X-Git-Url: http://git.indexdata.com/?a=blobdiff_plain;f=src%2Fcharsets.c;h=f867b6f4a31aa98cf6b34a1b79e1b0393ee0f48e;hb=1498181abb059f78afc0ad4f7bcea0b5c13c7964;hp=b753c4f776472561717bffdd1fa584ebd05888f3;hpb=e107b0011a295ccc61502d6e5ea79d9125a3fbb4;p=pazpar2-moved-to-github.git diff --git a/src/charsets.c b/src/charsets.c index b753c4f..f867b6f 100644 --- a/src/charsets.c +++ b/src/charsets.c @@ -1,5 +1,5 @@ /* This file is part of Pazpar2. - Copyright (C) 2006-2008 Index Data + Copyright (C) 2006-2010 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 @@ -22,39 +22,42 @@ Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA */ #if HAVE_CONFIG_H -#include "cconfig.h" +#include #endif #include #include #include +#include #include #include +#include #include "charsets.h" #include "normalize7bit.h" -#ifdef HAVE_ICU -#include "icu_I18N.h" -#endif // HAVE_ICU +#if YAZ_HAVE_ICU +#include +#endif /* charset handle */ struct pp2_charset_s { const char *(*token_next_handler)(pp2_relevance_token_t prt); - const char *(*get_sort_handler)(pp2_relevance_token_t prt, int skip); -#ifdef HAVE_ICU + const char *(*get_sort_handler)(pp2_relevance_token_t prt); + int ref_count; +#if YAZ_HAVE_ICU struct icu_chain * icu_chn; UErrorCode icu_sts; -#endif // HAVE_ICU +#endif }; 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, int skip_article); +static const char *pp2_get_sort_ascii(pp2_relevance_token_t prt); -#ifdef HAVE_ICU +#if YAZ_HAVE_ICU static const char *pp2_relevance_token_icu(pp2_relevance_token_t prt); -static const char *pp2_get_sort_icu(pp2_relevance_token_t prt, int skip_article); -#endif // HAVE_ICU +static const char *pp2_get_sort_icu(pp2_relevance_token_t prt); +#endif /* tokenzier handle */ struct pp2_relevance_token_s { @@ -63,16 +66,22 @@ 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 }; pp2_charset_t pp2_charset_create_xml(xmlNode *xml_node) { -#ifdef HAVE_ICU +#if YAZ_HAVE_ICU UErrorCode status = U_ZERO_ERROR; + struct icu_chain *chain = 0; + if (xml_node) + xml_node = xml_node->children; while (xml_node && xml_node->type != XML_ELEMENT_NODE) xml_node = xml_node->next; - struct icu_chain *chain = icu_chain_xml_config(xml_node, &status); + chain = icu_chain_xml_config(xml_node, 1, &status); if (!chain || U_FAILURE(status)){ //xmlDocPtr icu_doc = 0; //xmlChar *xmlstr = 0; @@ -85,19 +94,20 @@ pp2_charset_t pp2_charset_create_xml(xmlNode *xml_node) return 0; } return pp2_charset_create(chain); -#else // HAVE_ICU +#else // YAZ_HAVE_ICU yaz_log(YLOG_FATAL, "Error: ICU support requested with element:\n" "<%s>\n ... \n", xml_node->name, xml_node->name); yaz_log(YLOG_FATAL, - "But no ICU support compiled into pazpar2 server."); - yaz_log(YLOG_FATAL, - "Please install libicu36-dev and icu-doc or similar, " - "re-configure and re-compile"); + "But no ICU support is compiled into the YAZ library."); return 0; -#endif // HAVE_ICU +#endif // YAZ_HAVE_ICU } +void pp2_charset_incref(pp2_charset_t pct) +{ + (pct->ref_count)++; +} pp2_charset_t pp2_charset_create(struct icu_chain * icu_chn) { @@ -105,7 +115,8 @@ pp2_charset_t pp2_charset_create(struct icu_chain * icu_chn) pct->token_next_handler = pp2_relevance_token_a_to_z; pct->get_sort_handler = pp2_get_sort_ascii; -#ifdef HAVE_ICU + pct->ref_count = 1; +#if YAZ_HAVE_ICU pct->icu_chn = 0; if (icu_chn) { @@ -114,17 +125,27 @@ pp2_charset_t pp2_charset_create(struct icu_chain * icu_chn) pct->token_next_handler = pp2_relevance_token_icu; pct->get_sort_handler = pp2_get_sort_icu; } -#endif // HAVE_ICU +#endif // YAZ_HAVE_ICU return pct; } void pp2_charset_destroy(pp2_charset_t pct) { - xfree(pct); + if (pct) + { + assert(pct->ref_count >= 1); + --(pct->ref_count); + if (pct->ref_count == 0) + { +#if YAZ_HAVE_ICU + icu_chain_destroy(pct->icu_chn); +#endif + xfree(pct); + } + } } -pp2_relevance_token_t pp2_relevance_tokenize(pp2_charset_t pct, - const char *buf) +pp2_relevance_token_t pp2_relevance_tokenize(pp2_charset_t pct) { pp2_relevance_token_t prt = xmalloc(sizeof(*prt)); @@ -132,27 +153,59 @@ pp2_relevance_token_t pp2_relevance_tokenize(pp2_charset_t pct, prt->norm_str = wrbuf_alloc(); prt->sort_str = wrbuf_alloc(); - prt->cp = buf; + prt->cp = 0; prt->last_cp = 0; prt->pct = pct; -#ifdef HAVE_ICU +#if YAZ_HAVE_ICU + prt->iter = 0; if (pct->icu_chn) - { - pct->icu_sts = U_ZERO_ERROR; - int ok = 0; - ok = icu_chain_assign_cstr(pct->icu_chn, buf, &pct->icu_sts); - //printf("\nfield ok: %d '%s'\n", ok, buf); - prt->pct = pct; - } -#endif // HAVE_ICU + 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; + char firstword[64]; + char *pout = firstword; + char articles[] = "the den der die des an a "; // must end in space + + while (*p && !isalnum(*(unsigned char *)p)) + p++; + for (; *p && *p != ' ' && pout - firstword < (sizeof(firstword)-2); p++) + *pout++ = tolower(*(unsigned char *)p); + *pout++ = ' '; + *pout++ = '\0'; + if (strstr(articles, firstword)) + buf = p; + } + + wrbuf_rewind(prt->norm_str); + wrbuf_rewind(prt->sort_str); + prt->cp = buf; + prt->last_cp = 0; + +#if YAZ_HAVE_ICU + if (prt->iter) + { + icu_iter_first(prt->iter, buf); + } +#endif // YAZ_HAVE_ICU +} 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) @@ -166,9 +219,9 @@ const char *pp2_relevance_token_next(pp2_relevance_token_t prt) return (prt->pct->token_next_handler)(prt); } -const char *pp2_get_sort(pp2_relevance_token_t prt, int skip) +const char *pp2_get_sort(pp2_relevance_token_t prt) { - return prt->pct->get_sort_handler(prt, skip); + return prt->pct->get_sort_handler(prt); } #define raw_char(c) (((c) >= 'a' && (c) <= 'z') ? (c) : -1) @@ -181,7 +234,7 @@ static const char *pp2_relevance_token_a_to_z(pp2_relevance_token_t prt) int c; /* skip white space */ - while (*cp && (c = raw_char(tolower(*cp))) < 0) + while (*cp && (c = raw_char(tolower(*(const unsigned char *)cp))) < 0) cp++; if (*cp == '\0') { @@ -202,8 +255,7 @@ static const char *pp2_relevance_token_a_to_z(pp2_relevance_token_t prt) return wrbuf_cstr(prt->norm_str); } -static const char *pp2_get_sort_ascii(pp2_relevance_token_t prt, - int skip_article) +static const char *pp2_get_sort_ascii(pp2_relevance_token_t prt) { if (prt->last_cp == 0) return 0; @@ -211,7 +263,7 @@ static const char *pp2_get_sort_ascii(pp2_relevance_token_t prt, { char *tmp = xstrdup(prt->last_cp); char *result = 0; - result = normalize7bit_mergekey(tmp, skip_article); + result = normalize7bit_mergekey(tmp); wrbuf_rewind(prt->sort_str); wrbuf_puts(prt->sort_str, result); @@ -221,34 +273,30 @@ static const char *pp2_get_sort_ascii(pp2_relevance_token_t prt, } -#ifdef HAVE_ICU +#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_get_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, - int skip_article) +static const char *pp2_get_sort_icu(pp2_relevance_token_t prt) { - return icu_chain_get_sort(prt->pct->icu_chn); + return icu_iter_get_sortkey(prt->iter); } -#endif // HAVE_ICU - +#endif // YAZ_HAVE_ICU /* * Local variables: * c-basic-offset: 4 + * c-file-style: "Stroustrup" * indent-tabs-mode: nil * End: * vim: shiftwidth=4 tabstop=8 expandtab */ +