X-Git-Url: http://jsfdemo.indexdata.com/?a=blobdiff_plain;f=src%2Fcharsets.c;h=7bbe102e858788409e5f6d1b35ec70beb6d5145c;hb=c1a0d1230f98bfb0fd88cfc39eec90cbb8782fad;hp=f71f4c85faa12ba277c74c8586abe901eecc5718;hpb=572de728d257c3d2e6b11d0f60713c81b004c406;p=pazpar2-moved-to-github.git

diff --git a/src/charsets.c b/src/charsets.c
index f71f4c8..7bbe102 100644
--- a/src/charsets.c
+++ b/src/charsets.c
@@ -1,5 +1,5 @@
 /* This file is part of Pazpar2.
-   Copyright (C) 2006-2009 Index Data
+   Copyright (C) 2006-2011 Index Data
 
 Pazpar2 is free software; you can redistribute it and/or modify it under
 the terms of the GNU General Public License as published by the Free
@@ -28,6 +28,7 @@ Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
 #include <yaz/xmalloc.h>
 #include <yaz/wrbuf.h>
 #include <yaz/log.h>
+#include <yaz/yaz-version.h>
 #include <ctype.h>
 #include <assert.h>
 #include <string.h>
@@ -39,7 +40,6 @@ Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
 #include <yaz/icu.h>
 #endif
 
-
 /* charset handle */
 struct pp2_charset_s {
     const char *(*token_next_handler)(pp2_relevance_token_t prt);
@@ -51,6 +51,7 @@ struct pp2_charset_s {
 #endif
 };
 
+static const char *pp2_relevance_token_null(pp2_relevance_token_t prt);
 static const char *pp2_relevance_token_a_to_z(pp2_relevance_token_t prt);
 static const char *pp2_get_sort_ascii(pp2_relevance_token_t prt);
 
@@ -66,6 +67,9 @@ struct pp2_relevance_token_s {
     pp2_charset_t pct;  /* our main charset handle (type+config) */
     WRBUF norm_str;     /* normized string we return (temporarily) */
     WRBUF sort_str;     /* sort string we return (temporarily) */
+#if YAZ_HAVE_ICU
+    yaz_icu_iter_t iter;
+#endif
 };
 
 
@@ -106,11 +110,18 @@ void pp2_charset_incref(pp2_charset_t pct)
     (pct->ref_count)++;
 }
 
-pp2_charset_t pp2_charset_create(struct icu_chain * icu_chn)
+pp2_charset_t pp2_charset_create_a_to_z(void)
+{
+    pp2_charset_t pct = pp2_charset_create(0);
+    pct->token_next_handler = pp2_relevance_token_a_to_z;
+    return pct;
+}
+
+pp2_charset_t pp2_charset_create(struct icu_chain *icu_chn)
 {
     pp2_charset_t pct = xmalloc(sizeof(*pct));
 
-    pct->token_next_handler = pp2_relevance_token_a_to_z;
+    pct->token_next_handler = pp2_relevance_token_null;
     pct->get_sort_handler  = pp2_get_sort_ascii;
     pct->ref_count = 1;
 #if YAZ_HAVE_ICU
@@ -142,14 +153,30 @@ void pp2_charset_destroy(pp2_charset_t pct)
     }
 }
 
-pp2_relevance_token_t pp2_relevance_tokenize(pp2_charset_t pct,
-                                             const char *buf,
-                                             int skip_article)
+pp2_relevance_token_t pp2_relevance_tokenize(pp2_charset_t pct)
 {
     pp2_relevance_token_t prt = xmalloc(sizeof(*prt));
 
     assert(pct);
 
+    prt->norm_str = wrbuf_alloc();
+    prt->sort_str = wrbuf_alloc();
+    prt->cp = 0;
+    prt->last_cp = 0;
+    prt->pct = pct;
+
+#if YAZ_HAVE_ICU
+    prt->iter = 0;
+    if (pct->icu_chn)
+        prt->iter = icu_iter_create(pct->icu_chn);
+#endif
+    return prt;
+}
+
+void pp2_relevance_first(pp2_relevance_token_t prt,
+                         const char *buf,
+                         int skip_article)
+{ 
     if (skip_article)
     {
         const char *p = buf;
@@ -167,29 +194,26 @@ pp2_relevance_token_t pp2_relevance_tokenize(pp2_charset_t pct,
             buf = p;
     }
 
-    prt->norm_str = wrbuf_alloc();
-    prt->sort_str = wrbuf_alloc();
+    wrbuf_rewind(prt->norm_str);
+    wrbuf_rewind(prt->sort_str);
     prt->cp = buf;
     prt->last_cp = 0;
-    prt->pct = pct;
 
 #if YAZ_HAVE_ICU
-    if (pct->icu_chn)
+    if (prt->iter)
     {
-        int ok = 0;
-        pct->icu_sts = U_ZERO_ERROR;
-        ok = icu_chain_assign_cstr(pct->icu_chn, buf, &pct->icu_sts);
-        //printf("\nfield ok: %d '%s'\n", ok, buf);
-        prt->pct = pct;
+        icu_iter_first(prt->iter, buf);
     }
 #endif // YAZ_HAVE_ICU
-    return prt;
 }
 
-
 void pp2_relevance_token_destroy(pp2_relevance_token_t prt)
 {
     assert(prt);
+#if YAZ_HAVE_ICU
+    if (prt->iter)
+        icu_iter_destroy(prt->iter);
+#endif
     if(prt->norm_str) 
         wrbuf_destroy(prt->norm_str);
     if(prt->sort_str) 
@@ -256,24 +280,30 @@ static const char *pp2_get_sort_ascii(pp2_relevance_token_t prt)
     }
 }
 
+static const char *pp2_relevance_token_null(pp2_relevance_token_t prt)
+{
+    const char *cp = prt->cp;
+
+    prt->last_cp = *cp ? cp : 0;
+    while (*cp)
+        cp++;
+    prt->cp = cp;
+    return prt->last_cp;
+}
 
 #if YAZ_HAVE_ICU
 static const char *pp2_relevance_token_icu(pp2_relevance_token_t prt)
 {
-    if (icu_chain_next_token(prt->pct->icu_chn, &prt->pct->icu_sts))
+    if (icu_iter_next(prt->iter))
     {
-        if (U_FAILURE(prt->pct->icu_sts))
-        {
-            return 0;
-        }
-        return icu_chain_token_norm(prt->pct->icu_chn);
+        return icu_iter_get_norm(prt->iter);
     }
     return 0;
 }
 
 static const char *pp2_get_sort_icu(pp2_relevance_token_t prt)
 {
-    return icu_chain_token_sortkey(prt->pct->icu_chn);
+    return icu_iter_get_sortkey(prt->iter);
 }
 
 #endif // YAZ_HAVE_ICU