9 #include <yaz/yconfig.h>
11 #include <yaz/stemmer.h>
13 #include <yaz/xmalloc.h>
15 #include <libstemmer.h>
17 #include <unicode/ustring.h> /* some more string fcns*/
18 #include <unicode/uchar.h> /* char names */
20 enum stemmer_implementation {
27 // Required for cloning.
31 struct sb_stemmer *sb_stemmer;
35 const char* yaz_stemmer_lookup_charenc(const char *charenc, const char *rule) {
39 const char* yaz_stemmer_lookup_algorithm(const char *locale, const char *rule) {
43 yaz_stemmer_p yaz_stemmer_snowball_create(const char *locale, const char *rule, UErrorCode *status) {
44 const char *charenc = yaz_stemmer_lookup_charenc(locale, rule);
45 const char *algorithm = yaz_stemmer_lookup_algorithm(locale,rule);
46 struct sb_stemmer *stemmer = sb_stemmer_new(algorithm, charenc);
47 yaz_stemmer_p yaz_stemmer;
48 yaz_log(YLOG_DEBUG, "create snowball stemmer: algoritm %s charenc %s ", algorithm, charenc);
50 *status = U_ARGUMENT_TYPE_MISMATCH;
51 yaz_log(YLOG_DEBUG, "failed to create stemmer. Creating NOP stemmer");
55 yaz_stemmer = xmalloc(sizeof(*yaz_stemmer));
56 yaz_stemmer->implementation = yaz_snowball;
57 yaz_stemmer->locale = xstrdup(locale);
58 yaz_stemmer->rule = xstrdup(rule);
59 yaz_stemmer->sb_stemmer = stemmer;
60 yaz_log(YLOG_DEBUG, "created snowball stemmer: algoritm %s charenc %s ", algorithm, charenc);
64 yaz_stemmer_p yaz_stemmer_create(const char *locale, const char *rule, UErrorCode *status) {
65 *status = U_ZERO_ERROR;
66 // dispatch logic required if more algorithms is implemented.
67 yaz_log(YLOG_DEBUG, "create stemmer: locale %s rule %s ", locale, rule);
68 return yaz_stemmer_snowball_create(locale, rule, status);
71 yaz_stemmer_p yaz_stemmer_clone(yaz_stemmer_p stemmer) {
72 UErrorCode error = U_ZERO_ERROR;
73 return yaz_stemmer_create(stemmer->locale, stemmer->rule, &error);
76 void yaz_stemmer_stem(yaz_stemmer_p stemmer, struct icu_buf_utf16 *dst, struct icu_buf_utf16* src, UErrorCode *status)
78 switch(stemmer->implementation) {
80 struct icu_buf_utf8 *utf8_buf = icu_buf_utf8_create(0);
81 icu_utf16_to_utf8(utf8_buf, src, status);
82 if (*status == U_ZERO_ERROR) {
83 const sb_symbol *cstr = (const sb_symbol*) icu_buf_utf8_to_cstr(utf8_buf);
84 const sb_symbol *sb_symbol = sb_stemmer_stem(stemmer->sb_stemmer, cstr, utf8_buf->utf8_len);
86 icu_buf_utf16_copy(dst, src);
89 const char *cstr = (const char *) sb_symbol;
90 icu_utf16_from_utf8_cstr(dst, cstr , status);
93 icu_buf_utf8_destroy(utf8_buf);
98 // Default return the same as given.
99 icu_buf_utf16_copy(dst, src);
104 void yaz_stemmer_destroy(yaz_stemmer_p stemmer) {
105 switch (stemmer->implementation) {
107 sb_stemmer_delete(stemmer->sb_stemmer);
110 xfree(stemmer->locale);
111 xfree(stemmer->rule);
115 #endif /* YAZ_HAVE_ICU */