[log] log the offending line if token count does not match in language_classifier_io

This commit is contained in:
Al
2017-04-02 23:47:07 -04:00
parent 964ac15e51
commit 835d851310

View File

@@ -23,7 +23,6 @@ language_classifier_data_set_t *language_classifier_data_set_init(char *filename
return data_set; return data_set;
} }
bool language_classifier_data_set_next(language_classifier_data_set_t *self) { bool language_classifier_data_set_next(language_classifier_data_set_t *self) {
if (self == NULL) return false; if (self == NULL) return false;
@@ -36,12 +35,12 @@ bool language_classifier_data_set_next(language_classifier_data_set_t *self) {
cstring_array *fields = cstring_array_split(line, TAB_SEPARATOR, TAB_SEPARATOR_LEN, &token_count); cstring_array *fields = cstring_array_split(line, TAB_SEPARATOR, TAB_SEPARATOR_LEN, &token_count);
free(line);
if (token_count != LANGUAGE_CLASSIFIER_FILE_NUM_TOKENS) { if (token_count != LANGUAGE_CLASSIFIER_FILE_NUM_TOKENS) {
log_error("Token count did not match, ected %d, got %zu\n", LANGUAGE_CLASSIFIER_FILE_NUM_TOKENS, token_count); log_error("Token count did not match, expected %d, got %zu, line=%s\n", LANGUAGE_CLASSIFIER_FILE_NUM_TOKENS, token_count, line);
} }
free(line);
char *language = cstring_array_get_string(fields, LANGUAGE_CLASSIFIER_FIELD_LANGUAGE); char *language = cstring_array_get_string(fields, LANGUAGE_CLASSIFIER_FIELD_LANGUAGE);
char *country = cstring_array_get_string(fields, LANGUAGE_CLASSIFIER_FIELD_COUNTRY); char *country = cstring_array_get_string(fields, LANGUAGE_CLASSIFIER_FIELD_COUNTRY);
char *address = cstring_array_get_string(fields, LANGUAGE_CLASSIFIER_FIELD_ADDRESS); char *address = cstring_array_get_string(fields, LANGUAGE_CLASSIFIER_FIELD_ADDRESS);