-/* $Id: icu_I18N.c,v 1.8 2007-05-09 14:01:21 marc Exp $
+/* $Id: icu_I18N.c,v 1.22 2007-05-25 13:27:21 marc Exp $
Copyright (c) 2006-2007, Index Data.
This file is part of Pazpar2.
};
+struct icu_buf_utf16 * icu_buf_utf16_copy(struct icu_buf_utf16 * dest16,
+ struct icu_buf_utf16 * src16)
+{
+ if(!dest16 || !src16
+ || dest16 == src16)
+ return 0;
+
+ if (dest16->utf16_cap < src16->utf16_len)
+ icu_buf_utf16_resize(dest16, src16->utf16_len * 2);
+
+ u_strncpy(dest16->utf16, src16->utf16, src16->utf16_len);
+ dest16->utf16_len = src16->utf16_len;
+
+ return dest16;
+};
+
+
void icu_buf_utf16_destroy(struct icu_buf_utf16 * buf16)
{
if (buf16){
};
+struct icu_buf_utf8 * icu_buf_utf8_copy(struct icu_buf_utf8 * dest8,
+ struct icu_buf_utf8 * src8)
+{
+ if(!dest8 || !src8
+ || dest8 == src8)
+ return 0;
+
+
+ if (dest8->utf8_cap < src8->utf8_len)
+ icu_buf_utf8_resize(dest8, src8->utf8_len * 2);
+
+ strncpy((char*) dest8->utf8, (char*) src8->utf8, src8->utf8_len);
+
+ return dest8;
+};
+
+
void icu_buf_utf8_destroy(struct icu_buf_utf8 * buf8)
{
+struct icu_casemap * icu_casemap_create(const char *locale, char action,
+ UErrorCode *status)
+{
+ struct icu_casemap * casemap
+ = (struct icu_casemap *) malloc(sizeof(struct icu_casemap));
+ strcpy(casemap->locale, locale);
+ casemap->action = action;
+
+ switch(casemap->action) {
+ case 'l':
+ break;
+ case 'u':
+ break;
+ case 't':
+ break;
+ case 'f':
+ break;
+ default:
+ icu_casemap_destroy(casemap);
+ return 0;
+ }
+
+ return casemap;
+};
+
+void icu_casemap_destroy(struct icu_casemap * casemap)
+{
+ if (casemap)
+ free(casemap);
+};
+
+
+int icu_casemap_casemap(struct icu_casemap * casemap,
+ struct icu_buf_utf16 * dest16,
+ struct icu_buf_utf16 * src16,
+ UErrorCode *status)
+{
+ if(!casemap)
+ return 0;
+
+ return icu_utf16_casemap(dest16, src16,
+ casemap->locale, casemap->action, status);
+};
+
+
int icu_utf16_casemap(struct icu_buf_utf16 * dest16,
struct icu_buf_utf16 * src16,
const char *locale, char action,
// check for buffer overflow, resize and retry
if (*status == U_BUFFER_OVERFLOW_ERROR
+ && dest16 != src16 // do not resize if in-place conversion
//|| dest16_len > dest16->utf16_cap
){
icu_buf_utf16_resize(dest16, dest16_len * 2);
dest8->utf8_len = 0;
}
- return *status;
+ return sortkey_len;
};
tokenizer->action = action;
tokenizer->bi = 0;
tokenizer->buf16 = 0;
+ tokenizer->token_count = 0;
tokenizer->token_id = 0;
tokenizer->token_start = 0;
tokenizer->token_end = 0;
if (U_SUCCESS(*status))
return tokenizer;
- // reestablishing zero error state
- //if (*status == U_USING_DEFAULT_WARNING)
- // *status = U_ZERO_ERROR;
-
-
// freeing if failed
- free(tokenizer);
+ icu_tokenizer_destroy(tokenizer);
return 0;
};
void icu_tokenizer_destroy(struct icu_tokenizer * tokenizer)
{
-
if (tokenizer) {
if (tokenizer->bi)
ubrk_close(tokenizer->bi);
if (!tokenizer || !tokenizer->bi || !src16)
return 0;
+
tokenizer->buf16 = src16;
+ tokenizer->token_count = 0;
+ tokenizer->token_id = 0;
+ tokenizer->token_start = 0;
+ tokenizer->token_end = 0;
ubrk_setText(tokenizer->bi, src16->utf16, src16->utf16_len, status);
{
int32_t tkn_start = 0;
int32_t tkn_end = 0;
+ int32_t tkn_len = 0;
if (!tokenizer || !tokenizer->bi
// copy out if everything is well
if(U_FAILURE(*status))
return 0;
-
- tokenizer->token_id++;
+
+ // everything OK, now update internal state
+ tkn_len = tkn_end - tkn_start;
+
+ if (0 < tkn_len){
+ tokenizer->token_count++;
+ tokenizer->token_id++;
+ } else {
+ tokenizer->token_id = 0;
+ }
tokenizer->token_start = tkn_start;
tokenizer->token_end = tkn_end;
+
// copying into token buffer if it exists
if (tkn16){
- if (tkn16->utf16_cap < (tkn_end - tkn_start))
- icu_buf_utf16_resize(tkn16, (size_t) (tkn_end - tkn_start) * 2);
+ if (tkn16->utf16_cap < tkn_len)
+ icu_buf_utf16_resize(tkn16, (size_t) tkn_len * 2);
u_strncpy(tkn16->utf16, &(tokenizer->buf16->utf16)[tkn_start],
- (tkn_end - tkn_start));
+ tkn_len);
- tkn16->utf16_len = (tkn_end - tkn_start);
+ tkn16->utf16_len = tkn_len;
}
- return (tokenizer->token_end - tokenizer->token_start);
+ return tkn_len;
}
+struct icu_normalizer * icu_normalizer_create(const char *rules, char action,
+ UErrorCode *status)
+{
+
+ struct icu_normalizer * normalizer
+ = (struct icu_normalizer *) malloc(sizeof(struct icu_normalizer));
+
+ normalizer->action = action;
+ normalizer->trans = 0;
+ normalizer->rules16 = icu_buf_utf16_create(0);
+ icu_utf16_from_utf8_cstr(normalizer->rules16, rules, status);
+
+ switch(normalizer->action) {
+ case 'f':
+ normalizer->trans
+ = utrans_openU(normalizer->rules16->utf16,
+ normalizer->rules16->utf16_len,
+ UTRANS_FORWARD,
+ 0, 0,
+ normalizer->parse_error, status);
+ // yaz_log(YLOG_LOG, "utrans_open %p", normalizer->trans);
+ break;
+ case 'r':
+ normalizer->trans
+ = utrans_openU(normalizer->rules16->utf16,
+ normalizer->rules16->utf16_len,
+ UTRANS_REVERSE ,
+ 0, 0,
+ normalizer->parse_error, status);
+ // yaz_log(YLOG_LOG, "utrans_open %p", normalizer->trans);
+ break;
+ default:
+ *status = U_UNSUPPORTED_ERROR;
+ return 0;
+ break;
+ }
+
+ if (U_SUCCESS(*status))
+ return normalizer;
+
+ // freeing if failed
+ icu_normalizer_destroy(normalizer);
+ return 0;
+};
+
+
+void icu_normalizer_destroy(struct icu_normalizer * normalizer){
+ if (normalizer) {
+ if (normalizer->rules16)
+ icu_buf_utf16_destroy(normalizer->rules16);
+ if (normalizer->trans)
+ {
+ // yaz_log(YLOG_LOG, "utrans_close %p", normalizer->trans);
+ utrans_close(normalizer->trans);
+ }
+ free(normalizer);
+ }
+};
+
+
+
+int icu_normalizer_normalize(struct icu_normalizer * normalizer,
+ struct icu_buf_utf16 * dest16,
+ struct icu_buf_utf16 * src16,
+ UErrorCode *status)
+{
+ if (!normalizer || !normalizer->trans || !src16 || !dest16)
+ return 0;
+
+ if (!icu_buf_utf16_copy(dest16, src16))
+ return 0;
+
+ utrans_transUChars (normalizer->trans,
+ dest16->utf16, &(dest16->utf16_len),
+ dest16->utf16_cap,
+ 0, &(src16->utf16_len), status);
+
+ if (U_FAILURE(*status)){
+ dest16->utf16[0] = (UChar) 0;
+ dest16->utf16_len = 0;
+ }
+
+ return dest16->utf16_len;
+}
+
+
+
+
+struct icu_chain_step * icu_chain_step_create(struct icu_chain * chain,
+ enum icu_chain_step_type type,
+ const uint8_t * rule,
+ struct icu_buf_utf16 * buf16,
+ UErrorCode *status)
+{
+ struct icu_chain_step * step = 0;
+
+ if(!chain || !type || !rule)
+ return 0;
+
+ step = (struct icu_chain_step *) malloc(sizeof(struct icu_chain_step));
+
+ step->type = type;
+
+ step->buf16 = buf16;
+
+ // create auxilary objects
+ switch(step->type) {
+ case ICU_chain_step_type_display:
+ break;
+ case ICU_chain_step_type_index:
+ break;
+ case ICU_chain_step_type_sortkey:
+ break;
+ case ICU_chain_step_type_casemap:
+ step->u.casemap = icu_casemap_create((char *) chain->locale,
+ (char) rule[0], status);
+ break;
+ case ICU_chain_step_type_normalize:
+ step->u.normalizer = icu_normalizer_create((char *) rule, 'f', status);
+ break;
+ case ICU_chain_step_type_tokenize:
+ step->u.tokenizer = icu_tokenizer_create((char *) chain->locale,
+ (char) rule[0], status);
+ break;
+ default:
+ break;
+ }
+
+ return step;
+};
+
+
+void icu_chain_step_destroy(struct icu_chain_step * step){
+
+ if (!step)
+ return;
+
+ icu_chain_step_destroy(step->previous);
+
+ switch(step->type) {
+ case ICU_chain_step_type_display:
+ break;
+ case ICU_chain_step_type_index:
+ break;
+ case ICU_chain_step_type_sortkey:
+ break;
+ case ICU_chain_step_type_casemap:
+ icu_casemap_destroy(step->u.casemap);
+ icu_buf_utf16_destroy(step->buf16);
+ break;
+ case ICU_chain_step_type_normalize:
+ icu_normalizer_destroy(step->u.normalizer);
+ icu_buf_utf16_destroy(step->buf16);
+ break;
+ case ICU_chain_step_type_tokenize:
+ icu_tokenizer_destroy(step->u.tokenizer);
+ icu_buf_utf16_destroy(step->buf16);
+ break;
+ default:
+ break;
+ }
+ free(step);
+};
+
+
+
+struct icu_chain * icu_chain_create(const uint8_t * identifier,
+ const uint8_t * locale)
+{
+
+ struct icu_chain * chain
+ = (struct icu_chain *) malloc(sizeof(struct icu_chain));
+
+ strncpy((char *) chain->identifier, (const char *) identifier, 128);
+ chain->identifier[128 - 1] = '\0';
+ strncpy((char *) chain->locale, (const char *) locale, 16);
+ chain->locale[16 - 1] = '\0';
+
+ chain->token_count = 0;
+
+ chain->display8 = icu_buf_utf8_create(0);
+ chain->norm8 = icu_buf_utf8_create(0);
+ chain->sort8 = icu_buf_utf8_create(0);
+
+ chain->src16 = icu_buf_utf16_create(0);
+
+ chain->steps = 0;
+
+ return chain;
+};
+
+
+void icu_chain_destroy(struct icu_chain * chain)
+{
+ if (chain){
+ icu_buf_utf8_destroy(chain->display8);
+ icu_buf_utf8_destroy(chain->norm8);
+ icu_buf_utf8_destroy(chain->sort8);
+
+ icu_buf_utf16_destroy(chain->src16);
+
+ icu_chain_step_destroy(chain->steps);
+ free(chain);
+ }
+};
+
+
+
+struct icu_chain * icu_chain_xml_config(xmlNode *xml_node,
+ UErrorCode * status){
+
+ xmlNode *node = 0;
+ struct icu_chain * chain = 0;
+
+ if (!xml_node
+ ||xml_node->type != XML_ELEMENT_NODE
+ || strcmp((const char *) xml_node->name, "icu_chain"))
+
+ return 0;
+
+ xmlChar *xml_id = xmlGetProp(xml_node, (xmlChar *) "id");
+ xmlChar *xml_locale = xmlGetProp(xml_node, (xmlChar *) "locale");
+
+ if (!xml_id || !strlen((const char *) xml_id)
+ || !xml_locale || !strlen((const char *) xml_locale))
+ return 0;
+
+ chain = icu_chain_create((const uint8_t *) xml_id,
+ (const uint8_t *) xml_locale);
+
+ xmlFree(xml_id);
+ xmlFree(xml_locale);
+ if (!chain)
+ return 0;
+
+ for (node = xml_node->children; node; node = node->next)
+ {
+ if (node->type != XML_ELEMENT_NODE)
+ continue;
+
+ xmlChar *xml_rule = xmlGetProp(node, (xmlChar *) "rule");
+ struct icu_chain_step * step = 0;
+
+ if (!strcmp((const char *) node->name,
+ (const char *) "casemap")){
+ step = icu_chain_insert_step(chain, ICU_chain_step_type_casemap,
+ (const uint8_t *) xml_rule, status);
+ }
+ else if (!strcmp((const char *) node->name,
+ (const char *) "normalize")){
+ step = icu_chain_insert_step(chain, ICU_chain_step_type_normalize,
+ (const uint8_t *) xml_rule, status);
+ }
+ else if (!strcmp((const char *) node->name,
+ (const char *) "tokenize")){
+ step = icu_chain_insert_step(chain, ICU_chain_step_type_tokenize,
+ (const uint8_t *) xml_rule, status);
+ }
+ else if (!strcmp((const char *) node->name,
+ (const char *) "display")){
+ step = icu_chain_insert_step(chain, ICU_chain_step_type_display,
+ (const uint8_t *) "", status);
+ }
+ else if (!strcmp((const char *) node->name,
+ (const char *) "index")){
+ step = icu_chain_insert_step(chain, ICU_chain_step_type_index,
+ (const uint8_t *) "", status);
+ }
+ else if (!strcmp((const char *) node->name,
+ (const char *) "sortkey")){
+ step = icu_chain_insert_step(chain, ICU_chain_step_type_sortkey,
+ (const uint8_t *) "", status);
+ }
+
+ xmlFree(xml_rule);
+ if (!step || U_FAILURE(*status)){
+ icu_chain_destroy(chain);
+ return 0;
+ }
+
+
+ }
+
+ return chain;
+};
+
+
+
+struct icu_chain_step * icu_chain_insert_step(struct icu_chain * chain,
+ enum icu_chain_step_type type,
+ const uint8_t * rule,
+ UErrorCode *status)
+{
+ struct icu_chain_step * step = 0;
+ struct icu_buf_utf16 * src16 = 0;
+ struct icu_buf_utf16 * buf16 = 0;
+
+ if (!chain || !type || !rule)
+ return 0;
+
+ // assign utf16 src buffers as needed
+ if (chain->steps && chain->steps->buf16)
+ src16 = chain->steps->buf16;
+ else if (chain->src16)
+ src16 = chain->src16;
+ else
+ return 0;
+
+
+ // create utf16 destination buffers as needed, or
+ switch(type) {
+ case ICU_chain_step_type_display:
+ buf16 = src16;
+ break;
+ case ICU_chain_step_type_index:
+ buf16 = src16;
+ break;
+ case ICU_chain_step_type_sortkey:
+ buf16 = src16;
+ break;
+ case ICU_chain_step_type_casemap:
+ buf16 = icu_buf_utf16_create(0);
+ break;
+ case ICU_chain_step_type_normalize:
+ buf16 = icu_buf_utf16_create(0);
+ break;
+ case ICU_chain_step_type_tokenize:
+ buf16 = icu_buf_utf16_create(0);
+ break;
+ default:
+ break;
+ }
+
+ // create actual chain step with this buffer
+ step = icu_chain_step_create(chain, type, rule, buf16, status);
+
+ step->previous = chain->steps;
+ chain->steps = step;
+
+ return step;
+};
+
+
+int icu_chain_step_next_token(struct icu_chain * chain,
+ struct icu_chain_step * step,
+ UErrorCode *status)
+{
+ struct icu_buf_utf16 * src16 = 0;
+
+ //printf("icu_chain_step_next_token %d\n", (int) step);
+
+ if (!chain || !chain->src16 || !step || !step->more_tokens)
+ return 0;
+
+ // assign utf16 src buffers as neeed, advance in previous steps
+ // tokens until non-zero token met, and setting stop condition
+ if (step->previous){
+ src16 = step->previous->buf16;
+ if (step->need_new_token)
+ //while (step->more_tokens && !src16->utf16_len)
+ step->more_tokens
+ = icu_chain_step_next_token(chain, step->previous, status);
+ }
+ else { // first step can only work once on chain->src16 input buffer
+ src16 = chain->src16;
+ step->more_tokens = 1;
+ }
+
+ // stop if nothing to process
+ // i.e new token source was not properly assigned
+ if (!step->more_tokens || !src16) // || !src16->utf16_len
+ return 0;
+
+ //printf("icu_chain_step_next_token %d working\n", (int) step);
+
+
+ // perform the work, eventually put this steps output in
+ // step->buf16 or the chains UTF8 output buffers
+ switch(step->type) {
+ case ICU_chain_step_type_display:
+ icu_utf16_to_utf8(chain->display8, src16, status);
+ break;
+ case ICU_chain_step_type_index:
+ icu_utf16_to_utf8(chain->norm8, src16, status);
+ break;
+ case ICU_chain_step_type_sortkey:
+ icu_utf16_to_utf8(chain->sort8, src16, status);
+ break;
+ case ICU_chain_step_type_casemap:
+ icu_casemap_casemap(step->u.casemap,
+ step->buf16, src16, status);
+ break;
+ case ICU_chain_step_type_normalize:
+ icu_normalizer_normalize(step->u.normalizer,
+ step->buf16, src16, status);
+ break;
+ case ICU_chain_step_type_tokenize:
+ // attach to new src16 token only first time during splitting
+ if (step->need_new_token){
+ icu_tokenizer_attach(step->u.tokenizer, src16, status);
+ step->need_new_token = 0;
+ }
+ // splitting one src16 token into multiple buf16 tokens
+ step->more_tokens
+ = icu_tokenizer_next_token(step->u.tokenizer,
+ step->buf16, status);
+ // make sure to get new previous token if this one had been used up
+ if (step->previous && !step->more_tokens){
+ if (icu_chain_step_next_token(chain, step->previous, status)){
+ icu_tokenizer_attach(step->u.tokenizer, src16, status);
+ step->need_new_token = 0;
+ step->more_tokens
+ = icu_tokenizer_next_token(step->u.tokenizer,
+ step->buf16, status);
+ }
+ }
+ if (0 == step->more_tokens)
+ return 0;
+ break;
+ default:
+ return 0;
+ break;
+ }
+
+
+
+ // stop further token processing if last step and
+ // new tokens are needed from previous (non-existing) step
+ if (!step->previous && step->need_new_token)
+ step->more_tokens = 0;
+
+ //printf("%d %d %d\n",
+ // step->more_tokens, src16->utf16_len, step->buf16->utf16_len);
+
+
+ if (U_FAILURE(*status))
+ return 0;
+
+ return 1;
+};
+
+
+
+int icu_chain_assign_cstr(struct icu_chain * chain,
+ const char * src8cstr,
+ UErrorCode *status)
+{
+ struct icu_chain_step * stp = 0;
+
+ if (!chain || !src8cstr)
+ return 0;
+
+ stp = chain->steps;
+
+ // clear token count
+ chain->token_count = 0;
+
+ // clear all steps stop states
+
+ while (stp){
+ stp->more_tokens = 1;
+ stp->need_new_token = 1;
+ stp = stp->previous;
+ }
+
+ // finally convert UTF8 to UTF16 string
+ icu_utf16_from_utf8_cstr(chain->src16, src8cstr, status);
+
+ if (U_FAILURE(*status))
+ return 0;
+
+ return 1;
+};
+
+
+
+int icu_chain_next_token(struct icu_chain * chain,
+ UErrorCode *status)
+{
+ int success = 0;
+
+ if (!chain || !chain->steps)
+ return 0;
+
+ success = icu_chain_step_next_token(chain, chain->steps, status);
+
+ if (success){
+ chain->token_count++;
+ return chain->token_count;
+ }
+
+ return 0;
+};
+
+int icu_chain_get_token_count(struct icu_chain * chain)
+{
+ if (!chain)
+ return 0;
+
+ return chain->token_count;
+};
+
+
+
+const char * icu_chain_get_display(struct icu_chain * chain)
+{
+ if (chain->display8)
+ return (const char *) chain->display8->utf8;
+
+ return 0;
+};
+
+const char * icu_chain_get_norm(struct icu_chain * chain)
+{
+ if (chain->norm8)
+ return (const char *) chain->norm8->utf8;
+
+ return 0;
+};
+
+const char * icu_chain_get_sort(struct icu_chain * chain)
+{
+ if (chain->sort8)
+ return (const char *) chain->sort8->utf8;
+
+ return 0;
+};
+
+
+
#endif // HAVE_ICU