void add_postprocessed_string(cstring_array *strings, char *str, normalize_options_t options) {
bool add_string = true;
if (options.roman_numerals) {
char *numex_replaced = replace_numeric_expressions(str, LATIN_LANGUAGE_CODE);
if (numex_replaced != NULL) {
cstring_array_add_string(strings, numex_replaced);
free(numex_replaced);
add_string = false;
}
}
if (add_string) {
cstring_array_add_string(strings, str);
}
}
bool address_dictionary_add_canonical(char *canonical) {
if (address_dict == NULL || address_dict->canonical == NULL) {
log_error(ADDRESS_DICTIONARY_SETUP_ERROR);
return false;
}
cstring_array_add_string(address_dict->canonical, canonical);
return true;
}
void add_normalized_strings_token(cstring_array *strings, char *str, token_t token, normalize_options_t options) {
uint64_t normalize_token_options = get_normalize_token_options(options);
if (token.type != WHITESPACE ) {
bool contains_hyphen = string_contains_hyphen_len(str + token.offset, token.len);
if (!contains_hyphen || token.type == HYPHEN) {
normalize_token(strings, str, token, normalize_token_options);
} else if (is_word_token(token.type)) {
normalize_token(strings, str, token, normalize_token_options);
if (options.replace_word_hyphens) {
normalize_token_options |= NORMALIZE_TOKEN_REPLACE_HYPHENS;
normalize_token(strings, str, token, normalize_token_options);
normalize_token_options ^= NORMALIZE_TOKEN_REPLACE_HYPHENS;
}
if (options.delete_word_hyphens) {
normalize_token_options |= NORMALIZE_TOKEN_DELETE_HYPHENS;
normalize_token(strings, str, token, normalize_token_options);
normalize_token_options ^= NORMALIZE_TOKEN_DELETE_HYPHENS;
}
} else if (is_numeric_token(token.type)) {
normalize_token(strings, str, token, normalize_token_options);
if (options.replace_numeric_hyphens) {
normalize_token_options |= NORMALIZE_TOKEN_REPLACE_HYPHENS;
normalize_token(strings, str, token, normalize_token_options);
normalize_token_options ^= NORMALIZE_TOKEN_REPLACE_HYPHENS;
}
if (options.delete_numeric_hyphens) {
normalize_token_options |= NORMALIZE_TOKEN_DELETE_HYPHENS;
normalize_token(strings, str, token, normalize_token_options);
normalize_token_options ^= NORMALIZE_TOKEN_DELETE_HYPHENS;
}
}
if (is_numeric_token(token.type) && options.split_alpha_from_numeric) {
normalize_token_options |= NORMALIZE_TOKEN_SPLIT_ALPHA_FROM_NUMERIC;
normalize_token(strings, str, token, normalize_token_options);
normalize_token_options ^= NORMALIZE_TOKEN_SPLIT_ALPHA_FROM_NUMERIC;
}
} else {
cstring_array_add_string(strings, " ");
}
}
int main(int argc, char **argv) {
char *filename;
if (argc == 2) {
filename = argv[1];
} else {
filename = DEFAULT_TRANSLITERATION_PATH;
}
FILE *f = fopen(filename, "wb");
if (f == NULL) {
log_error("File could not be opened, ensure directory exists: %s", filename);
exit(1);
}
size_t num_source_transliterators = sizeof(transliterators_source) / sizeof(transliterator_source_t);
char *key;
size_t key_len;
context_type_t pre_context_type;
size_t pre_context_max_len;
char *pre_context;
size_t pre_context_len;
context_type_t post_context_type;
size_t post_context_max_len;
char *post_context;
size_t post_context_len;
char *replacement;
size_t replacement_len;
char *revisit;
size_t revisit_len;
char *group_regex_str;
size_t group_regex_len;
transliteration_module_init();
transliteration_table_t *trans_table = get_transliteration_table();
trie_t *trie = trans_table->trie;
for (int i = 0; i < num_source_transliterators; i++) {
transliterator_source_t trans_source = transliterators_source[i];
size_t trans_name_len = strlen(trans_source.name);
log_info("Doing transliterator: %s\n", trans_source.name);
char_array *trans_key = char_array_from_string(trans_source.name);
char_array_cat(trans_key, NAMESPACE_SEPARATOR_CHAR);
char *trans_name = strdup(trans_source.name);
if (trans_name == NULL) {
log_error("strdup returned NULL on trans_source.name\n");
goto exit_teardown;
}
transliterator_t *trans = transliterator_new(trans_name, trans_source.internal, trans_table->steps->n, trans_source.steps_length);
for (int j = 0; j < trans_source.steps_length; j++) {
transliteration_step_source_t step_source = steps_source[trans_source.steps_start + j];
size_t step_name_len = strlen(step_source.name);
log_debug("Doing step: %s, type=%d\n", step_source.name, step_source.type);
if (!transliteration_table_add_step(trans_table, step_source.type, step_source.name)) {
log_error("Step couldn't be added\n");
goto exit_teardown;
}
if (step_source.type != STEP_RULESET) {
continue;
}
char_array *step_key = char_array_from_string(char_array_get_string(trans_key));
char_array_cat(step_key, step_source.name);
char_array_cat(step_key, NAMESPACE_SEPARATOR_CHAR);
char *step_key_str = char_array_get_string(step_key);
size_t step_key_len = strlen(step_key_str);
for (int k = 0; k < step_source.rules_length; k++) {
transliteration_rule_source_t rule_source = rules_source[step_source.rules_start + k];
key = rule_source.key;
key_len = rule_source.key_len;
pre_context_type = rule_source.pre_context_type;
pre_context_max_len = rule_source.pre_context_max_len;
pre_context = rule_source.pre_context;
pre_context_len = rule_source.pre_context_len;
post_context_type = rule_source.post_context_type;
post_context_max_len = rule_source.post_context_max_len;
post_context = rule_source.post_context;
post_context_len = rule_source.post_context_len;
replacement = rule_source.replacement;
replacement_len = rule_source.replacement_len;
revisit = rule_source.revisit;
revisit_len = rule_source.revisit_len;
group_regex_str = rule_source.group_regex_str;
group_regex_len = rule_source.group_regex_len;
uint32_t data = trans_table->replacements->n;
char_array *rule_key = char_array_from_string(step_key_str);
uint32_t replacement_string_index = cstring_array_num_strings(trans_table->replacement_strings);
cstring_array_add_string_len(trans_table->replacement_strings, replacement, replacement_len);
uint32_t revisit_index = 0;
if (revisit != NULL && revisit_len > 0) {
revisit_index = cstring_array_num_strings(trans_table->revisit_strings);
cstring_array_add_string_len(trans_table->revisit_strings, revisit, revisit_len);
}
group_capture_array *groups = parse_groups(group_regex_str, group_regex_len);
transliteration_replacement_t *trans_repl = transliteration_replacement_new(replacement_string_index, revisit_index, groups);
uint32_t replacement_index = trans_table->replacements->n;
transliteration_replacement_array_push(trans_table->replacements, trans_repl);
int c;
char *token;
log_debug("Doing rule: %s\n", key);
string_tree_t *tree = regex_string_tree(key, key_len);
string_tree_t *pre_context_tree = NULL;
string_tree_iterator_t *pre_context_iter = NULL;
cstring_array *pre_context_strings = NULL;
if (pre_context_type != CONTEXT_TYPE_NONE) {
pre_context_strings = cstring_array_new();
}
if (pre_context_type == CONTEXT_TYPE_REGEX) {
log_debug("pre_context_type == CONTEXT_TYPE_REGEX\n");
pre_context_tree = regex_string_tree(pre_context, pre_context_len);
pre_context_iter = string_tree_iterator_new(pre_context_tree);
char_array *pre_context_perm = char_array_new_size(pre_context_len);
for (; !string_tree_iterator_done(pre_context_iter); string_tree_iterator_next(pre_context_iter)) {
char_array_clear(pre_context_perm);
for (c = 0; c < pre_context_iter->num_tokens; c++) {
token = string_tree_iterator_get_string(pre_context_iter, c);
if (token == NULL || strlen(token) == 0) {
log_warn("pre_token_context is NULL or 0 length: %s\n", token);
}
char_array_cat(pre_context_perm, token);
}
token = char_array_get_string(pre_context_perm);
if (token == NULL || strlen(token) == 0) {
log_warn("pre_perm is NULL or 0 length\n");
}
cstring_array_add_string(pre_context_strings, token);
}
char_array_destroy(pre_context_perm);
string_tree_iterator_destroy(pre_context_iter);
string_tree_destroy(pre_context_tree);
} else if (pre_context_type == CONTEXT_TYPE_STRING) {
if (pre_context == NULL || strlen(pre_context) == 0) {
log_warn("pre_context STRING NULL or 0 length\n");
}
cstring_array_add_string(pre_context_strings, pre_context);
} else if (pre_context_type == CONTEXT_TYPE_WORD_BOUNDARY) {
cstring_array_add_string(pre_context_strings, WORD_BOUNDARY_CHAR);
}
size_t num_pre_context_strings = 0;
if (pre_context_type != CONTEXT_TYPE_NONE) {
num_pre_context_strings = cstring_array_num_strings(pre_context_strings);
log_debug("num_pre_context_strings = %zu\n", num_pre_context_strings);
}
string_tree_t *post_context_tree = NULL;
string_tree_iterator_t *post_context_iter = NULL;
cstring_array *post_context_strings = NULL;
if (post_context_type != CONTEXT_TYPE_NONE) {
post_context_strings = cstring_array_new();
}
if (post_context_type == CONTEXT_TYPE_REGEX) {
log_debug("post_context_type == CONTEXT_TYPE_REGEX\n");
post_context_tree = regex_string_tree(post_context, post_context_len);
post_context_iter = string_tree_iterator_new(post_context_tree);
char_array *post_context_perm = char_array_new_size(post_context_len);
for (; !string_tree_iterator_done(post_context_iter); string_tree_iterator_next(post_context_iter)) {
char_array_clear(post_context_perm);
for (c = 0; c < post_context_iter->num_tokens; c++) {
token = string_tree_iterator_get_string(post_context_iter, c);
if (token == NULL) {
log_error ("post_token_context is NULL\n");
} else if (strlen(token) == 0) {
log_error("post_token_context is 0 length\n");
}
char_array_cat(post_context_perm, token);
}
cstring_array_add_string(post_context_strings, char_array_get_string(post_context_perm));
}
char_array_destroy(post_context_perm);
string_tree_iterator_destroy(post_context_iter);
string_tree_destroy(post_context_tree);
} else if (post_context_type == CONTEXT_TYPE_STRING) {
if (post_context == NULL || strlen(post_context) == 0) {
log_error("post_context STRING NULL or 0 length\n");
}
cstring_array_add_string(post_context_strings, post_context);
} else if (post_context_type == CONTEXT_TYPE_WORD_BOUNDARY) {
cstring_array_add_string(post_context_strings, WORD_BOUNDARY_CHAR);
}
size_t num_post_context_strings = 0;
if (post_context_type != CONTEXT_TYPE_NONE) {
num_post_context_strings = cstring_array_num_strings(post_context_strings);
log_debug("num_post_context_strings = %zu\n", num_post_context_strings);
}
cstring_array *context_strings = NULL;
size_t num_context_strings = 0;
char *context_start_char = NULL;
bool combined_context_strings = false;
int ante, post;
if (num_pre_context_strings > 0 && num_post_context_strings > 0) {
context_start_char = PRE_CONTEXT_CHAR;
combined_context_strings = true;
size_t max_string_size = 2 * MAX_UTF8_CHAR_SIZE +
((pre_context_max_len * MAX_UTF8_CHAR_SIZE) *
(post_context_max_len * MAX_UTF8_CHAR_SIZE));
num_context_strings = num_pre_context_strings * num_post_context_strings;
char_array *context = char_array_new_size(max_string_size);
context_strings = cstring_array_new_size(num_context_strings * max_string_size + num_context_strings);
for (ante = 0; ante < num_pre_context_strings; ante++) {
char_array_clear(context);
token = cstring_array_get_string(pre_context_strings, ante);
if (token == NULL || strlen(token) == 0) {
log_error("pre_context token was NULL or 0 length\n");
goto exit_teardown;
}
char_array_cat(context, token);
size_t context_len = strlen(char_array_get_string(context));
for (post = 0; post < num_post_context_strings; post++) {
context->n = context_len;
char_array_cat(context, POST_CONTEXT_CHAR);
token = cstring_array_get_string(post_context_strings, post);
char_array_cat(context, token);
if (token == NULL || strlen(token) == 0) {
log_error("post_context token was NULL or 0 length\n");
goto exit_teardown;
}
token = char_array_get_string(context);
cstring_array_add_string(context_strings, token);
}
}
char_array_destroy(context);
} else if (num_pre_context_strings > 0) {
context_start_char = PRE_CONTEXT_CHAR;
num_context_strings = num_pre_context_strings;
context_strings = pre_context_strings;
} else if (num_post_context_strings > 0) {
context_start_char = POST_CONTEXT_CHAR;
num_context_strings = num_post_context_strings;
context_strings = post_context_strings;
}
if (num_context_strings > 0) {
log_debug("num_context_strings = %zu\n", num_context_strings);
}
if (tree == NULL) {
log_error("Tree was NULL, rule=%s\n", key);
goto exit_teardown;
}
string_tree_iterator_t *iter = string_tree_iterator_new(tree);
log_debug("iter->remaining=%d\n", iter->remaining);
char *key_str;
for (; !string_tree_iterator_done(iter); string_tree_iterator_next(iter)) {
rule_key->n = step_key_len;
for (c = 0; c < iter->num_tokens; c++) {
token = string_tree_iterator_get_string(iter, c);
if (token == NULL) {
log_error("string_tree_iterator_get_string was NULL: %s\n", key);
goto exit_teardown;
}
char_array_cat(rule_key, token);
log_debug("string_tree token was %s\n", token);
}
log_debug("rule_key=%s\n", char_array_get_string(rule_key));
size_t context_key_len;
if (num_context_strings == 0) {
token = char_array_get_string(rule_key);
if (trie_get(trie, token) == NULL_NODE_ID) {
trie_add(trie, token, replacement_index);
} else {
log_warn("Key exists: %s, skipping\n", token);
}
} else {
char_array_cat(rule_key, context_start_char);
context_key_len = strlen(char_array_get_string(rule_key));
for (c = 0; c < num_context_strings; c++) {
rule_key->n = context_key_len;
token = cstring_array_get_string(context_strings, c);
if (token == NULL) {
log_error("token was NULL for c=%d\n", c);
}
char_array_cat(rule_key, token);
token = char_array_get_string(rule_key);
if (trie_get(trie, token) == NULL_NODE_ID) {
trie_add(trie, token, replacement_index);
} else {
log_warn("Key exists: %s, skipping\n", token);
}
}
}
}
string_tree_iterator_destroy(iter);
string_tree_destroy(tree);
char_array_destroy(rule_key);
if (pre_context_strings != NULL) {
cstring_array_destroy(pre_context_strings);
}
if (post_context_strings != NULL) {
cstring_array_destroy(post_context_strings);
}
// Only needed if we created a combined context array
if (combined_context_strings) {
cstring_array_destroy(context_strings);
}
}
char_array_destroy(step_key);
}
char_array_destroy(trans_key);
if (!transliteration_table_add_transliterator(trans)) {
goto exit_teardown;
}
}
size_t num_source_scripts = sizeof(script_transliteration_rules) / sizeof(script_transliteration_rule_t);
for (int i = 0; i < num_source_scripts; i++) {
script_transliteration_rule_t rule = script_transliteration_rules[i];
if (!transliteration_table_add_script_language(rule.script_language, rule.index)) {
goto exit_teardown;
}
transliterator_index_t index = rule.index;
for (int j = index.transliterator_index; j < index.transliterator_index + index.num_transliterators; j++) {
char *trans_name = script_transliterators[j];
if (trans_name == NULL) {
goto exit_teardown;
}
cstring_array_add_string(trans_table->transliterator_names, trans_name);
}
}
transliteration_table_write(f);
fclose(f);
transliteration_module_teardown();
log_info("Done!\n");
exit(EXIT_SUCCESS);
exit_teardown:
log_error("FAIL\n");
transliteration_module_teardown();
exit(EXIT_FAILURE);
}
void add_affix_expansions(string_tree_t *tree, char *str, char *lang, token_t token, phrase_t prefix, phrase_t suffix, normalize_options_t options) {
cstring_array *strings = tree->strings;
bool have_suffix = suffix.len > 0;
bool have_prefix = prefix.len > 0;
address_expansion_array *prefix_expansions = NULL;
address_expansion_array *suffix_expansions = NULL;
address_expansion_t prefix_expansion;
address_expansion_t suffix_expansion;
char_array *key = char_array_new_size(token.len);
char *expansion;
uint64_t num_strings = 0;
char *root_word = NULL;
size_t root_len;
token_t root_token;
cstring_array *root_strings = NULL;
int add_space = 0;
int spaces = 0;
size_t prefix_start, prefix_end, root_end, suffix_start;
if (have_prefix) {
prefix_expansions = get_affix_expansions(key, str, lang, token, prefix, false, options);
if (prefix_expansions == NULL) have_prefix = false;
}
if (have_suffix) {
suffix_expansions = get_affix_expansions(key, str, lang, token, suffix, true, options);
if (suffix_expansions == NULL) have_suffix = false;
}
if (have_prefix && have_suffix) {
for (int i = 0; i < prefix_expansions->n; i++) {
prefix_expansion = prefix_expansions->a[i];
char_array_clear(key);
cat_affix_expansion(key, str, prefix_expansion, token, prefix);
prefix_start = key->n - 1;
add_space = (int)prefix_expansion.separable;
if (prefix.len + suffix.len < token.len && !prefix_expansion.separable) {
add_space = suffix_expansion.separable;
}
for (spaces = 0; spaces <= add_space; spaces++) {
key->n = prefix_start;
if (spaces) {
char_array_cat(key, " ");
}
prefix_end = key->n;
if (prefix.len + suffix.len < token.len) {
root_len = token.len - suffix.len - prefix.len;
root_token = (token_t){token.offset + prefix.len, root_len, token.type};
root_strings = cstring_array_new_size(root_len);
add_normalized_strings_token(root_strings, str, root_token, options);
num_strings = cstring_array_num_strings(root_strings);
for (int j = 0; j < num_strings; j++) {
key->n = prefix_end;
root_word = cstring_array_get_string(root_strings, j);
char_array_cat(key, root_word);
root_end = key->n - 1;
for (int k = 0; k < suffix_expansions->n; k++) {
key->n = root_end;
suffix_expansion = suffix_expansions->a[k];
int add_suffix_space = suffix_expansion.separable;
suffix_start = key->n;
for (int suffix_spaces = 0; suffix_spaces <= add_suffix_space; suffix_spaces++) {
key->n = suffix_start;
if (suffix_spaces) {
char_array_cat(key, " ");
}
cat_affix_expansion(key, str, suffix_expansion, token, suffix);
expansion = char_array_get_string(key);
cstring_array_add_string(strings, expansion);
}
}
}
} else {
for (int j = 0; j < suffix_expansions->n; j++) {
key->n = prefix_end;
suffix_expansion = suffix_expansions->a[j];
cat_affix_expansion(key, str, suffix_expansion, token, suffix);
expansion = char_array_get_string(key);
cstring_array_add_string(tree->strings, expansion);
}
}
}
}
} else if (have_suffix) {
root_len = suffix.start;
root_token = (token_t){token.offset, root_len, token.type};
root_strings = cstring_array_new_size(root_len);
add_normalized_strings_token(root_strings, str, root_token, options);
num_strings = cstring_array_num_strings(root_strings);
for (int j = 0; j < num_strings; j++) {
char_array_clear(key);
root_word = cstring_array_get_string(root_strings, j);
char_array_cat(key, root_word);
root_end = key->n - 1;
for (int k = 0; k < suffix_expansions->n; k++) {
key->n = root_end;
suffix_expansion = suffix_expansions->a[k];
add_space = suffix_expansion.separable;
suffix_start = key->n;
for (int spaces = 0; spaces <= add_space; spaces++) {
key->n = suffix_start;
if (spaces) {
char_array_cat(key, " ");
}
cat_affix_expansion(key, str, suffix_expansion, token, suffix);
expansion = char_array_get_string(key);
cstring_array_add_string(tree->strings, expansion);
}
}
}
} else if (have_prefix) {
root_len = token.len - prefix.len;
root_token = (token_t){token.offset + prefix.len, root_len, token.type};
root_strings = cstring_array_new_size(root_len);
add_normalized_strings_token(root_strings, str, root_token, options);
num_strings = cstring_array_num_strings(root_strings);
for (int j = 0; j < prefix_expansions->n; j++) {
char_array_clear(key);
prefix_expansion = prefix_expansions->a[j];
cat_affix_expansion(key, str, prefix_expansion, token, prefix);
prefix_end = key->n - 1;
add_space = prefix_expansion.separable;
for (int spaces = 0; spaces <= add_space; spaces++) {
key->n = prefix_end;
if (spaces) {
char_array_cat(key, " ");
}
for (int k = 0; k < num_strings; k++) {
root_word = cstring_array_get_string(root_strings, k);
char_array_cat(key, root_word);
expansion = char_array_get_string(key);
cstring_array_add_string(tree->strings, expansion);
}
}
}
}
char_array_destroy(key);
if (root_strings != NULL) {
cstring_array_destroy(root_strings);
}
}
