X-Git-Url: http://jsfdemo.indexdata.com/?a=blobdiff_plain;f=src%2Ficu_I18N.c;h=d791abadd5f4b6e7e4c439439dc185327e82f83f;hb=d293bd2470d6d93f44c89ca292ae3dfbb22e9b09;hp=76b9ed7d30a3f901faa74fdcd58c3f9b933142c4;hpb=316bd768750e074a7ba30e24a46f7e95d0302c8c;p=pazpar2-moved-to-github.git

diff --git a/src/icu_I18N.c b/src/icu_I18N.c
index 76b9ed7..d791aba 100644
--- a/src/icu_I18N.c
+++ b/src/icu_I18N.c
@@ -1,4 +1,4 @@
-/* $Id: icu_I18N.c,v 1.10 2007-05-11 09:35:50 marc Exp $
+/* $Id: icu_I18N.c,v 1.16 2007-05-16 19:50:01 marc Exp $
    Copyright (c) 2006-2007, Index Data.
 
    This file is part of Pazpar2.
@@ -111,6 +111,23 @@ struct icu_buf_utf16 * icu_buf_utf16_resize(struct icu_buf_utf16 * buf16,
 };
 
 
+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){
@@ -172,6 +189,23 @@ struct icu_buf_utf8 * icu_buf_utf8_resize(struct icu_buf_utf8 * buf8,
 };
 
 
+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)
 {
@@ -404,7 +438,7 @@ UErrorCode icu_sortkey8_from_utf16(UCollator *coll,
         dest8->utf8_len = 0;
     }
 
-    return *status;
+    return sortkey_len;
 };
 
 
@@ -461,13 +495,8 @@ struct icu_tokenizer * icu_tokenizer_create(const char *locale, char action,
     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;
 };
 
@@ -611,8 +640,9 @@ struct icu_normalizer * icu_normalizer_create(const char *rules, char action,
 
     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
@@ -622,14 +652,14 @@ struct icu_normalizer * icu_normalizer_create(const char *rules, char action,
                            0, 0, 
                            normalizer->parse_error, status);
         break;
-/*     case 'b': */
-/*         normalizer->trans */
-/*             = utrans_openU(normalizer->rules16->utf16,  */
-/*                            normalizer->rules16->utf16_len, */
-/*                            UTRANS_BACKWARD, */
-/*                            0, 0,  */
-/*                            normalizer->parse_error, status); */
-/*         break; */
+    case 'r':
+        normalizer->trans
+            = utrans_openU(normalizer->rules16->utf16,
+                           normalizer->rules16->utf16_len,
+                           UTRANS_REVERSE ,
+                           0, 0,
+                           normalizer->parse_error, status);
+        break;
     default:
         *status = U_UNSUPPORTED_ERROR;
         return 0;
@@ -640,21 +670,383 @@ struct icu_normalizer * icu_normalizer_create(const char *rules, char action,
         return normalizer;
 
     // freeing if failed
-    free(normalizer);
+    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)
-            utrans_close (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->more_tokens = 0;
+
+    if (buf16)
+        step->buf16 = buf16;
+    else
+        step->buf16 = 0;
+
+    // create auxilary objects
+    switch(step->type) {
+    case ICU_chain_step_type_display:
+        break;
+    case ICU_chain_step_type_norm:
+        break;
+    case ICU_chain_step_type_sort:
+        break;
+    case ICU_chain_step_type_charmap:
+        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_norm:
+        break;
+    case ICU_chain_step_type_sort:
+        break;
+    case ICU_chain_step_type_charmap:
+        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;
+    }
+
+
+};
+
+
+
+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)
+{
+    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);
+};
+
+
+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;
+
+    
+    // assign utf16 destination buffers as needed, or
+    // re-use previous uft18 buffer if this step does not touch it
+    switch(type) {
+    case ICU_chain_step_type_display:
+        buf16 = src16;
+        break;
+    case ICU_chain_step_type_norm:
+        buf16 = src16;
+        break;
+    case ICU_chain_step_type_sort:
+        buf16 = src16;
+        break;
+    case ICU_chain_step_type_charmap:
+        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, and setting stop condition
+    if (step->previous){
+        src16 = step->previous->buf16;
+        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_norm:
+        icu_utf16_to_utf8(chain->norm8, src16, status);
+        break;
+    case ICU_chain_step_type_sort:
+        icu_utf16_to_utf8(chain->sort8, src16, status);
+        break;
+    case ICU_chain_step_type_charmap:
+        break;
+    case ICU_chain_step_type_normalize:
+        icu_normalizer_normalize(step->u.normalizer,
+                                 step->buf16, src16, status);
+        break;
+    case ICU_chain_step_type_tokenize:
+        icu_tokenizer_attach(step->u.tokenizer, src16, status);
+        step->more_tokens
+            = icu_tokenizer_next_token(step->u.tokenizer,
+                                       step->buf16, status);
+        break;
+    default:
+        return 0;
+        break;
+    }
+    
+
+    // stop further token processing if last step
+    if (!step->previous)
+        step->more_tokens = 0;
+
+
+    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 = chain->steps;
+
+    if (!chain || !src8cstr)
+        return 0;
+    
+    // clear token count
+    chain->token_count = 0;
+
+    // clear all steps stop states
+
+    while (stp){
+        stp->more_tokens = 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