int
varnam_is_known_word(varnam* handle, const char* word)
{
int rc;
sqlite3_int64 word_id;
if (handle == NULL || word == NULL)
return 0;
if (!is_words_store_available (handle)) {
return 0;
}
reset_pool (handle);
rc = vwt_get_word_id (handle, word, &word_id);
if (rc != VARNAM_SUCCESS) {
return 0;
}
if (word_id > 0)
return 1;
else
return 0;
}
int
varnam_train(varnam *handle, const char *pattern, const char *word)
{
int rc;
sqlite3_int64 word_id;
reset_pool (handle);
rc = vwt_start_changes (handle);
if (rc != VARNAM_SUCCESS) return rc;
rc = varnam_learn_internal(handle, word, 1);
if (rc != VARNAM_SUCCESS) {
vwt_discard_changes (handle);
return rc;
}
rc = vwt_get_word_id (handle, word, &word_id);
if (rc) return rc;
rc = vwt_persist_pattern (handle, pattern, word_id, false);
if (rc)
return rc;
vwt_end_changes (handle);
return VARNAM_SUCCESS;
}
int
varnam_learn(varnam *handle, const char *word)
{
int rc;
#ifdef _RECORD_EXEC_TIME
V_BEGIN_TIMING
#endif
reset_pool (handle);
if (!is_words_store_available (handle)) {
return VARNAM_ERROR;
}
rc = vwt_start_changes (handle);
if (rc != VARNAM_SUCCESS) return rc;
rc = varnam_learn_internal(handle, word, 1);
if (rc != VARNAM_SUCCESS) {
vwt_discard_changes (handle);
return rc;
}
rc = vwt_end_changes (handle);
if (rc != VARNAM_SUCCESS)
return rc;
#ifdef _RECORD_EXEC_TIME
V_REPORT_TIME_TAKEN("varnam_learn")
#endif
return VARNAM_SUCCESS;
}
/*
* The public function to reset pool.
*/
PJ_DEF(void) pj_pool_reset(pj_pool_t *pool)
{
LOG((pool->obj_name, "reset(): cap=%d, used=%d(%d%%)",
pool->capacity, pj_pool_get_used_size(pool),
pj_pool_get_used_size(pool)*100/pool->capacity));
reset_pool(pool);
}
int
varnam_delete_word(varnam *handle, const char *word)
{
if (handle == NULL || word == NULL) {
return VARNAM_ARGS_ERROR;
}
reset_pool (handle);
return vwt_delete_word (handle, word);
}
int
varnam_export_words(varnam* handle, int words_per_file, const char* out_dir, int export_type,
void (*callback)(int total_words, int processed, const char *current_word))
{
if (handle == NULL || out_dir == NULL || words_per_file <= 0) {
return VARNAM_ARGS_ERROR;
}
reset_pool (handle);
if (export_type == VARNAM_EXPORT_FULL)
return vwt_full_export (handle, words_per_file, out_dir, callback);
else
return vwt_export_words (handle, words_per_file, out_dir, callback);
}
/*
* Destroy the pool.
*/
PJ_DEF(void) pj_pool_destroy_int(pj_pool_t *pool)
{
pj_size_t initial_size;
LOG((pool->obj_name, "destroy(): cap=%d, used=%d(%d%%), block0=%p-%p",
pool->capacity, pj_pool_get_used_size(pool),
pj_pool_get_used_size(pool)*100/pool->capacity,
((pj_pool_block*)pool->block_list.next)->buf,
((pj_pool_block*)pool->block_list.next)->end));
reset_pool(pool);
initial_size = ((pj_pool_block*)pool->block_list.next)->end -
(unsigned char*)pool;
if (pool->factory->policy.block_free)
(*pool->factory->policy.block_free)(pool->factory, pool, initial_size);
}
int
varnam_import_learnings_from_file(varnam *handle, const char *filepath,
void (*onfailure)(const char* line))
{
int rc, filetype = -1;
FILE *infile;
if (handle == NULL || filepath == NULL)
return VARNAM_ARGS_ERROR;
reset_pool (handle);
infile = fopen(filepath, "r");
if (!infile) {
set_last_error (handle, "Couldn't open file '%s' for reading", filepath);
return VARNAM_ERROR;
}
rc = vwt_optimize_for_huge_transaction(handle);
if (rc) {
fclose (infile);
return rc;
}
varnam_log (handle, "Starting to import from %s", filepath);
rc = vwt_start_changes (handle);
if (rc) {
vwt_turn_off_optimization_for_huge_transaction(handle);
fclose (infile);
return rc;
}
filetype = get_file_type (infile);
switch (filetype) {
case _WORDS_IMPORT:
rc = vwt_import_words (handle, infile, onfailure);
if (rc != VARNAM_SUCCESS) {
fclose (infile);
vwt_turn_off_optimization_for_huge_transaction (handle);
return rc;
}
break;
case _PATTERNS_IMPORT:
rc = vwt_import_patterns (handle, infile, onfailure);
if (rc != VARNAM_SUCCESS) {
fclose (infile);
vwt_turn_off_optimization_for_huge_transaction (handle);
return rc;
}
break;
case -1:
set_last_error (handle, "Couldn't read file '%s'. Unknown file type", filepath);
fclose (infile);
return VARNAM_ERROR;
}
fclose (infile);
varnam_log (handle, "Writing changes to disk");
rc = vwt_end_changes (handle);
if (rc != VARNAM_SUCCESS) {
varnam_log (handle, "Writing changes to disk failed");
return rc;
}
varnam_log (handle, "Ensuring file integrity");
rc = vwt_turn_off_optimization_for_huge_transaction(handle);
if (rc) {
varnam_log (handle, "Failed to check file integrity");
return rc;
}
return VARNAM_SUCCESS;
}
int
varnam_learn_from_file(varnam *handle,
const char *filepath,
vlearn_status *status,
void (*callback)(varnam *handle, const char *word, int status_code, void *object),
void *object)
{
int rc;
FILE *infile;
char line_buffer[10000];
strbuf *word;
varray *word_parts;
int confidence;
int parts;
infile = fopen(filepath, "r");
if (!infile) {
set_last_error (handle, "Couldn't open file '%s' for reading.\n", filepath);
return VARNAM_ERROR;
}
if (status != NULL)
{
status->total_words = 0;
status->failed = 0;
}
rc = vwt_optimize_for_huge_transaction(handle);
if (rc) {
fclose (infile);
return rc;
}
/* Learning from file will be mostly new words. Optimizing for that */
v_->_config_mostly_learning_new_words = 1;
varnam_log (handle, "Starting to learn from %s", filepath);
rc = vwt_start_changes (handle);
if (rc) {
vwt_turn_off_optimization_for_huge_transaction(handle);
fclose (infile);
return rc;
}
while (fgets(line_buffer, sizeof(line_buffer), infile))
{
reset_pool (handle);
word = get_pooled_string (handle);
strbuf_add (word, trimwhitespace (line_buffer));
word_parts = strbuf_split (word, handle, ' ');
parts = varray_length (word_parts);
if (parts > 0 && parts <= 2)
{
confidence = 1;
if (parts == 2) {
word = varray_get (word_parts, 1);
confidence = atoi (strbuf_to_s (word));
}
word = varray_get (word_parts, 0);
rc = varnam_learn_internal (handle, strbuf_to_s (word), confidence);
if (rc) {
if (status != NULL) status->failed++;
}
}
else {
rc = VARNAM_ERROR;
if (status != NULL) status->failed++;
}
if (status != NULL) status->total_words++;
if (callback != NULL) callback (handle, strbuf_to_s (word), rc, object);
}
varnam_log (handle, "Writing changes to disk");
rc = vwt_end_changes (handle);
if (rc) {
varnam_log (handle, "Writing changes to disk failed");
}
varnam_log (handle, "Ensuring file integrity");
rc = vwt_turn_off_optimization_for_huge_transaction(handle);
if (rc) {
varnam_log (handle, "Failed to check file integrity");
}
varnam_log (handle, "Compacting file");
rc = vwt_compact_file (handle);
if (rc) return rc;
fclose (infile);
return rc;
}
lock_free_pool(size_t poolsize = 0) { reset_pool(poolsize); }
