tmpsq = esl_sq_CreateDigital(abc);
esl_sqfile_SetDigital(sqfp, abc);
block = esl_sq_CreateDigitalBlock(FM_BLOCK_COUNT, abc);
block->complete = FALSE;
// max_block_size = FM_BLOCK_OVERLAP+block_size+1 + block_size*.2; // +1 for the '$'
max_block_size = FM_BLOCK_OVERLAP+block_size+1 + block_size; // temporary hack to avoid memory over-runs (see end of 1101_fmindex_benchmarking/00NOTES)
if (alphatype == fm_DNA)
fm_initAmbiguityList(&ambig_list);
/* Allocate BWT, Text, SA, and FM-index data structures, allowing storage of maximally large sequence*/
ESL_ALLOC (T, max_block_size * sizeof(uint8_t));
ESL_ALLOC (BWT, max_block_size * sizeof(uint8_t));
ESL_ALLOC (SA, max_block_size * sizeof(int));
ESL_ALLOC (SAsamp, (1+floor((double)max_block_size/meta->freq_SA) ) * sizeof(uint32_t));
ESL_ALLOC (occCnts_sb, (1+ceil((double)max_block_size/meta->freq_cnt_sb)) * meta->alph_size * sizeof(uint32_t)); // every freq_cnt_sb positions, store an array of ints
ESL_ALLOC (cnts_sb, meta->alph_size * sizeof(uint32_t));
ESL_ALLOC (occCnts_b, ( 1+ceil((double)max_block_size/meta->freq_cnt_b)) * meta->alph_size * sizeof(uint16_t)); // every freq_cnt_b positions, store an array of 8-byte ints
ESL_ALLOC (cnts_b, meta->alph_size * sizeof(uint16_t));
if((T == NULL) || (BWT == NULL) || (SA==NULL) || (SAsamp==NULL) || (BWT==NULL) || (cnts_b==NULL) || (occCnts_b==NULL) || (cnts_sb==NULL) || (occCnts_sb==NULL) ) {
esl_fatal( "%s: Cannot allocate memory.\n", argv[0]);
}
// Open a temporary file, to which FM-index data will be written
if (esl_tmpfile(tmp_filename, &fptmp) != eslOK) esl_fatal("unable to open fm-index tmpfile");
/* Main loop: */
while (status == eslOK ) {
//reset block as an empty vessel
for (i=0; i<block->count; i++)
esl_sq_Reuse(block->list + i);
if (use_tmpsq) {
esl_sq_Copy(tmpsq , block->list);
block->complete = FALSE; //this lets ReadBlock know that it needs to append to a small bit of previously-read seqeunce
block->list->C = FM_BLOCK_OVERLAP; // overload the ->C value, which ReadBlock uses to determine how much
// overlap should be retained in the ReadWindow step
} else {
block->complete = TRUE;
}
status = esl_sqio_ReadBlock(sqfp, block, block_size, -1, alphatype != eslAMINO);
if (status == eslEOF) continue;
if (status != eslOK) ESL_XEXCEPTION(status, "failure reading sequence block");
seq_offset = numseqs;
ambig_offset = meta->ambig_list->count;
if (block->complete || block->count == 0) {
use_tmpsq = FALSE;
} else {
/* The final sequence on the block was a probably-incomplete window of the active sequence.
* Grab a copy of the end for use in the next pass, to ensure we don't miss hits crossing
* the boundary between two blocks.
*/
esl_sq_Copy(block->list + (block->count - 1) , tmpsq);
use_tmpsq = TRUE;
}
block->first_seqidx = sq_cnt;
sq_cnt += block->count - (use_tmpsq ? 1 : 0);// if there's an incomplete sequence read into the block wait to count it until it's complete.
/* Read dseqs from block into text element T.
* Convert the dsq from esl-alphabet to fm-alphabet (1..k for alphabet of size k).
* (a) collapsing upper/lower case for appropriate sorting.
* (b) reserving 0 for '$', which must be lexicographically smallest
* (these will later be shifted to 0-based alphabet, once SA has been built)
*
*/
block_length = 0;
for (i=0; i<block->count; i++) {
//start a new block, with space for the name
allocateSeqdata(meta, block->list+i, numseqs, &allocedseqs);
//meta data
meta->seq_data[numseqs].target_id = block->first_seqidx + i ;
meta->seq_data[numseqs].target_start = block->list[i].start;
meta->seq_data[numseqs].fm_start = block_length;
if (block->list[i].name == NULL) meta->seq_data[numseqs].name[0] = '\0';
else strcpy(meta->seq_data[numseqs].name, block->list[i].name );
if (block->list[i].acc == NULL) meta->seq_data[numseqs].acc[0] = '\0';
else strcpy(meta->seq_data[numseqs].acc, block->list[i].acc );
if (block->list[i].source == NULL) meta->seq_data[numseqs].source[0] = '\0';
else strcpy(meta->seq_data[numseqs].source, block->list[i].source );
if (block->list[i].desc == NULL) meta->seq_data[numseqs].desc[0] = '\0';
else strcpy(meta->seq_data[numseqs].desc, block->list[i].desc );
for (j=1; j<=block->list[i].n; j++) {
c = abc->sym[block->list[i].dsq[j]];
if ( meta->alph_type == fm_DNA) {
if (meta->inv_alph[c] == -1) {
// replace ambiguity characters by rotating through A,C,G, and T.
c = meta->alph[ambig_repl];
ambig_repl = (ambig_repl+1)%4;
if (!in_ambig_run) {
fm_addAmbiguityRange(meta->ambig_list, block_length, block_length);
in_ambig_run=1;
} else {
meta->ambig_list->ranges[meta->ambig_list->count - 1].upper = block_length;
}
} else {
in_ambig_run=0;
}
} else if (meta->inv_alph[c] == -1) {
esl_fatal("requested alphabet doesn't match input text\n");
}
T[block_length] = meta->inv_alph[c];
block_length++;
if (j>block->list[i].C) total_char_count++; // add to total count, only if it's not redundant with earlier read
meta->seq_data[numseqs].length++;
}
numseqs++;
}
T[block_length] = 0; // last character 0 is effectively '$' for suffix array
block_length++;
seq_cnt = numseqs-seq_offset;
ambig_cnt = meta->ambig_list->count - ambig_offset;
//build and write FM-index for T. This will be a BWT on the reverse of the sequence, required for reverse-traversal of the BWT
buildAndWriteFMIndex(meta, seq_offset, ambig_offset, seq_cnt, ambig_cnt, (uint32_t)block->list[0].C, T, BWT, SA, SAsamp,
occCnts_sb, cnts_sb, occCnts_b, cnts_b, block_length, fptmp);
if ( ! meta->fwd_only ) {
//build and write FM-index for un-reversed T (used to find reverse hits using forward traversal of the BWT
buildAndWriteFMIndex(meta, seq_offset, ambig_offset, seq_cnt, ambig_cnt, 0, T, BWT, SA, NULL,
occCnts_sb, cnts_sb, occCnts_b, cnts_b, block_length, fptmp);
}
prev_numseqs = numseqs;
numblocks++;
}
esl_sqfile_Close(sqfp);
esl_alphabet_Destroy(abc);
esl_sq_Destroy(sq);
esl_sq_Destroy(tmpsq);
esl_sq_DestroyBlock(block);
meta->seq_count = numseqs;
meta->block_count = numblocks;
/* Finished writing the FM-index data to a temporary file. Now write
* metadata to fname_out, than append FM-index data from temp file
*/
if((fp = fopen(fname_out, "wb")) == NULL)
esl_fatal( "%s: Cannot open file `%s': ", argv[0], fname_out);
//write out meta data
if( fwrite(&(meta->fwd_only), sizeof(meta->fwd_only), 1, fp) != 1 ||
fwrite(&(meta->alph_type), sizeof(meta->alph_type), 1, fp) != 1 ||
fwrite(&(meta->alph_size), sizeof(meta->alph_size), 1, fp) != 1 ||
fwrite(&(meta->charBits), sizeof(meta->charBits), 1, fp) != 1 ||
fwrite(&(meta->freq_SA), sizeof(meta->freq_SA), 1, fp) != 1 ||
fwrite(&(meta->freq_cnt_sb), sizeof(meta->freq_cnt_sb), 1, fp) != 1 ||
fwrite(&(meta->freq_cnt_b), sizeof(meta->freq_cnt_b), 1, fp) != 1 ||
fwrite(&(meta->block_count), sizeof(meta->block_count), 1, fp) != 1 ||
fwrite(&(meta->seq_count), sizeof(meta->seq_count), 1, fp) != 1 ||
fwrite(&(meta->ambig_list->count), sizeof(meta->ambig_list->count), 1, fp) != 1 ||
fwrite(&total_char_count, sizeof(total_char_count), 1, fp) != 1
)
esl_fatal( "%s: Error writing meta data for FM index.\n", argv[0]);
for (i=0; i<meta->seq_count; i++) {
if( fwrite(&(meta->seq_data[i].target_id), sizeof(meta->seq_data[i].target_id), 1, fp) != 1 ||
fwrite(&(meta->seq_data[i].target_start), sizeof(meta->seq_data[i].target_start), 1, fp) != 1 ||
fwrite(&(meta->seq_data[i].fm_start), sizeof(meta->seq_data[i].fm_start), 1, fp) != 1 ||
fwrite(&(meta->seq_data[i].length), sizeof(meta->seq_data[i].length), 1, fp) != 1 ||
fwrite(&(meta->seq_data[i].name_length), sizeof(meta->seq_data[i].name_length), 1, fp) != 1 ||
fwrite(&(meta->seq_data[i].acc_length), sizeof(meta->seq_data[i].acc_length), 1, fp) != 1 ||
fwrite(&(meta->seq_data[i].source_length),sizeof(meta->seq_data[i].source_length), 1, fp) != 1 ||
fwrite(&(meta->seq_data[i].desc_length), sizeof(meta->seq_data[i].desc_length), 1, fp) != 1 ||
fwrite(meta->seq_data[i].name, sizeof(char), meta->seq_data[i].name_length+1 , fp) != meta->seq_data[i].name_length+1 ||
fwrite(meta->seq_data[i].acc, sizeof(char), meta->seq_data[i].acc_length+1 , fp) != meta->seq_data[i].acc_length+1 ||
fwrite(meta->seq_data[i].source, sizeof(char), meta->seq_data[i].source_length+1, fp) != meta->seq_data[i].source_length+1 ||
fwrite(meta->seq_data[i].desc, sizeof(char), meta->seq_data[i].desc_length+1 , fp) != meta->seq_data[i].desc_length+1
)
esl_fatal( "%s: Error writing meta data for FM index.\n", argv[0]);
}
for (i=0; i<meta->ambig_list->count; i++) {
if( fwrite(&(meta->ambig_list->ranges[i].lower), sizeof(meta->ambig_list->ranges[i].lower), 1, fp) != 1 ||
fwrite(&(meta->ambig_list->ranges[i].upper), sizeof(meta->ambig_list->ranges[i].upper), 1, fp) != 1
)
esl_fatal( "%s: Error writing ambiguity data for FM index.\n", argv[0]);
}
/* now append the FM-index data in fptmp to the desired output file, fp */
rewind(fptmp);
for (i=0; i<numblocks; i++) {
for(j=0; j< (meta->fwd_only?1:2); j++ ) { //do this once or twice, once for forward-T index, and possibly once for reversed
//first, read
if(fread(&block_length, sizeof(block_length), 1, fptmp) != 1)
esl_fatal( "%s: Error reading block_length in FM index.\n", argv[0]);
if(fread(&term_loc, sizeof(term_loc), 1, fptmp) != 1)
esl_fatal( "%s: Error reading terminal location in FM index.\n", argv[0]);
if(fread(&seq_offset, sizeof(seq_offset), 1, fptmp) != 1)
esl_fatal( "%s: Error reading seq_offset in FM index.\n", argv[0]);
if(fread(&ambig_offset, sizeof(ambig_offset ), 1, fptmp) != 1)
esl_fatal( "%s: Error reading ambig_offset in FM index.\n", argv[0]);
if(fread(&overlap, sizeof(overlap), 1, fptmp) != 1)
esl_fatal( "%s: Error reading overlap in FM index.\n", argv[0]);
if(fread(&seq_cnt, sizeof(seq_cnt), 1, fptmp) != 1)
esl_fatal( "%s: Error reading seq_cnt in FM index.\n", argv[0]);
if(fread(&ambig_cnt, sizeof(ambig_cnt), 1, fptmp) != 1)
esl_fatal( "%s: Error reading ambig_cnt in FM index.\n", argv[0]);
compressed_bytes = ((chars_per_byte-1+block_length)/chars_per_byte);
num_freq_cnts_b = 1+ceil((double)block_length/meta->freq_cnt_b);
num_freq_cnts_sb = 1+ceil((double)block_length/meta->freq_cnt_sb);
num_SA_samples = 1+floor((double)block_length/meta->freq_SA);
//j==0 test cause T and SA to be written only for forward sequence
if(j==0 && fread(T, sizeof(uint8_t), compressed_bytes, fptmp) != compressed_bytes)
esl_fatal( "%s: Error reading T in FM index.\n", argv[0]);
if(fread(BWT, sizeof(uint8_t), compressed_bytes, fptmp) != compressed_bytes)
esl_fatal( "%s: Error reading BWT in FM index.\n", argv[0]);
if(j==0 && fread(SAsamp, sizeof(uint32_t), (size_t)num_SA_samples, fptmp) != (size_t)num_SA_samples)
esl_fatal( "%s: Error reading SA in FM index.\n", argv[0]);
if(fread(occCnts_b, sizeof(uint16_t)*(meta->alph_size), (size_t)num_freq_cnts_b, fptmp) != (size_t)num_freq_cnts_b)
esl_fatal( "%s: Error reading occCnts_b in FM index.\n", argv[0]);
if(fread(occCnts_sb, sizeof(uint32_t)*(meta->alph_size), (size_t)num_freq_cnts_sb, fptmp) != (size_t)num_freq_cnts_sb)
esl_fatal( "%s: Error reading occCnts_sb in FM index.\n", argv[0]);
//then, write
if(fwrite(&block_length, sizeof(block_length), 1, fp) != 1)
esl_fatal( "%s: Error writing block_length in FM index.\n", argv[0]);
if(fwrite(&term_loc, sizeof(term_loc), 1, fp) != 1)
esl_fatal( "%s: Error writing terminal location in FM index.\n", argv[0]);
if(fwrite(&seq_offset, sizeof(seq_offset), 1, fp) != 1)
esl_fatal( "%s: Error writing seq_offset in FM index.\n", argv[0]);
if(fwrite(&ambig_offset, sizeof(ambig_offset), 1, fp) != 1)
esl_fatal( "%s: Error writing ambig_offset in FM index.\n", argv[0]);
if(fwrite(&overlap, sizeof(overlap), 1, fp) != 1)
esl_fatal( "%s: Error writing overlap in FM index.\n", argv[0]);
if(fwrite(&seq_cnt, sizeof(seq_cnt), 1, fp) != 1)
esl_fatal( "%s: Error writing seq_cnt in FM index.\n", argv[0]);
if(fwrite(&ambig_cnt, sizeof(ambig_cnt), 1, fp) != 1)
esl_fatal( "%s: Error writing ambig_cnt in FM index.\n", argv[0]);
if(j==0 && fwrite(T, sizeof(uint8_t), compressed_bytes, fp) != compressed_bytes)
esl_fatal( "%s: Error writing T in FM index.\n", argv[0]);
if(fwrite(BWT, sizeof(uint8_t), compressed_bytes, fp) != compressed_bytes)
esl_fatal( "%s: Error writing BWT in FM index.\n", argv[0]);
if(j==0 && fwrite(SAsamp, sizeof(uint32_t), (size_t)num_SA_samples, fp) != (size_t)num_SA_samples)
esl_fatal( "%s: Error writing SA in FM index.\n", argv[0]);
if(fwrite(occCnts_b, sizeof(uint16_t)*(meta->alph_size), (size_t)num_freq_cnts_b, fp) != (size_t)num_freq_cnts_b)
esl_fatal( "%s: Error writing occCnts_b in FM index.\n", argv[0]);
if(fwrite(occCnts_sb, sizeof(uint32_t)*(meta->alph_size), (size_t)num_freq_cnts_sb, fp) != (size_t)num_freq_cnts_sb)
esl_fatal( "%s: Error writing occCnts_sb in FM index.\n", argv[0]);
}
}
fprintf (stderr, "Number of characters in index: %ld\n", (long)total_char_count);
fprintf (stderr, "Number of FM-index blocks: %ld\n", (long)meta->block_count);
fclose(fp);
fclose(fptmp);
free(T);
free(BWT);
free(SA);
free(SAsamp);
free(occCnts_b);
free(cnts_b);
free(occCnts_sb);
free(cnts_sb);
fm_metaDestroy(meta);
esl_getopts_Destroy(go);
// compute and print the elapsed time in millisec
t2 = times(&ts2);
{
double clk_ticks = sysconf(_SC_CLK_TCK);
double elapsedTime = (t2-t1)/clk_ticks;
fprintf (stderr, "run time: %.2f seconds\n", elapsedTime);
}
return (eslOK);
ERROR:
/* Deallocate memory. */
if (fp) fclose(fp);
if (T) free(T);
if (BWT) free(BWT);
if (SA) free(SA);
if (SAsamp) free(SAsamp);
if (occCnts_b) free(occCnts_b);
if (cnts_b) free(cnts_b);
if (occCnts_sb) free(occCnts_sb);
if (cnts_sb) free(cnts_sb);
if (ambig_list.ranges) free(ambig_list.ranges);
fm_metaDestroy(meta);
esl_getopts_Destroy(go);
esl_sqfile_Close(sqfp);
esl_alphabet_Destroy(abc);
esl_sq_Destroy(sq);
if (tmpsq) esl_sq_Destroy(tmpsq);
if (block) esl_sq_DestroyBlock(block);
fprintf (stderr, "failure during memory allocation\n");
exit(status);
}
