int
main(int argc, char **argv)
{
ESL_GETOPTS *go = esl_getopts_CreateDefaultApp(options, 0, argc, argv, banner, usage);
ESL_RANDOMNESS *r = esl_randomness_Create(esl_opt_GetInteger(go, "-s"));
int be_verbose = esl_opt_GetBoolean(go, "-v");
int N = esl_opt_GetInteger(go, "-N");
if (be_verbose) printf("seed = %" PRIu32 "\n", esl_randomness_GetSeed(r));
utest_LogGamma(r, N, be_verbose);
utest_LinearRegression(r, TRUE, be_verbose);
utest_LinearRegression(r, FALSE, be_verbose);
esl_getopts_Destroy(go);
esl_randomness_Destroy(r);
exit(0);
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = NULL; /* application configuration */
ESL_RANDOMNESS *r = NULL; /* random number generator */
FILE *ofp = NULL; /* data output stream */
int outfmt = eslSQFILE_FASTA;
/* 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);
/* Open the output data file, if any */
if (esl_opt_GetString(go, "-o") != NULL)
{
if ((ofp = fopen(esl_opt_GetString(go, "-o"), "w")) == NULL)
esl_fatal("Failed to open output file %s\n", esl_opt_GetString(go, "-o"));
}
else ofp = stdout;
/* Initialize */
r = esl_randomness_Create(esl_opt_GetInteger(go, "--seed"));
/* Hand off execution to one of the three modes */
if (esl_opt_GetBoolean(go, "-A")) /* Alignment shuffling */
{
if (esl_opt_ArgNumber(go) != 1)
cmdline_failure(argv[0], "Incorrect number of command line arguments.\n");
msa_shuffling(go, r, ofp, outfmt);
}
else if (esl_opt_GetBoolean(go, "-G")) /* Sequence generation */
{
if (esl_opt_ArgNumber(go) != 0)
cmdline_failure(argv[0], "Incorrect number of command line arguments.\n");
seq_generation(go, r, ofp, outfmt);
}
else if (esl_opt_GetBoolean(go, "-S")) /* Sequence shuffling */
{
if (esl_opt_ArgNumber(go) != 1)
cmdline_failure(argv[0], "Incorrect number of command line arguments.\n");
seq_shuffling(go, r, ofp, outfmt);
}
if (esl_opt_GetString(go, "-o") != NULL) fclose(ofp);
esl_randomness_Destroy(r);
esl_getopts_Destroy(go);
return 0;
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = p7_CreateDefaultApp(options, 0, argc, argv, banner, usage);
ESL_RANDOMNESS *r = esl_randomness_CreateFast(esl_opt_GetInteger(go, "-s"));
ESL_ALPHABET *abc = esl_alphabet_Create(eslAMINO);
P7_HMM *hmm = NULL;
P7_BG *bg = NULL;
P7_PROFILE *gm = NULL;
P7_GMX *fwd = NULL;
P7_GMX *bck = NULL;
ESL_DSQ *dsq = NULL;
int M = esl_opt_GetInteger(go, "-M");
int L = esl_opt_GetInteger(go, "-L");
/* Sample a random HMM */
p7_hmm_Sample(r, M, abc, &hmm);
/* Configure a profile from the sampled HMM */
bg = p7_bg_Create(abc);
p7_bg_SetLength(bg, L);
gm = p7_profile_Create(hmm->M, abc);
p7_ProfileConfig(hmm, bg, gm, L, p7_LOCAL);
/* Other initial allocations */
dsq = malloc(sizeof(ESL_DSQ) * (L+2));
fwd = p7_gmx_Create(gm->M, L);
bck = p7_gmx_Create(gm->M, L);
p7_FLogsumInit();
utest_correct_normalization(r, gm, bg, dsq, L, fwd, bck);
free(dsq);
p7_gmx_Destroy(fwd);
p7_gmx_Destroy(bck);
p7_profile_Destroy(gm);
p7_bg_Destroy(bg);
p7_hmm_Destroy(hmm);
esl_alphabet_Destroy(abc);
esl_randomness_Destroy(r);
esl_getopts_Destroy(go);
return 0;
}
/* msa_shuffling()
*
* Shuffling multiple sequence alignments
*/
static int
msa_shuffling(ESL_GETOPTS *go, ESL_RANDOMNESS *r, FILE *ofp, int outfmt)
{
char *msafile = esl_opt_GetArg(go, 1);
int infmt = eslMSAFILE_UNKNOWN;
ESL_MSAFILE *afp = NULL;
ESL_MSA *msa = NULL;
ESL_MSA *shuf = NULL;
int N = esl_opt_GetInteger(go, "-N");
int i;
int status;
if ( (status = esl_msafile_Open(NULL, msafile, NULL, infmt, NULL, &afp)) != eslOK)
esl_msafile_OpenFailure(afp, status);
while ((status = esl_msafile_Read(afp, &msa)) != eslEOF)
{
if (status != eslOK) esl_msafile_ReadFailure(afp, status);
shuf = esl_msa_Clone(msa);
for (i = 0; i < N; i++)
{
if (esl_opt_GetBoolean(go, "--boot")) esl_msashuffle_Bootstrap(r, msa, shuf);
else esl_msashuffle_Shuffle (r, msa, shuf);
/* Set the name of the shuffled alignment */
if (msa->name != NULL) {
if (esl_opt_GetBoolean(go, "--boot")) {
if (N > 1) esl_msa_FormatName(shuf, "%s-sample-%d", msa->name, i);
else esl_msa_FormatName(shuf, "%s-sample", msa->name);
} else {
if (N > 1) esl_msa_FormatName(shuf, "%s-shuffle-%d", msa->name, i);
else esl_msa_FormatName(shuf, "%s-shuffle", msa->name);
}
} else {
if (esl_opt_GetBoolean(go, "--boot")) {
if (N > 1) esl_msa_FormatName(shuf, "sample-%d", i);
else esl_msa_FormatName(shuf, "sample");
} else {
if (N > 1) esl_msa_FormatName(shuf, "shuffle-%d", i);
else esl_msa_FormatName(shuf, "shuffle");
}
}
esl_msafile_Write(ofp, shuf, afp->format);
}
esl_msa_Destroy(shuf);
esl_msa_Destroy(msa);
}
esl_msafile_Close(afp);
return eslOK;
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = esl_getopts_CreateDefaultApp(options, 1, argc, argv, banner, usage);
char *hmmfile = esl_opt_GetArg(go, 1);
ESL_STOPWATCH *w = esl_stopwatch_Create();
ESL_ALPHABET *abc = NULL;
P7_HMMFILE *hfp = NULL;
P7_HMM *hmm = NULL;
P7_BG *bg = NULL;
P7_PROFILE *gm = NULL;
P7_OPROFILE *om = NULL;
int L = esl_opt_GetInteger(go, "-L");
int N = esl_opt_GetInteger(go, "-N");
int i;
if (p7_hmmfile_Open(hmmfile, NULL, &hfp) != eslOK) p7_Fail("Failed to open HMM file %s", hmmfile);
if (p7_hmmfile_Read(hfp, &abc, &hmm) != eslOK) p7_Fail("Failed to read HMM");
bg = p7_bg_Create(abc);
p7_bg_SetLength(bg, L);
gm = p7_profile_Create(hmm->M, abc);
p7_ProfileConfig(hmm, bg, gm, L, p7_LOCAL);
om = p7_oprofile_Create(gm->M, abc);
esl_stopwatch_Start(w);
for (i = 0; i < N; i++)
p7_oprofile_Convert(gm, om);
esl_stopwatch_Stop(w);
esl_stopwatch_Display(stdout, w, "# CPU time: ");
printf("# M = %d\n", gm->M);
p7_oprofile_Destroy(om);
p7_profile_Destroy(gm);
p7_bg_Destroy(bg);
p7_hmm_Destroy(hmm);
p7_hmmfile_Close(hfp);
esl_alphabet_Destroy(abc);
esl_stopwatch_Destroy(w);
esl_getopts_Destroy(go);
return 0;
}
static void
read_tabfile(ESL_GETOPTS *go, char *tabfile, ESL_KEYHASH *kh, ESL_DMATRIX *D)
{
ESL_FILEPARSER *efp = NULL;
int nline = 0;
int vfield = esl_opt_GetInteger(go, "-v");
int qfield = esl_opt_GetInteger(go, "-q");
int tfield = esl_opt_GetInteger(go, "-t");
char *tok;
int toklen;
int ntok;
double value;
int qidx, tidx;
if (esl_fileparser_Open(tabfile, NULL, &efp) != eslOK) esl_fatal("File open failed");
esl_fileparser_SetCommentChar(efp, '#');
esl_dmatrix_Set(D, eslINFINITY);
while (esl_fileparser_NextLine(efp) == eslOK)
{
nline++;
ntok = 0;
qidx = tidx = -1;
value = eslNaN;
while (esl_fileparser_GetTokenOnLine(efp, &tok, &toklen) == eslOK)
{
ntok++;
if (ntok == vfield) value = atof(tok);
if (ntok == qfield && esl_keyhash_Lookup(kh, tok, toklen, &qidx) != eslOK) esl_fatal("failed to find query key %s", tok);
if (ntok == tfield && esl_keyhash_Lookup(kh, tok, toklen, &tidx) != eslOK) esl_fatal("failed to find target key %s", tok);
}
if (qidx == -1) esl_fatal("Failed to find query name on line %d (looking for field %d)\n", nline, qfield);
if (tidx == -1) esl_fatal("Failed to find target name on line %d (looking for field %d)\n", nline, tfield);
if (isnan(value)) esl_fatal("Failed to find value on line %d (looking for field %d)\n", nline, vfield);
D->mx[qidx][tidx] = value;
if (D->mx[tidx][qidx] == eslINFINITY) D->mx[tidx][qidx] = value;
}
esl_fileparser_Close(efp);
}
static int
output_header(const ESL_GETOPTS *go, const struct cfg_s *cfg)
{
if (cfg->my_rank > 0) return eslOK;
/* if (! cfg->be_verbose) return eslOK; */
p7_banner(cfg->ofp, go->argv[0], banner);
fprintf(cfg->ofp, "# input alignment file: %s\n", cfg->alifile);
fprintf(cfg->ofp, "# output HMM file: %s\n", cfg->hmmfile);
if (esl_opt_IsUsed(go, "-n")) fprintf(cfg->ofp, "# name (the single) HMM: %s\n", esl_opt_GetString(go, "-n"));
if (esl_opt_IsUsed(go, "-o")) fprintf(cfg->ofp, "# output directed to file: %s\n", esl_opt_GetString(go, "-o"));
if (esl_opt_IsUsed(go, "-O")) fprintf(cfg->ofp, "# processed alignment resaved to: %s\n", esl_opt_GetString(go, "-O"));
if (esl_opt_IsUsed(go, "--amino")) fprintf(cfg->ofp, "# input alignment is asserted as: protein\n");
if (esl_opt_IsUsed(go, "--dna")) fprintf(cfg->ofp, "# input alignment is asserted as: DNA\n");
if (esl_opt_IsUsed(go, "--rna")) fprintf(cfg->ofp, "# input alignment is asserted as: RNA\n");
if (esl_opt_IsUsed(go, "--fast")) fprintf(cfg->ofp, "# model architecture construction: fast/heuristic\n");
if (esl_opt_IsUsed(go, "--hand")) fprintf(cfg->ofp, "# model architecture construction: hand-specified by RF annotation\n");
if (esl_opt_IsUsed(go, "--symfrac")) fprintf(cfg->ofp, "# sym fraction for model structure: %.3f\n", esl_opt_GetReal(go, "--symfrac"));
if (esl_opt_IsUsed(go, "--fragthresh"))fprintf(cfg->ofp, "# seq called fragment if < xL : %.3f\n", esl_opt_GetReal(go, "--fragthresh"));
if (esl_opt_IsUsed(go, "--wpb")) fprintf(cfg->ofp, "# relative weighting scheme: Henikoff PB\n");
if (esl_opt_IsUsed(go, "--wgsc")) fprintf(cfg->ofp, "# relative weighting scheme: G/S/C\n");
if (esl_opt_IsUsed(go, "--wblosum")) fprintf(cfg->ofp, "# relative weighting scheme: BLOSUM filter\n");
if (esl_opt_IsUsed(go, "--wnone")) fprintf(cfg->ofp, "# relative weighting scheme: none\n");
if (esl_opt_IsUsed(go, "--wid")) fprintf(cfg->ofp, "# frac id cutoff for BLOSUM wgts: %f\n", esl_opt_GetReal(go, "--wid"));
if (esl_opt_IsUsed(go, "--eent")) fprintf(cfg->ofp, "# effective seq number scheme: entropy weighting\n");
if (esl_opt_IsUsed(go, "--eclust")) fprintf(cfg->ofp, "# effective seq number scheme: single linkage clusters\n");
if (esl_opt_IsUsed(go, "--enone")) fprintf(cfg->ofp, "# effective seq number scheme: none\n");
if (esl_opt_IsUsed(go, "--eset")) fprintf(cfg->ofp, "# effective seq number: set to %f\n", esl_opt_GetReal(go, "--eset"));
if (esl_opt_IsUsed(go, "--ere") ) fprintf(cfg->ofp, "# minimum rel entropy target: %f bits\n", esl_opt_GetReal(go, "--ere"));
if (esl_opt_IsUsed(go, "--esigma") ) fprintf(cfg->ofp, "# entropy target sigma parameter: %f bits\n", esl_opt_GetReal(go, "--esigma"));
if (esl_opt_IsUsed(go, "--eid") ) fprintf(cfg->ofp, "# frac id cutoff for --eclust: %f\n", esl_opt_GetReal(go, "--eid"));
if (esl_opt_IsUsed(go, "--EmL") ) fprintf(cfg->ofp, "# seq length for MSV Gumbel mu fit: %d\n", esl_opt_GetInteger(go, "--EmL"));
if (esl_opt_IsUsed(go, "--EmN") ) fprintf(cfg->ofp, "# seq number for MSV Gumbel mu fit: %d\n", esl_opt_GetInteger(go, "--EmN"));
if (esl_opt_IsUsed(go, "--EvL") ) fprintf(cfg->ofp, "# seq length for Vit Gumbel mu fit: %d\n", esl_opt_GetInteger(go, "--EvL"));
if (esl_opt_IsUsed(go, "--EvN") ) fprintf(cfg->ofp, "# seq number for Vit Gumbel mu fit: %d\n", esl_opt_GetInteger(go, "--EvN"));
if (esl_opt_IsUsed(go, "--EfL") ) fprintf(cfg->ofp, "# seq length for Fwd exp tau fit: %d\n", esl_opt_GetInteger(go, "--EfL"));
if (esl_opt_IsUsed(go, "--EfN") ) fprintf(cfg->ofp, "# seq number for Fwd exp tau fit: %d\n", esl_opt_GetInteger(go, "--EfN"));
if (esl_opt_IsUsed(go, "--Eft") ) fprintf(cfg->ofp, "# tail mass for Fwd exp tau fit: %f\n", esl_opt_GetReal(go, "--Eft"));
#ifdef HMMER_THREADS
if (esl_opt_IsUsed(go, "--cpu")) fprintf(cfg->ofp, "# number of worker threads: %d\n", esl_opt_GetInteger(go, "--cpu"));
#endif
#ifdef HAVE_MPI
if (esl_opt_IsUsed(go, "--mpi") ) fprintf(cfg->ofp, "# parallelization mode: MPI\n");
#endif
if (esl_opt_IsUsed(go, "--seed")) {
if (esl_opt_GetInteger(go, "--seed") == 0) fprintf(cfg->ofp,"# random number seed: one-time arbitrary\n");
else fprintf(cfg->ofp,"# random number seed set to: %d\n", esl_opt_GetInteger(go, "--seed"));
}
if (esl_opt_IsUsed(go, "--laplace") ) fprintf(cfg->ofp, "# prior: Laplace +1\n");
fprintf(cfg->ofp, "# - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -\n\n");
return eslOK;
}
static int
output_header(FILE *ofp, ESL_GETOPTS *go, char *hmmfile, char *seqfile)
{
p7_banner(ofp, go->argv[0], banner);
if (fprintf(ofp, "# query sequence file: %s\n", seqfile) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (fprintf(ofp, "# target HMM database: %s\n", hmmfile) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "-o") && fprintf(ofp, "# output directed to file: %s\n", esl_opt_GetString(go, "-o")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--tblout") && fprintf(ofp, "# per-seq hits tabular output: %s\n", esl_opt_GetString(go, "--tblout")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--dfamtblout") && fprintf(ofp, "# hits output in Dfam format: %s\n", esl_opt_GetString(go, "--dfamtblout")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--aliscoresout") && fprintf(ofp, "# alignment scores output: %s\n", esl_opt_GetString(go, "--aliscoresout")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--acc") && fprintf(ofp, "# prefer accessions over names: yes\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--noali") && fprintf(ofp, "# show alignments in output: no\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--notextw") && fprintf(ofp, "# max ASCII text line length: unlimited\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--textw") && fprintf(ofp, "# max ASCII text line length: %d\n", esl_opt_GetInteger(go, "--textw")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "-E") && fprintf(ofp, "# profile reporting threshold: E-value <= %g\n", esl_opt_GetReal(go, "-E")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "-T") && fprintf(ofp, "# profile reporting threshold: score >= %g\n", esl_opt_GetReal(go, "-T")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--incE") && fprintf(ofp, "# profile inclusion threshold: E-value <= %g\n", esl_opt_GetReal(go, "--incE")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--incT") && fprintf(ofp, "# profile inclusion threshold: score >= %g\n", esl_opt_GetReal(go, "--incT")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--cut_ga") && fprintf(ofp, "# model-specific thresholding: GA cutoffs\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--cut_nc") && fprintf(ofp, "# model-specific thresholding: NC cutoffs\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--cut_tc") && fprintf(ofp, "# model-specific thresholding: TC cutoffs\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--max") && fprintf(ofp, "# Max sensitivity mode: on [all heuristic filters off]\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--F1") && fprintf(ofp, "# MSV filter P threshold: <= %g\n", esl_opt_GetReal(go, "--F1")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--F2") && fprintf(ofp, "# Vit filter P threshold: <= %g\n", esl_opt_GetReal(go, "--F2")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--F3") && fprintf(ofp, "# Fwd filter P threshold: <= %g\n", esl_opt_GetReal(go, "--F3")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--nobias") && fprintf(ofp, "# biased composition HMM filter: off\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--B1") && fprintf(ofp, "# biased comp MSV window len: %d\n", esl_opt_GetInteger(go, "--B1")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--B2") && fprintf(ofp, "# biased comp Viterbi window len: %d\n", esl_opt_GetInteger(go, "--B2")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--B3") && fprintf(ofp, "# biased comp Forward window len: %d\n", esl_opt_GetInteger(go, "--B3")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--nonull2") && fprintf(ofp, "# null2 bias corrections: off\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--toponly") && fprintf(ofp, "# search only top strand: on\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--bottomonly") && fprintf(ofp, "# search only bottom strand: on\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "-Z") && fprintf(ofp, "# sequence search space set to: %.0f\n", esl_opt_GetReal(go, "-Z")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--seed")) {
if (esl_opt_GetInteger(go, "--seed")==0 && fprintf(ofp, "# random number seed: one-time arbitrary\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
else if ( fprintf(ofp, "# random number seed set to: %d\n", esl_opt_GetInteger(go, "--seed")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
}
if (esl_opt_IsUsed(go, "--qformat") && fprintf(ofp, "# input seqfile format asserted: %s\n", esl_opt_GetString(go, "--qformat")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--w_beta") && fprintf(ofp, "# window length beta value: %g\n", esl_opt_GetReal(go, "--w_beta")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--w_length") && fprintf(ofp, "# window length : %d\n", esl_opt_GetInteger(go, "--w_length")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
// if (esl_opt_IsUsed(go, "--daemon") && fprintf(ofp, "run as a daemon process\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
#ifdef HMMER_THREADS
if (esl_opt_IsUsed(go, "--cpu") && fprintf(ofp, "# number of worker threads: %d\n", esl_opt_GetInteger(go, "--cpu")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
#endif
if (fprintf(ofp, "# - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -\n\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
return eslOK;
}
/* Fetch in a random sequence of length <L> from the the pre-digitized
* concatenated sequence database, select a random subseq, shuffle it
* by the chosen algorithm; set dsq[1..L] to the resulting randomized
* segment.
*
* If <logfp> is non-NULL, append one or more "<sqname> <from> <to>"
* fields to current line, to record where the random segment was
* selected from. This is useful in cases where we want to track back
* the origin of a high-scoring segment, in case the randomization
* wasn't good enough to obscure the identity of a segment.
*
*/
static int
set_random_segment(ESL_GETOPTS *go, struct cfg_s *cfg, FILE *logfp, ESL_DSQ *dsq, int L)
{
ESL_SQ *sq = esl_sq_CreateDigital(cfg->abc);
int minDPL = esl_opt_GetInteger(go, "--minDPL");
int db_dependent = (esl_opt_GetBoolean(go, "--iid") == TRUE ? FALSE : TRUE);
char *pkey = NULL;
int start, end;
int64_t Lseq;
int status;
if (L==0) return eslOK;
if (L > cfg->db_maxL) esl_fatal("can't fetch a segment of length %d; database max is %d\n", L, cfg->db_maxL);
/* fetch a random subseq from the source database */
esl_sq_GrowTo(sq, L);
if (db_dependent)
{
do {
if (pkey != NULL) free(pkey);
if (esl_ssi_FindNumber(cfg->dbfp->data.ascii.ssi, esl_rnd_Roll(cfg->r, cfg->db_nseq), NULL, NULL, NULL, &Lseq, &pkey) != eslOK)
esl_fatal("failed to look up a random seq");
} while (Lseq < L);
start = 1 + esl_rnd_Roll(cfg->r, Lseq-L);
end = start + L - 1;
if (esl_sqio_FetchSubseq(cfg->dbfp, pkey, start, end, sq) != eslOK) esl_fatal("failed to fetch subseq");
esl_sq_ConvertDegen2X(sq);
}
/* log sequence source info: <name> <start> <end> */
if (logfp != NULL && db_dependent)
fprintf(logfp, " %-15s %5d %5d", pkey, start, end);
/* Now apply the appropriate randomization algorithm */
if (esl_opt_GetBoolean(go, "--mono")) status = esl_rsq_XShuffle (cfg->r, sq->dsq, L, sq->dsq);
else if (esl_opt_GetBoolean(go, "--di")) {
if (L < minDPL) status = esl_rsq_XShuffle (cfg->r, sq->dsq, L, sq->dsq);
else status = esl_rsq_XShuffleDP(cfg->r, sq->dsq, L, cfg->abc->Kp, sq->dsq);
}
else if (esl_opt_GetBoolean(go, "--markov0")) status = esl_rsq_XMarkov0 (cfg->r, sq->dsq, L, cfg->abc->Kp, sq->dsq);
else if (esl_opt_GetBoolean(go, "--markov1")) status = esl_rsq_XMarkov1 (cfg->r, sq->dsq, L, cfg->abc->Kp, sq->dsq);
else if (esl_opt_GetBoolean(go, "--reverse")) status = esl_rsq_XReverse (sq->dsq, L, sq->dsq);
else if (esl_opt_GetBoolean(go, "--iid")) status = esl_rsq_xIID (cfg->r, cfg->fq, cfg->abc->K, L, sq->dsq);
else status = eslEINCONCEIVABLE;
if (status != eslOK) esl_fatal("esl's shuffling failed");
memcpy(dsq, sq->dsq+1, sizeof(ESL_DSQ) * L);
esl_sq_Destroy(sq);
free(pkey);
return eslOK;
}
/* utest_odds(): test accuracy of logf, expf on odds ratios,
* our main intended use.
*/
static void
utest_odds(ESL_GETOPTS *go, ESL_RANDOMNESS *r)
{
int N = esl_opt_GetInteger(go, "-N");
int verbose = esl_opt_GetBoolean(go, "-v");
int very_verbose = esl_opt_GetBoolean(go, "--vv");
int i;
float p1, p2, odds;
union { __m128 v; float x[4]; } r1;
union { __m128 v; float x[4]; } r2;
float scalar_r1, scalar_r2;
double err1, maxerr1 = 0.0, avgerr1 = 0.0; /* errors on logf() */
double err2, maxerr2 = 0.0, avgerr2 = 0.0; /* errors on expf() */
for (i = 0; i < N; i++)
{
p1 = esl_rnd_UniformPositive(r);
p2 = esl_rnd_UniformPositive(r);
odds = p1 / p2;
if (odds == 0.0) esl_fatal("whoa, odds ratio can't be 0!\n");
r1.v = esl_sse_logf(_mm_set1_ps(odds)); /* r1.x[z] = log(p1/p2) */
scalar_r1 = log(odds);
err1 = (r1.x[0] == 0. && scalar_r1 == 0.) ? 0.0 : 2 * fabs(r1.x[0] - scalar_r1) / fabs(r1.x[0] + scalar_r1);
if (err1 > maxerr1) maxerr1 = err1;
avgerr1 += err1 / (float) N;
if (isnan(avgerr1)) esl_fatal("whoa, what?\n");
r2.v = esl_sse_expf(r1.v); /* and back to odds */
scalar_r2 = exp(r1.x[0]);
err2 = (r2.x[0] == 0. && scalar_r2 == 0.) ? 0.0 : 2 * fabs(r2.x[0] - scalar_r2) / fabs(r2.x[0] + scalar_r2);
if (err2 > maxerr2) maxerr2 = err2;
avgerr2 += err2 / (float) N;
if (very_verbose)
printf("%13.7g %13.7g %13.7g %13.7g %13.7g %13.7g %13.7g\n", odds, scalar_r1, r1.x[0], scalar_r2, r2.x[0], err1, err2);
}
if (verbose) {
printf("Average [max] logf() relative error in %d odds trials: %13.8g [%13.8g]\n", N, avgerr1, maxerr1);
printf("Average [max] expf() relative error in %d odds trials: %13.8g [%13.8g]\n", N, avgerr2, maxerr2);
printf("(random seed : %" PRIu32 ")\n", esl_randomness_GetSeed(r));
}
if (avgerr1 > 1e-8) esl_fatal("average error on logf() is intolerable\n");
if (maxerr1 > 1e-6) esl_fatal("maximum error on logf() is intolerable\n");
if (avgerr2 > 1e-8) esl_fatal("average error on expf() is intolerable\n");
if (maxerr2 > 1e-6) esl_fatal("maximum error on expf() is intolerable\n");
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = esl_getopts_CreateDefaultApp(options, 0, argc, argv, banner, usage);
ESL_RANDOMNESS *r = esl_randomness_CreateFast(esl_opt_GetInteger(go, "-s"));
ESL_ALPHABET *abc = NULL;
P7_BG *bg = NULL;
int M = esl_opt_GetInteger(go, "-M");
int L = esl_opt_GetInteger(go, "-L");
int N = esl_opt_GetInteger(go, "-N");
/* First round of tests for DNA alphabets. */
if ((abc = esl_alphabet_Create(eslDNA)) == NULL) esl_fatal("failed to create alphabet");
if ((bg = p7_bg_Create(abc)) == NULL) esl_fatal("failed to create null model");
if (esl_opt_GetBoolean(go, "-v")) printf("ViterbiFilter() tests, DNA\n");
utest_viterbi_filter(r, abc, bg, M, L, N);
utest_viterbi_filter(r, abc, bg, 1, L, 10);
utest_viterbi_filter(r, abc, bg, M, 1, 10);
esl_alphabet_Destroy(abc);
p7_bg_Destroy(bg);
/* Second round of tests for amino alphabets. */
if ((abc = esl_alphabet_Create(eslAMINO)) == NULL) esl_fatal("failed to create alphabet");
if ((bg = p7_bg_Create(abc)) == NULL) esl_fatal("failed to create null model");
if (esl_opt_GetBoolean(go, "-v")) printf("ViterbiFilter() tests, protein\n");
utest_viterbi_filter(r, abc, bg, M, L, N);
utest_viterbi_filter(r, abc, bg, 1, L, 10);
utest_viterbi_filter(r, abc, bg, M, 1, 10);
esl_alphabet_Destroy(abc);
p7_bg_Destroy(bg);
esl_getopts_Destroy(go);
esl_randomness_Destroy(r);
return eslOK;
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = esl_getopts_CreateDefaultApp(options, 0, argc, argv, banner, usage);
ESL_RANDOMNESS *r = esl_randomness_CreateFast(esl_opt_GetInteger(go, "-s"));
ESL_ALPHABET *abc = esl_alphabet_Create(eslAMINO);
P7_BG *bg = p7_bg_Create(abc);
int M = esl_opt_GetInteger(go, "-M");
int L = esl_opt_GetInteger(go, "-L");
int N = esl_opt_GetInteger(go, "-N");
float tol = esl_opt_GetReal (go, "-t");
p7_FLogsumInit();
utest_decoding(r, abc, bg, M, L, N, tol);
esl_getopts_Destroy(go);
esl_randomness_Destroy(r);
esl_alphabet_Destroy(abc);
p7_bg_Destroy(bg);
return eslOK;
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = esl_getopts_CreateDefaultApp(options, 0, argc, argv, banner, usage);
ESL_RANDOMNESS *r = NULL;
char *bitfile = esl_opt_GetString (go, "--bitfile");
int nbins = esl_opt_GetInteger(go, "-b");
int n = esl_opt_GetInteger(go, "-n");
int be_verbose = esl_opt_GetBoolean(go, "-v");
int seed = esl_opt_GetInteger(go, "-s");
if (esl_opt_GetBoolean(go, "-r")) r = esl_randomness_CreateTimeseeded();
else r = esl_randomness_Create(seed);
utest_random(seed, n, nbins, be_verbose);
utest_choose(r, n, nbins, be_verbose);
if (bitfile != NULL) save_bitfile(bitfile, r, n);
esl_randomness_Destroy(r);
esl_getopts_Destroy(go);
return 0;
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = esl_getopts_CreateDefaultApp(options, 0, argc, argv, banner, usage);
ESL_RANDOMNESS *r = esl_randomness_Create(esl_opt_GetInteger(go, "-s"));;
utest_logf(go);
utest_expf(go);
utest_odds(go, r);
esl_randomness_Destroy(r);
esl_getopts_Destroy(go);
return 0;
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = p7_CreateDefaultApp(options, 0, argc, argv, banner, usage);
ESL_RANDOMNESS *r = esl_randomness_CreateFast(esl_opt_GetInteger(go, "-s"));
ESL_ALPHABET *abc = NULL;
P7_BG *bg = NULL;
int M = esl_opt_GetInteger(go, "-M");
int L = esl_opt_GetInteger(go, "-L");
int N = esl_opt_GetInteger(go, "-N");
/* first round of tests for DNA alphabets. */
if ((abc = esl_alphabet_Create(eslDNA)) == NULL) esl_fatal("failed to create alphabet");
if ((bg = p7_bg_Create(abc)) == NULL) esl_fatal("failed to create null model");
utest_optacc(go, r, abc, bg, M, L, N); /* normal sized models */
utest_optacc(go, r, abc, bg, 1, L, 10); /* size 1 models */
utest_optacc(go, r, abc, bg, M, 1, 10); /* size 1 sequences */
esl_alphabet_Destroy(abc);
p7_bg_Destroy(bg);
/* Second round of tests for amino alphabets. */
if ((abc = esl_alphabet_Create(eslAMINO)) == NULL) esl_fatal("failed to create alphabet");
if ((bg = p7_bg_Create(abc)) == NULL) esl_fatal("failed to create null model");
utest_optacc(go, r, abc, bg, M, L, N);
utest_optacc(go, r, abc, bg, 1, L, 10);
utest_optacc(go, r, abc, bg, M, 1, 10);
esl_alphabet_Destroy(abc);
p7_bg_Destroy(bg);
esl_getopts_Destroy(go);
esl_randomness_Destroy(r);
return eslOK;
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = p7_CreateDefaultApp(options, 0, argc, argv, banner, usage);
ESL_RANDOMNESS *r = esl_randomness_CreateFast(esl_opt_GetInteger(go, "-s"));
p7_FLogsumInit();
utest_FLogsumError(go, r);
utest_FLogsumSpecials();
esl_randomness_Destroy(r);
esl_getopts_Destroy(go);
return eslOK;
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = esl_getopts_CreateDefaultApp(options, 0, argc, argv, banner, usage);
ESL_RANDOMNESS *rng = esl_randomness_CreateFast(esl_opt_GetInteger(go, "-s"));
int be_verbose = esl_opt_GetBoolean(go, "-v");
if (be_verbose) printf("p7_bg unit test: rng seed %" PRIu32 "\n", esl_randomness_GetSeed(rng));
utest_ReadWrite(rng);
esl_randomness_Destroy(rng);
esl_getopts_Destroy(go);
return 0;
}
static int
output_header(FILE *ofp, const ESL_GETOPTS *go, char *hmmfile, char *seqfile)
{
p7_banner(ofp, go->argv[0], banner);
if (fprintf(ofp, "# query HMM file: %s\n", hmmfile) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (fprintf(ofp, "# target sequence database: %s\n", seqfile) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "-o") && fprintf(ofp, "# output directed to file: %s\n", esl_opt_GetString(go, "-o")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "-A") && fprintf(ofp, "# MSA of all hits saved to file: %s\n", esl_opt_GetString(go, "-A")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--tblout") && fprintf(ofp, "# per-seq hits tabular output: %s\n", esl_opt_GetString(go, "--tblout")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--domtblout") && fprintf(ofp, "# per-dom hits tabular output: %s\n", esl_opt_GetString(go, "--domtblout")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--pfamtblout") && fprintf(ofp, "# pfam-style tabular hit output: %s\n", esl_opt_GetString(go, "--pfamtblout")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--acc") && fprintf(ofp, "# prefer accessions over names: yes\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--noali") && fprintf(ofp, "# show alignments in output: no\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--notextw") && fprintf(ofp, "# max ASCII text line length: unlimited\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--textw") && fprintf(ofp, "# max ASCII text line length: %d\n", esl_opt_GetInteger(go, "--textw")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "-E") && fprintf(ofp, "# sequence reporting threshold: E-value <= %g\n", esl_opt_GetReal(go, "-E")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "-T") && fprintf(ofp, "# sequence reporting threshold: score >= %g\n", esl_opt_GetReal(go, "-T")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--domE") && fprintf(ofp, "# domain reporting threshold: E-value <= %g\n", esl_opt_GetReal(go, "--domE")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--domT") && fprintf(ofp, "# domain reporting threshold: score >= %g\n", esl_opt_GetReal(go, "--domT")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--incE") && fprintf(ofp, "# sequence inclusion threshold: E-value <= %g\n", esl_opt_GetReal(go, "--incE")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--incT") && fprintf(ofp, "# sequence inclusion threshold: score >= %g\n", esl_opt_GetReal(go, "--incT")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--incdomE") && fprintf(ofp, "# domain inclusion threshold: E-value <= %g\n", esl_opt_GetReal(go, "--incdomE")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--incdomT") && fprintf(ofp, "# domain inclusion threshold: score >= %g\n", esl_opt_GetReal(go, "--incdomT")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--cut_ga") && fprintf(ofp, "# model-specific thresholding: GA cutoffs\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--cut_nc") && fprintf(ofp, "# model-specific thresholding: NC cutoffs\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--cut_tc") && fprintf(ofp, "# model-specific thresholding: TC cutoffs\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--max") && fprintf(ofp, "# Max sensitivity mode: on [all heuristic filters off]\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--F1") && fprintf(ofp, "# MSV filter P threshold: <= %g\n", esl_opt_GetReal(go, "--F1")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--F2") && fprintf(ofp, "# Vit filter P threshold: <= %g\n", esl_opt_GetReal(go, "--F2")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--F3") && fprintf(ofp, "# Fwd filter P threshold: <= %g\n", esl_opt_GetReal(go, "--F3")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--nobias") && fprintf(ofp, "# biased composition HMM filter: off\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--restrictdb_stkey") && fprintf(ofp, "# Restrict db to start at seq key: %s\n", esl_opt_GetString(go, "--restrictdb_stkey")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--restrictdb_n") && fprintf(ofp, "# Restrict db to # target seqs: %d\n", esl_opt_GetInteger(go, "--restrictdb_n")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--ssifile") && fprintf(ofp, "# Override ssi file to: %s\n", esl_opt_GetString(go, "--ssifile")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--nonull2") && fprintf(ofp, "# null2 bias corrections: off\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "-Z") && fprintf(ofp, "# sequence search space set to: %.0f\n", esl_opt_GetReal(go, "-Z")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--domZ") && fprintf(ofp, "# domain search space set to: %.0f\n", esl_opt_GetReal(go, "--domZ")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--seed")) {
if (esl_opt_GetInteger(go, "--seed") == 0 && fprintf(ofp, "# random number seed: one-time arbitrary\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
else if ( fprintf(ofp, "# random number seed set to: %d\n", esl_opt_GetInteger(go, "--seed")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
}
if (esl_opt_IsUsed(go, "--tformat") && fprintf(ofp, "# targ <seqfile> format asserted: %s\n", esl_opt_GetString(go, "--tformat")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (fprintf(ofp, "# - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -\n\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
return eslOK;
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = p7_CreateDefaultApp(options, 0, argc, argv, banner, usage);
ESL_RANDOMNESS *rng = esl_randomness_CreateFast(esl_opt_GetInteger(go, "-s"));
fprintf(stderr, "## %s\n", argv[0]);
fprintf(stderr, "# rng seed = %" PRIu32 "\n", esl_randomness_GetSeed(rng));
utest_sampling(rng);
fprintf(stderr, "# status = ok\n");
esl_randomness_Destroy(rng);
esl_getopts_Destroy(go);
exit(0); /* success */
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = esl_getopts_CreateDefaultApp(options, 0, argc, argv, banner, usage);
ESL_RANDOMNESS *r = esl_randomness_CreateFast(esl_opt_GetInteger(go, "-s"));
ESL_ALPHABET *abc = NULL;
P7_HMM *hmm = NULL;
P7_PROFILE *gm = NULL;
P7_OPROFILE *om = NULL;
P7_BG *bg = NULL;
ESL_DSQ *dsq = NULL;
ESL_SQ *sq = NULL;
int M = 6;
int L = 10;
int ntrace = 1000;
if ((abc = esl_alphabet_Create(eslAMINO)) == NULL) esl_fatal("failed to create alphabet");
if (p7_hmm_Sample(r, M, abc, &hmm) != eslOK) esl_fatal("failed to sample an HMM");
if ((bg = p7_bg_Create(abc)) == NULL) esl_fatal("failed to create null model");
if ((gm = p7_profile_Create(hmm->M, abc)) == NULL) esl_fatal("failed to create profile");
if (p7_ProfileConfig(hmm, bg, gm, L, p7_LOCAL) != eslOK) esl_fatal("failed to config profile");
if ((om = p7_oprofile_Create(gm->M, abc)) == NULL) esl_fatal("failed to create optimized profile");
if (p7_oprofile_Convert(gm, om) != eslOK) esl_fatal("failed to convert profile");
/* Test with randomly generated (iid) sequence */
if ((dsq = malloc(sizeof(ESL_DSQ) *(L+2))) == NULL) esl_fatal("malloc failed");
if (esl_rsq_xfIID(r, bg->f, abc->K, L, dsq) != eslOK) esl_fatal("seq generation failed");
utest_stotrace(go, r, abc, gm, om, dsq, L, ntrace);
/* Test with seq sampled from profile */
if ((sq = esl_sq_CreateDigital(abc)) == NULL) esl_fatal("sequence allocation failed");
if (p7_ProfileEmit(r, hmm, gm, bg, sq, NULL) != eslOK) esl_fatal("profile emission failed");
utest_stotrace(go, r, abc, gm, om, sq->dsq, sq->n, ntrace);
esl_sq_Destroy(sq);
free(dsq);
p7_oprofile_Destroy(om);
p7_profile_Destroy(gm);
p7_bg_Destroy(bg);
p7_hmm_Destroy(hmm);
esl_alphabet_Destroy(abc);
esl_randomness_Destroy(r);
esl_getopts_Destroy(go);
return 0;
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = NULL;
struct cfg_s cfg;
int status = eslOK;
impl_Init(); /* processor specific initialization */
p7_FLogsumInit(); /* we're going to use table-driven Logsum() approximations at times */
/* Initialize what we can in the config structure (without knowing the alphabet yet)
*/
cfg.hmmfile = NULL;
cfg.dbfile = NULL;
cfg.do_mpi = FALSE; /* this gets reset below, if we init MPI */
cfg.nproc = 0; /* this gets reset below, if we init MPI */
cfg.my_rank = 0; /* this gets reset below, if we init MPI */
cfg.firstseq_key = NULL;
cfg.n_targetseq = -1;
process_commandline(argc, argv, &go, &cfg.hmmfile, &cfg.dbfile);
/* is the range restricted? */
if (esl_opt_IsUsed(go, "--restrictdb_stkey") )
if ((cfg.firstseq_key = esl_opt_GetString(go, "--restrictdb_stkey")) == NULL) p7_Fail("Failure capturing --restrictdb_stkey\n");
if (esl_opt_IsUsed(go, "--restrictdb_n") )
cfg.n_targetseq = esl_opt_GetInteger(go, "--restrictdb_n");
if ( cfg.n_targetseq != -1 && cfg.n_targetseq < 1 )
p7_Fail("--restrictdb_n must be >= 1\n");
{
status = serial_master(go, &cfg);
}
esl_getopts_Destroy(go);
return status;
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = p7_CreateDefaultApp(options, 0, argc, argv, banner, usage);
ESL_RANDOMNESS *rng = esl_randomness_Create(esl_opt_GetInteger(go, "-s"));
ESL_ALPHABET *abc = esl_alphabet_Create(eslAMINO);
int M = 50;
fprintf(stderr, "## %s\n", argv[0]);
fprintf(stderr, "# rng seed = %" PRIu32 "\n", esl_randomness_GetSeed(rng));
utest_generation (rng, M, abc, 10); // test a bunch of seqs to try to make sure we exercise exact domain score recalculation
utest_singlemulti(rng, M, abc, 10);
fprintf(stderr, "# status = ok\n");
esl_alphabet_Destroy(abc);
esl_randomness_Destroy(rng);
esl_getopts_Destroy(go);
return 0;
}
void
worker_process(ESL_GETOPTS *go)
{
HMMD_COMMAND *cmd = NULL; /* see hmmpgmd.h */
int shutdown = 0;
WORKER_ENV env;
int status;
/* Initializations */
impl_Init();
p7_FLogsumInit(); /* we're going to use table-driven Logsum() approximations at times */
if (esl_opt_IsOn(go, "--cpu")) env.ncpus = esl_opt_GetInteger(go, "--cpu");
else esl_threads_CPUCount(&env.ncpus);
env.hmm_db = NULL;
env.seq_db = NULL;
env.fd = setup_masterside_comm(go);
while (!shutdown)
{
if ((status = read_Command(&cmd, &env)) != eslOK) break;
switch (cmd->hdr.command) {
case HMMD_CMD_INIT: process_InitCmd (cmd, &env); break;
case HMMD_CMD_SCAN: process_SearchCmd(cmd, &env); break;
case HMMD_CMD_SEARCH: process_SearchCmd(cmd, &env); break;
case HMMD_CMD_SHUTDOWN: process_Shutdown (cmd, &env); shutdown = 1; break;
default: p7_syslog(LOG_ERR,"[%s:%d] - unknown command %d (%d)\n", __FILE__, __LINE__, cmd->hdr.command, cmd->hdr.length);
}
free(cmd);
cmd = NULL;
}
if (env.hmm_db) p7_hmmcache_Close(env.hmm_db);
if (env.seq_db) p7_seqcache_Close(env.seq_db);
if (env.fd != -1) close(env.fd);
return;
}
static void
serial_master(ESL_GETOPTS *go, struct cfg_s *cfg)
{
P7_HMM *hmm = NULL;
double *xv = NULL; /* results: array of N scores */
int *av = NULL; /* optional results: array of N alignment lengths */
char errbuf[eslERRBUFSIZE];
int status;
if ((status = init_master_cfg(go, cfg, errbuf)) != eslOK) p7_Fail(errbuf);
if ((xv = malloc(sizeof(double) * cfg->N)) == NULL) p7_Fail("allocation failed");
if (esl_opt_GetBoolean(go, "-a") &&
(av = malloc(sizeof(int) * cfg->N)) == NULL) p7_Fail("allocation failed");
while ((status = p7_hmmfile_Read(cfg->hfp, &(cfg->abc), &hmm)) != eslEOF)
{
if (status == eslEOD) p7_Fail("read failed, HMM file %s may be truncated?", cfg->hmmfile);
else if (status == eslEFORMAT) p7_Fail("bad file format in HMM file %s", cfg->hmmfile);
else if (status == eslEINCOMPAT) p7_Fail("HMM file %s contains different alphabets", cfg->hmmfile);
else if (status != eslOK) p7_Fail("Unexpected error in reading HMMs from %s", cfg->hmmfile);
if (cfg->bg == NULL) {
if (esl_opt_GetBoolean(go, "--bgflat")) cfg->bg = p7_bg_CreateUniform(cfg->abc);
else cfg->bg = p7_bg_Create(cfg->abc);
p7_bg_SetLength(cfg->bg, esl_opt_GetInteger(go, "-L")); /* set the null model background length in both master and workers. */
}
if (esl_opt_GetBoolean(go, "--recal")) {
if (recalibrate_model(go, cfg, errbuf, hmm) != eslOK) p7_Fail(errbuf);
}
if (process_workunit(go, cfg, errbuf, hmm, xv, av) != eslOK) p7_Fail(errbuf);
if (output_result (go, cfg, errbuf, hmm, xv, av) != eslOK) p7_Fail(errbuf);
p7_hmm_Destroy(hmm);
}
free(xv);
if (av != NULL) free(av);
}
static int
output_header(const ESL_GETOPTS *go, FILE *ofp, char *alifile, char *postmsafile)
{
p7_banner(ofp, go->argv[0], banner);
if (fprintf(ofp, "# input alignment file: %s\n", alifile) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--alirange") || esl_opt_IsUsed(go, "--modelrange") ) {
if (fprintf(ofp, "# output alignment file: %s\n", postmsafile) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
}
if (esl_opt_IsUsed(go, "--alirange") && fprintf(ofp, "# alignment range: %s\n", esl_opt_GetString(go, "--alirange")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--modelrange") && fprintf(ofp, "# model range: %s\n", esl_opt_GetString(go, "--modelrange"))< 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--apendmask") && fprintf(ofp, "# add to existing mask: [on]\n" )< 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--model2ali") && fprintf(ofp, "# ali ranges for model range: %s\n", esl_opt_GetString(go, "--model2ali")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--ali2model") && fprintf(ofp, "# model ranges for ali range: %s\n", esl_opt_GetString(go, "--ali2model")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "-o") && fprintf(ofp, "# output directed to file: %s\n", esl_opt_GetString(go, "-o")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--amino") && fprintf(ofp, "# input alignment is asserted as: protein\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--dna") && fprintf(ofp, "# input alignment is asserted as: DNA\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--rna") && fprintf(ofp, "# input alignment is asserted as: RNA\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--fast") && fprintf(ofp, "# model architecture construction: fast/heuristic\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--hand") && fprintf(ofp, "# model architecture construction: hand-specified by RF annotation\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--symfrac") && fprintf(ofp, "# sym fraction for model structure: %.3f\n", esl_opt_GetReal(go, "--symfrac")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--fragthresh") && fprintf(ofp, "# seq called frag if L <= x*alen: %.3f\n", esl_opt_GetReal(go, "--fragthresh")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--wpb") && fprintf(ofp, "# relative weighting scheme: Henikoff PB\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--wgsc") && fprintf(ofp, "# relative weighting scheme: G/S/C\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--wblosum") && fprintf(ofp, "# relative weighting scheme: BLOSUM filter\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--wnone") && fprintf(ofp, "# relative weighting scheme: none\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--wid") && fprintf(ofp, "# frac id cutoff for BLOSUM wgts: %f\n", esl_opt_GetReal(go, "--wid")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
if (esl_opt_IsUsed(go, "--seed")) {
if (esl_opt_GetInteger(go, "--seed") == 0 && fprintf(ofp,"# random number seed: one-time arbitrary\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
else if ( fprintf(ofp,"# random number seed set to: %d\n", esl_opt_GetInteger(go, "--seed")) < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
}
if (fprintf(ofp, "# - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -\n\n") < 0) ESL_EXCEPTION_SYS(eslEWRITE, "write failed");
return eslOK;
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = esl_getopts_CreateDefaultApp(options, 0, argc, argv, banner, usage);
ESL_STOPWATCH *w = esl_stopwatch_Create();
int N = esl_opt_GetInteger(go, "-N");
float origx = 2.0;
float x = origx;
__m128 xv = _mm_set1_ps(x);
int i;
/* First, serial time. */
esl_stopwatch_Start(w);
for (i = 0; i < N; i++) {
x = logf(x);
x = expf(x);
}
esl_stopwatch_Stop(w);
esl_stopwatch_Display(stdout, w, "# serial CPU time: ");
/* Vector time */
esl_stopwatch_Start(w);
for (i = 0; i < N; i++) {
xv = esl_sse_logf(xv);
xv = esl_sse_expf(xv);
}
esl_stopwatch_Stop(w);
esl_stopwatch_Display(stdout, w, "# vector CPU time: ");
/* If you don't do something with x and xv, the compiler may optimize them away */
printf("%g => many scalar logf,expf cycles => %g\n", origx, x);
printf("%g => many vector logf,expf cycles => ", origx);
esl_sse_dump_ps(stdout, xv);
printf("\n");
esl_stopwatch_Destroy(w);
esl_getopts_Destroy(go);
return 0;
}
static int
setup_masterside_comm(ESL_GETOPTS *opts)
{
int fd = -1;
int cnt;
int sec;
int connected;
struct sockaddr_in addr;
/* create a reliable, stream socket using TCP */
if ((fd = socket(AF_INET, SOCK_STREAM, 0)) < 0) LOG_FATAL_MSG("socket", errno);
/* construct the server address structure */
memset(&addr, 0, sizeof(addr));
addr.sin_family = AF_INET;
addr.sin_port = htons(esl_opt_GetInteger(opts, "--wport"));
if ((inet_pton(AF_INET, esl_opt_GetString(opts, "--worker"), &addr.sin_addr)) < 0) LOG_FATAL_MSG("inet pton", errno);
/* try to connect to the master server */
cnt = 0;
sec = 1;
connected = -1;
while (connected < 0) {
/* establish the connection to the master server */
if ((connected = connect(fd, (struct sockaddr *) &addr, sizeof(addr))) < 0) {
if (errno != ECONNREFUSED) LOG_FATAL_MSG("connect", errno);
/* the master server is not listening. sleep and try again */
sleep(sec);
if (++cnt > 10) {
cnt = 0;
if (sec < 64) sec *= 2;
}
}
}
return fd;
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = p7_CreateDefaultApp(options, 0, argc, argv, banner, usage);
ESL_RANDOMNESS *rng = esl_randomness_CreateFast(esl_opt_GetInteger(go, "-s"));
fprintf(stderr, "## %s\n", argv[0]);
fprintf(stderr, "# rng seed = %" PRIu32 "\n", esl_randomness_GetSeed(rng));
utest_ReadWrite(rng);
utest_alphabet_config(eslAMINO);
utest_alphabet_config(eslDNA);
utest_alphabet_config(eslRNA);
utest_alphabet_config(eslCOINS);
utest_alphabet_config(eslDICE);
esl_randomness_Destroy(rng);
esl_getopts_Destroy(go);
fprintf(stderr, "# status = ok\n");
return 0;
}
int
main(int argc, char **argv)
{
ESL_GETOPTS *go = esl_getopts_CreateDefaultApp(options, 1, argc, argv, banner, usage);
ESL_STOPWATCH *w = esl_stopwatch_Create();
ESL_RECORDER *rc = NULL;
char *filename = esl_opt_GetArg(go, 1);
int N = esl_opt_GetInteger(go, "-N");
FILE *fp = NULL;
char *buf = NULL;
int balloc = 0;
int status;
if ((fp = fopen(filename, "r")) == NULL) esl_fatal("no such file %s\n", filename);
rc = esl_recorder_Create(fp, N);
esl_stopwatch_Start(w);
while ((status = esl_recorder_Read(rc, &buf)) == eslOK);
esl_recorder_Destroy(rc);
fclose(fp);
esl_stopwatch_Stop(w);
esl_stopwatch_Display(stdout, w, "recorder time: ");
if ((fp = fopen(filename, "r")) == NULL) esl_fatal("no such file %s\n", filename);
esl_stopwatch_Start(w);
while ((status = esl_fgets(&buf, &balloc, fp)) == eslOK);
free(buf);
fclose(fp);
esl_stopwatch_Stop(w);
esl_stopwatch_Display(stdout, w, "esl_fgets() time: ");
esl_stopwatch_Destroy(w);
esl_getopts_Destroy(go);
return 0;
}
static void
utest_FLogsumError(ESL_GETOPTS *go, ESL_RANDOMNESS *r)
{
int N = esl_opt_GetInteger(go, "-N");
float maxval = esl_opt_GetReal(go, "-S");
int be_verbose = esl_opt_GetBoolean(go, "-v");
float maxerr = 0.0;
float avgerr = 0.0;
int i;
float a,b,result,exact,err;
for (i = 0; i < N; i++)
{
a = (esl_random(r) - 0.5) * maxval * 2.; /* uniform draws on -maxval..maxval */
b = (esl_random(r) - 0.5) * maxval * 2.;
exact = log(exp(a) + exp(b));
result = p7_FLogsum(a,b);
err = fabs(exact-result) / maxval;
avgerr += err;
maxerr = ESL_MAX(maxerr, err);
if (be_verbose)
printf("%8.4f %8.4f %8.4f %8.4f %8.4f\n", a, b, exact, result, err);
}
avgerr /= (float) N;
if (be_verbose) {
printf("average error = %f\n", avgerr);
printf("max error = %f\n", maxerr);
}
if (maxerr > 0.0001) esl_fatal("maximum error of %f is too high: logsum unit test fails", maxerr);
if (avgerr > 0.0001) esl_fatal("average error of %f is too high: logsum unit test fails", avgerr);
}
