static int gt_condenseq_compress_runner(GT_UNUSED int argc, const char **argv,
int parsed_args, void *tool_arguments,
GtError *err)
{
GtCondenseqCompressArguments *arguments = tool_arguments;
GtLogger *logger,
*kdb_logger;
FILE *kmer_fp = NULL;
int had_err = 0;
gt_error_check(err);
gt_assert(arguments);
logger = gt_logger_new(arguments->verbose, GT_LOGGER_DEFLT_PREFIX, stderr);
kdb_logger = gt_logger_new(arguments->kdb, GT_LOGGER_DEFLT_PREFIX, stderr);
if (arguments->kdb) {
kmer_fp = gt_fa_fopen("kmer_db.out", "w", err);
gt_logger_set_target(kdb_logger, kmer_fp);
}
if (gt_str_length(arguments->indexname) == 0UL) {
char *basenameptr;
basenameptr = gt_basename(argv[parsed_args]);
gt_str_set(arguments->indexname, basenameptr);
gt_free(basenameptr);
}
if (!had_err) {
GtEncseqLoader *es_l = gt_encseq_loader_new();
arguments->input_es = gt_encseq_loader_load(es_l, argv[parsed_args], err);
if (arguments->input_es == NULL)
had_err = -1;
gt_encseq_loader_delete(es_l);
}
if (!had_err) {
if (arguments->minalignlength == GT_UNDEF_UWORD)
arguments->minalignlength = arguments->initsize != GT_UNDEF_UWORD ?
arguments->initsize / (GtUword) 3UL :
GT_UNDEF_UWORD;
if (arguments->windowsize == GT_UNDEF_UINT)
arguments->windowsize = arguments->minalignlength != GT_UNDEF_UWORD ?
(unsigned int) (arguments->minalignlength / 5U) :
GT_UNDEF_UINT;
if (arguments->windowsize < 4U)
arguments->windowsize = 4U;
if (arguments->kmersize == GT_UNDEF_UINT) {
unsigned int size =
gt_alphabet_num_of_chars(gt_encseq_alphabet(arguments->input_es));
/* size^k ~= 100000 */
gt_safe_assign(arguments->kmersize,
gt_round_to_long(gt_log_base(100000.0, (double) size)));
gt_logger_log(logger, "|A|: %u, k: %u",
size, arguments->kmersize);
}
if (arguments->windowsize == GT_UNDEF_UINT) {
arguments->windowsize = 5U * arguments->kmersize;
}
if (arguments->minalignlength == GT_UNDEF_UWORD) {
arguments->minalignlength = (GtUword) (3UL * arguments->windowsize);
}
if (arguments->initsize == GT_UNDEF_UWORD) {
arguments->initsize = (GtUword) (3UL * arguments->minalignlength);
}
}
if (!had_err &&
arguments->windowsize <= arguments->kmersize) {
gt_error_set(err, "-windowsize (%u) must be larger -kmersize (%u)!",
arguments->windowsize, arguments->kmersize);
had_err = -1;
}
if (!had_err &&
arguments->minalignlength < (GtUword) arguments->windowsize) {
gt_error_set(err, "-alignlength (" GT_WU ") must be at least "
"-windowsize (%u)!", arguments->minalignlength,
arguments->windowsize);
had_err = -1;
}
if (!had_err && (arguments->initsize < arguments->minalignlength)) {
gt_error_set(err, "-initsize (" GT_WU ") must be at least "
"-alignlength (" GT_WU ")!", arguments->initsize,
arguments->minalignlength);
had_err = -1;
}
if (!had_err) {
GtCondenseqCreator *ces_c;
if (!had_err) {
ces_c = gt_condenseq_creator_new(arguments->initsize,
arguments->minalignlength,
arguments->xdrop,
&(arguments->scores),
arguments->kmersize,
arguments->windowsize,
logger,
err);
if (ces_c == NULL)
had_err = -1;
}
if (!had_err) {
if (arguments->cutoff_value == GT_UNDEF_UWORD)
gt_condenseq_creator_use_mean_cutoff(ces_c);
else if (arguments->cutoff_value == 0)
gt_condenseq_creator_disable_cutoff(ces_c);
else
gt_condenseq_creator_set_cutoff(ces_c, arguments->cutoff_value);
gt_condenseq_creator_set_mean_fraction(ces_c, arguments->fraction);
if (arguments->prune)
gt_condenseq_creator_disable_prune(ces_c);
if (arguments->brute)
gt_condenseq_creator_enable_brute_force(ces_c);
if (!arguments->diags)
gt_condenseq_creator_disable_diagonals(ces_c);
if (arguments->full_diags)
gt_condenseq_creator_enable_full_diagonals(ces_c);
if (arguments->clean_percent != GT_UNDEF_UINT)
gt_condenseq_creator_set_diags_clean_limit(ces_c,
arguments->clean_percent);
had_err = gt_condenseq_creator_create(ces_c,
arguments->indexname,
arguments->input_es,
logger, kdb_logger, err);
gt_condenseq_creator_delete(ces_c);
}
}
gt_logger_delete(logger);
gt_logger_delete(kdb_logger);
if (arguments->kdb)
gt_fa_fclose(kmer_fp);
return had_err;
}
static int gt_kmer_database_runner(GT_UNUSED int argc, const char **argv,
int parsed_args, void *tool_arguments,
GtError *err)
{
GtKmerDatabaseArguments *arguments = tool_arguments;
int had_err = 0;
GtEncseq *es;
GtUword es_length,
nu_kmer_codes = 0;
GtKmerDatabase *compare_db = NULL,
*db = NULL;
GtLogger *logger;
FILE *fp = NULL;
GtHashmap *kmer_hash = NULL;
GtTimer *timer = NULL;
gt_error_check(err);
gt_assert(arguments);
if (arguments->use_hash)
kmer_hash = gt_hashmap_new(GT_HASH_DIRECT, NULL,
(GtFree) gt_kmer_database_delete_hash_value);
if (arguments->bench)
timer = gt_timer_new_with_progress_description("loading encoded sequence");
logger = gt_logger_new(arguments->verbose, GT_LOGGER_DEFLT_PREFIX, stderr);
if (arguments->verbose && gt_str_length(arguments->print_filename) > 0UL) {
fp = gt_fa_fopen(gt_str_get(arguments->print_filename), "w", err);
gt_logger_set_target(logger, fp);
}
if (!had_err) {
GtEncseqLoader *es_l;
if (arguments->bench)
gt_timer_start(timer);
es_l = gt_encseq_loader_new();
es = gt_encseq_loader_load(es_l, argv[parsed_args], err);
if (arguments->bench)
gt_timer_show_progress(timer, "saving kmers (+iterating over file)",
stdout);
if (es == NULL) {
had_err = -1;
}
gt_encseq_loader_delete(es_l);
}
if (!had_err) {
es_length = gt_encseq_total_length(es);
if (es_length < (GtUword) arguments->kmersize) {
gt_error_set(err, "Input is too short for used kmersize. File length: "
GT_WU " kmersize: %u", es_length, arguments->kmersize);
had_err = -1;
}
}
if (!had_err) {
GtAlphabet *alphabet;
alphabet = gt_encseq_alphabet(es);
if (arguments->bench)
nu_kmer_codes = gt_power_for_small_exponents(
gt_alphabet_num_of_chars(alphabet),
arguments->kmersize);
if (!arguments->merge_only && !arguments->use_hash && !arguments->bench) {
compare_db = gt_kmer_database_new(gt_alphabet_num_of_chars(alphabet),
arguments->kmersize, arguments->sb_size, es);
}
if (!arguments->use_hash) {
db = gt_kmer_database_new(gt_alphabet_num_of_chars(alphabet),
arguments->kmersize,
arguments->sb_size, es);
if (arguments->cutoff) {
if (arguments->mean_cutoff)
gt_kmer_database_use_mean_cutoff(db, (GtUword) 2,
arguments->cutoff_value);
else
gt_kmer_database_set_cutoff(db, arguments->cutoff_value);
if (!arguments->prune)
gt_kmer_database_set_prune(db);
}
}
}
if (!had_err) {
GtUword startpos = 0,
endpos;
GtKmercodeiterator *iter;
const GtKmercode *kmercode = NULL;
iter = gt_kmercodeiterator_encseq_new(es, GT_READMODE_FORWARD,
arguments->kmersize, 0);
while (!had_err && startpos < es_length - (arguments->kmersize - 1)) {
GtUword startpos_add_kmer = startpos;
if (arguments->merge_only) {
endpos = startpos + (arguments->kmersize - 1) +
(gt_rand_max((arguments->sb_size - 1) * 2));
if (endpos > es_length)
endpos = es_length;
}
else {
endpos = startpos + (arguments->kmersize - 1) +
(gt_rand_max(arguments->sb_size - 1));
}
gt_kmercodeiterator_reset(iter, GT_READMODE_FORWARD, startpos);
while ((kmercode = gt_kmercodeiterator_encseq_next(iter)) != NULL &&
startpos_add_kmer <= endpos - (arguments->kmersize - 1)) {
if (!arguments->merge_only && !arguments->use_hash &&
!kmercode->definedspecialposition && !arguments->bench) {
gt_kmer_database_add_kmer(compare_db, kmercode->code,
startpos_add_kmer);
}
if (arguments->use_hash && !kmercode->definedspecialposition) {
gt_kmer_database_add_to_hash(kmer_hash, kmercode->code,
startpos_add_kmer);
}
startpos_add_kmer++;
}
if (!arguments->use_hash) {
gt_kmer_database_add_interval(db, startpos, endpos);
gt_kmer_database_print_buffer(db, logger);
if (!arguments->bench)
had_err = gt_kmer_database_check_consistency(db, err);
}
startpos = endpos + 1;
}
if (!arguments->use_hash) {
gt_kmer_database_flush(db);
gt_kmer_database_print_buffer(db, logger);
if (!had_err && !arguments->bench)
had_err = gt_kmer_database_check_consistency(db, err);
if (!arguments->merge_only && !had_err && !arguments->bench)
had_err = gt_kmer_database_check_consistency(compare_db, err);
if (!arguments->merge_only && !arguments->bench)
gt_kmer_database_print(compare_db, logger, true);
if (!arguments->merge_only && !had_err && !arguments->bench)
had_err = gt_kmer_database_compare(compare_db, db, err);
gt_kmer_database_print(db, logger, true);
}
gt_kmercodeiterator_delete(iter);
}
if (arguments->bench) {
GtKmerStartpos pos;
GtArrayGtUword *pos_hash;
GtUword rand_access = (GtUword) 50000000,
rand_code,
i,
sum = 0;
gt_timer_show_progress(timer, "random access", stdout);
for (i = 0; i < rand_access; i++) {
rand_code = gt_rand_max(nu_kmer_codes - 1);
if (arguments->use_hash) {
pos_hash = gt_hashmap_get(kmer_hash, (const void *) rand_code);
if (pos_hash != NULL)
sum += pos_hash->spaceGtUword[pos_hash->nextfreeGtUword - 1];
}
else {
pos = gt_kmer_database_get_startpos(db, rand_code);
if (pos.no_positions > 0)
sum += pos.startpos[pos.no_positions - 1];
}
}
printf("sum: " GT_WU "\n", sum);
gt_timer_show_progress(timer, "", stdout);
gt_timer_stop(timer);
gt_timer_delete(timer);
}
if (arguments->use_hash)
gt_hashmap_delete(kmer_hash);
gt_encseq_delete(es);
if (!arguments->use_hash)
gt_kmer_database_delete(db);
if (!arguments->merge_only && !arguments->bench)
gt_kmer_database_delete(compare_db);
gt_logger_delete(logger);
gt_fa_fclose(fp);
return had_err;
}
