isomorfeus-ferret 0.12.4 → 0.12.5

data/ext/isomorfeus_ferret_ext/stem_UTF_8_yiddish.h

@@ -0,0 +1,6 @@
+/* Generated by Snowball 2.2.0 - https://snowballstem.org/ */
+
+extern struct SN_env * yiddish_UTF_8_create_env(void);
+extern void yiddish_UTF_8_close_env(struct SN_env * z);
+
+extern int yiddish_UTF_8_stem(struct SN_env * z);

data/ext/isomorfeus_ferret_ext/stem_api.c

@@ -2,7 +2,7 @@
 #include <stdlib.h> /* for calloc, free */
 #include "stem_header.h"
 
-extern struct SN_env * SN_create_env(int S_size, int I_size, int B_size)
+extern struct SN_env * SN_create_env(int S_size, int I_size)
 {
     struct SN_env * z = (struct SN_env *) calloc(1, sizeof(struct SN_env));
     if (z == NULL) return NULL;
@@ -27,12 +27,6 @@ extern struct SN_env * SN_create_env(int S_size, int I_size, int B_size)
         if (z->I == NULL) goto error;
     }
 
-    if (B_size)
-    {
-        z->B = (unsigned char *) calloc(B_size, sizeof(unsigned char));
-        if (z->B == NULL) goto error;
-    }
-
     return z;
 error:
     SN_close_env(z, S_size);
@@ -52,7 +46,6 @@ extern void SN_close_env(struct SN_env * z, int S_size)
         free(z->S);
     }
     free(z->I);
-    free(z->B);
     if (z->p) lose_s(z->p);
     free(z);
 }
@@ -63,4 +56,3 @@ extern int SN_set_current(struct SN_env * z, int size, const symbol * s)
     z->c = 0;
     return err;
 }
-

data/ext/isomorfeus_ferret_ext/stem_api.h

@@ -16,11 +16,9 @@ struct SN_env {
     int c; int l; int lb; int bra; int ket;
     symbol * * S;
     int * I;
-    unsigned char * B;
 };
 
-extern struct SN_env * SN_create_env(int S_size, int I_size, int B_size);
+extern struct SN_env * SN_create_env(int S_size, int I_size);
 extern void SN_close_env(struct SN_env * z, int S_size);
 
 extern int SN_set_current(struct SN_env * z, int size, const symbol * s);
-

data/ext/isomorfeus_ferret_ext/stem_header.h

@@ -20,38 +20,42 @@ struct among
     int (* function)(struct SN_env *);
 };
 
-symbol * create_s(void);
-void lose_s(symbol * p);
+extern symbol * create_s(void);
+extern void lose_s(symbol * p);
 
-int skip_utf8(const symbol * p, int c, int lb, int l, int n);
+extern int skip_utf8(const symbol * p, int c, int limit, int n);
 
-int in_grouping_U(struct SN_env * z, const unsigned char * s, int min, int max, int repeat);
-int in_grouping_b_U(struct SN_env * z, const unsigned char * s, int min, int max, int repeat);
-int out_grouping_U(struct SN_env * z, const unsigned char * s, int min, int max, int repeat);
-int out_grouping_b_U(struct SN_env * z, const unsigned char * s, int min, int max, int repeat);
+extern int skip_b_utf8(const symbol * p, int c, int limit, int n);
 
-int in_grouping(struct SN_env * z, const unsigned char * s, int min, int max, int repeat);
-int in_grouping_b(struct SN_env * z, const unsigned char * s, int min, int max, int repeat);
-int out_grouping(struct SN_env * z, const unsigned char * s, int min, int max, int repeat);
-int out_grouping_b(struct SN_env * z, const unsigned char * s, int min, int max, int repeat);
+extern int in_grouping_U(struct SN_env * z, const unsigned char * s, int min, int max, int repeat);
+extern int in_grouping_b_U(struct SN_env * z, const unsigned char * s, int min, int max, int repeat);
+extern int out_grouping_U(struct SN_env * z, const unsigned char * s, int min, int max, int repeat);
+extern int out_grouping_b_U(struct SN_env * z, const unsigned char * s, int min, int max, int repeat);
 
-int eq_s(struct SN_env * z, int s_size, const symbol * s);
-int eq_s_b(struct SN_env * z, int s_size, const symbol * s);
-int eq_v(struct SN_env * z, const symbol * p);
-int eq_v_b(struct SN_env * z, const symbol * p);
+extern int in_grouping(struct SN_env * z, const unsigned char * s, int min, int max, int repeat);
+extern int in_grouping_b(struct SN_env * z, const unsigned char * s, int min, int max, int repeat);
+extern int out_grouping(struct SN_env * z, const unsigned char * s, int min, int max, int repeat);
+extern int out_grouping_b(struct SN_env * z, const unsigned char * s, int min, int max, int repeat);
 
-int find_among(struct SN_env * z, const struct among * v, int v_size);
-int find_among_b(struct SN_env * z, const struct among * v, int v_size);
+extern int eq_s(struct SN_env * z, int s_size, const symbol * s);
+extern int eq_s_b(struct SN_env * z, int s_size, const symbol * s);
+extern int eq_v(struct SN_env * z, const symbol * p);
+extern int eq_v_b(struct SN_env * z, const symbol * p);
 
-int replace_s(struct SN_env * z, int c_bra, int c_ket, int s_size, const symbol * s, int * adjustment);
-int slice_from_s(struct SN_env * z, int s_size, const symbol * s);
-int slice_from_v(struct SN_env * z, const symbol * p);
-int slice_del(struct SN_env * z);
+extern int find_among(struct SN_env * z, const struct among * v, int v_size);
+extern int find_among_b(struct SN_env * z, const struct among * v, int v_size);
 
-int insert_s(struct SN_env * z, int bra, int ket, int s_size, const symbol * s);
-int insert_v(struct SN_env * z, int bra, int ket, const symbol * p);
+extern int replace_s(struct SN_env * z, int c_bra, int c_ket, int s_size, const symbol * s, int * adjustment);
+extern int slice_from_s(struct SN_env * z, int s_size, const symbol * s);
+extern int slice_from_v(struct SN_env * z, const symbol * p);
+extern int slice_del(struct SN_env * z);
 
-symbol * slice_to(struct SN_env * z, symbol * p);
-symbol * assign_to(struct SN_env * z, symbol * p);
+extern int insert_s(struct SN_env * z, int bra, int ket, int s_size, const symbol * s);
+extern int insert_v(struct SN_env * z, int bra, int ket, const symbol * p);
 
-void debug(struct SN_env * z, int number, int line_count);
+extern symbol * slice_to(struct SN_env * z, symbol * p);
+extern symbol * assign_to(struct SN_env * z, symbol * p);
+
+extern int len_utf8(const symbol * p);
+
+extern void debug(struct SN_env * z, int number, int line_count);

data/ext/isomorfeus_ferret_ext/stem_modules.h

@@ -3,11 +3,19 @@
  * This file is generated by mkmodules.pl from a list of module names.
  * Do not edit manually.
  *
- * Modules included by this file are: danish, dutch, english, finnish, french,
- * german, hungarian, italian, norwegian, porter, portuguese, romanian,
- * russian, spanish, swedish, turkish
+ * Modules included by this file are: arabic, armenian, basque, catalan,
+ * danish, dutch, english, finnish, french, german, greek, hindi, hungarian,
+ * indonesian, irish, italian, lithuanian, nepali, norwegian, porter,
+ * portuguese, romanian, russian, serbian, spanish, swedish, tamil, turkish,
+ * yiddish
  */
 
+#include "stem_UTF_8_arabic.h"
+#include "stem_UTF_8_armenian.h"
+#include "stem_ISO_8859_1_basque.h"
+#include "stem_UTF_8_basque.h"
+#include "stem_ISO_8859_1_catalan.h"
+#include "stem_UTF_8_catalan.h"
 #include "stem_ISO_8859_1_danish.h"
 #include "stem_UTF_8_danish.h"
 #include "stem_ISO_8859_1_dutch.h"
@@ -20,10 +28,18 @@
 #include "stem_UTF_8_french.h"
 #include "stem_ISO_8859_1_german.h"
 #include "stem_UTF_8_german.h"
-#include "stem_ISO_8859_1_hungarian.h"
+#include "stem_UTF_8_greek.h"
+#include "stem_UTF_8_hindi.h"
+#include "stem_ISO_8859_2_hungarian.h"
 #include "stem_UTF_8_hungarian.h"
+#include "stem_ISO_8859_1_indonesian.h"
+#include "stem_UTF_8_indonesian.h"
+#include "stem_ISO_8859_1_irish.h"
+#include "stem_UTF_8_irish.h"
 #include "stem_ISO_8859_1_italian.h"
 #include "stem_UTF_8_italian.h"
+#include "stem_UTF_8_lithuanian.h"
+#include "stem_UTF_8_nepali.h"
 #include "stem_ISO_8859_1_norwegian.h"
 #include "stem_UTF_8_norwegian.h"
 #include "stem_ISO_8859_1_porter.h"
@@ -34,11 +50,14 @@
 #include "stem_UTF_8_romanian.h"
 #include "stem_KOI8_R_russian.h"
 #include "stem_UTF_8_russian.h"
+#include "stem_UTF_8_serbian.h"
 #include "stem_ISO_8859_1_spanish.h"
 #include "stem_UTF_8_spanish.h"
 #include "stem_ISO_8859_1_swedish.h"
 #include "stem_UTF_8_swedish.h"
+#include "stem_UTF_8_tamil.h"
 #include "stem_UTF_8_turkish.h"
+#include "stem_UTF_8_yiddish.h"
 
 typedef enum {
   ENC_UNKNOWN=0,
@@ -52,7 +71,7 @@ struct stemmer_encoding {
   const char * name;
   stemmer_encoding_t enc;
 };
-static struct stemmer_encoding encodings[] = {
+static const struct stemmer_encoding encodings[] = {
   {"ISO_8859_1", ENC_ISO_8859_1},
   {"ISO_8859_2", ENC_ISO_8859_2},
   {"KOI8_R", ENC_KOI8_R},
@@ -62,12 +81,27 @@ static struct stemmer_encoding encodings[] = {
 
 struct stemmer_modules {
   const char * name;
-  stemmer_encoding_t enc; 
+  stemmer_encoding_t enc;
   struct SN_env * (*create)(void);
   void (*close)(struct SN_env *);
   int (*stem)(struct SN_env *);
 };
-static struct stemmer_modules modules[] = {
+static const struct stemmer_modules modules[] = {
+  {"ar", ENC_UTF_8, arabic_UTF_8_create_env, arabic_UTF_8_close_env, arabic_UTF_8_stem},
+  {"ara", ENC_UTF_8, arabic_UTF_8_create_env, arabic_UTF_8_close_env, arabic_UTF_8_stem},
+  {"arabic", ENC_UTF_8, arabic_UTF_8_create_env, arabic_UTF_8_close_env, arabic_UTF_8_stem},
+  {"arm", ENC_UTF_8, armenian_UTF_8_create_env, armenian_UTF_8_close_env, armenian_UTF_8_stem},
+  {"armenian", ENC_UTF_8, armenian_UTF_8_create_env, armenian_UTF_8_close_env, armenian_UTF_8_stem},
+  {"baq", ENC_ISO_8859_1, basque_ISO_8859_1_create_env, basque_ISO_8859_1_close_env, basque_ISO_8859_1_stem},
+  {"baq", ENC_UTF_8, basque_UTF_8_create_env, basque_UTF_8_close_env, basque_UTF_8_stem},
+  {"basque", ENC_ISO_8859_1, basque_ISO_8859_1_create_env, basque_ISO_8859_1_close_env, basque_ISO_8859_1_stem},
+  {"basque", ENC_UTF_8, basque_UTF_8_create_env, basque_UTF_8_close_env, basque_UTF_8_stem},
+  {"ca", ENC_ISO_8859_1, catalan_ISO_8859_1_create_env, catalan_ISO_8859_1_close_env, catalan_ISO_8859_1_stem},
+  {"ca", ENC_UTF_8, catalan_UTF_8_create_env, catalan_UTF_8_close_env, catalan_UTF_8_stem},
+  {"cat", ENC_ISO_8859_1, catalan_ISO_8859_1_create_env, catalan_ISO_8859_1_close_env, catalan_ISO_8859_1_stem},
+  {"cat", ENC_UTF_8, catalan_UTF_8_create_env, catalan_UTF_8_close_env, catalan_UTF_8_stem},
+  {"catalan", ENC_ISO_8859_1, catalan_ISO_8859_1_create_env, catalan_ISO_8859_1_close_env, catalan_ISO_8859_1_stem},
+  {"catalan", ENC_UTF_8, catalan_UTF_8_create_env, catalan_UTF_8_close_env, catalan_UTF_8_stem},
   {"da", ENC_ISO_8859_1, danish_ISO_8859_1_create_env, danish_ISO_8859_1_close_env, danish_ISO_8859_1_stem},
   {"da", ENC_UTF_8, danish_UTF_8_create_env, danish_UTF_8_close_env, danish_UTF_8_stem},
   {"dan", ENC_ISO_8859_1, danish_ISO_8859_1_create_env, danish_ISO_8859_1_close_env, danish_ISO_8859_1_stem},
@@ -82,6 +116,8 @@ static struct stemmer_modules modules[] = {
   {"dut", ENC_UTF_8, dutch_UTF_8_create_env, dutch_UTF_8_close_env, dutch_UTF_8_stem},
   {"dutch", ENC_ISO_8859_1, dutch_ISO_8859_1_create_env, dutch_ISO_8859_1_close_env, dutch_ISO_8859_1_stem},
   {"dutch", ENC_UTF_8, dutch_UTF_8_create_env, dutch_UTF_8_close_env, dutch_UTF_8_stem},
+  {"el", ENC_UTF_8, greek_UTF_8_create_env, greek_UTF_8_close_env, greek_UTF_8_stem},
+  {"ell", ENC_UTF_8, greek_UTF_8_create_env, greek_UTF_8_close_env, greek_UTF_8_stem},
   {"en", ENC_ISO_8859_1, english_ISO_8859_1_create_env, english_ISO_8859_1_close_env, english_ISO_8859_1_stem},
   {"en", ENC_UTF_8, english_UTF_8_create_env, english_UTF_8_close_env, english_UTF_8_stem},
   {"eng", ENC_ISO_8859_1, english_ISO_8859_1_create_env, english_ISO_8859_1_close_env, english_ISO_8859_1_stem},
@@ -92,6 +128,10 @@ static struct stemmer_modules modules[] = {
   {"es", ENC_UTF_8, spanish_UTF_8_create_env, spanish_UTF_8_close_env, spanish_UTF_8_stem},
   {"esl", ENC_ISO_8859_1, spanish_ISO_8859_1_create_env, spanish_ISO_8859_1_close_env, spanish_ISO_8859_1_stem},
   {"esl", ENC_UTF_8, spanish_UTF_8_create_env, spanish_UTF_8_close_env, spanish_UTF_8_stem},
+  {"eu", ENC_ISO_8859_1, basque_ISO_8859_1_create_env, basque_ISO_8859_1_close_env, basque_ISO_8859_1_stem},
+  {"eu", ENC_UTF_8, basque_UTF_8_create_env, basque_UTF_8_close_env, basque_UTF_8_stem},
+  {"eus", ENC_ISO_8859_1, basque_ISO_8859_1_create_env, basque_ISO_8859_1_close_env, basque_ISO_8859_1_stem},
+  {"eus", ENC_UTF_8, basque_UTF_8_create_env, basque_UTF_8_close_env, basque_UTF_8_stem},
   {"fi", ENC_ISO_8859_1, finnish_ISO_8859_1_create_env, finnish_ISO_8859_1_close_env, finnish_ISO_8859_1_stem},
   {"fi", ENC_UTF_8, finnish_UTF_8_create_env, finnish_UTF_8_close_env, finnish_UTF_8_stem},
   {"fin", ENC_ISO_8859_1, finnish_ISO_8859_1_create_env, finnish_ISO_8859_1_close_env, finnish_ISO_8859_1_stem},
@@ -106,22 +146,47 @@ static struct stemmer_modules modules[] = {
   {"fre", ENC_UTF_8, french_UTF_8_create_env, french_UTF_8_close_env, french_UTF_8_stem},
   {"french", ENC_ISO_8859_1, french_ISO_8859_1_create_env, french_ISO_8859_1_close_env, french_ISO_8859_1_stem},
   {"french", ENC_UTF_8, french_UTF_8_create_env, french_UTF_8_close_env, french_UTF_8_stem},
+  {"ga", ENC_ISO_8859_1, irish_ISO_8859_1_create_env, irish_ISO_8859_1_close_env, irish_ISO_8859_1_stem},
+  {"ga", ENC_UTF_8, irish_UTF_8_create_env, irish_UTF_8_close_env, irish_UTF_8_stem},
   {"ger", ENC_ISO_8859_1, german_ISO_8859_1_create_env, german_ISO_8859_1_close_env, german_ISO_8859_1_stem},
   {"ger", ENC_UTF_8, german_UTF_8_create_env, german_UTF_8_close_env, german_UTF_8_stem},
   {"german", ENC_ISO_8859_1, german_ISO_8859_1_create_env, german_ISO_8859_1_close_env, german_ISO_8859_1_stem},
   {"german", ENC_UTF_8, german_UTF_8_create_env, german_UTF_8_close_env, german_UTF_8_stem},
-  {"hu", ENC_ISO_8859_1, hungarian_ISO_8859_1_create_env, hungarian_ISO_8859_1_close_env, hungarian_ISO_8859_1_stem},
+  {"gle", ENC_ISO_8859_1, irish_ISO_8859_1_create_env, irish_ISO_8859_1_close_env, irish_ISO_8859_1_stem},
+  {"gle", ENC_UTF_8, irish_UTF_8_create_env, irish_UTF_8_close_env, irish_UTF_8_stem},
+  {"gre", ENC_UTF_8, greek_UTF_8_create_env, greek_UTF_8_close_env, greek_UTF_8_stem},
+  {"greek", ENC_UTF_8, greek_UTF_8_create_env, greek_UTF_8_close_env, greek_UTF_8_stem},
+  {"hi", ENC_UTF_8, hindi_UTF_8_create_env, hindi_UTF_8_close_env, hindi_UTF_8_stem},
+  {"hin", ENC_UTF_8, hindi_UTF_8_create_env, hindi_UTF_8_close_env, hindi_UTF_8_stem},
+  {"hindi", ENC_UTF_8, hindi_UTF_8_create_env, hindi_UTF_8_close_env, hindi_UTF_8_stem},
+  {"hu", ENC_ISO_8859_2, hungarian_ISO_8859_2_create_env, hungarian_ISO_8859_2_close_env, hungarian_ISO_8859_2_stem},
   {"hu", ENC_UTF_8, hungarian_UTF_8_create_env, hungarian_UTF_8_close_env, hungarian_UTF_8_stem},
-  {"hun", ENC_ISO_8859_1, hungarian_ISO_8859_1_create_env, hungarian_ISO_8859_1_close_env, hungarian_ISO_8859_1_stem},
+  {"hun", ENC_ISO_8859_2, hungarian_ISO_8859_2_create_env, hungarian_ISO_8859_2_close_env, hungarian_ISO_8859_2_stem},
   {"hun", ENC_UTF_8, hungarian_UTF_8_create_env, hungarian_UTF_8_close_env, hungarian_UTF_8_stem},
-  {"hungarian", ENC_ISO_8859_1, hungarian_ISO_8859_1_create_env, hungarian_ISO_8859_1_close_env, hungarian_ISO_8859_1_stem},
+  {"hungarian", ENC_ISO_8859_2, hungarian_ISO_8859_2_create_env, hungarian_ISO_8859_2_close_env, hungarian_ISO_8859_2_stem},
   {"hungarian", ENC_UTF_8, hungarian_UTF_8_create_env, hungarian_UTF_8_close_env, hungarian_UTF_8_stem},
+  {"hy", ENC_UTF_8, armenian_UTF_8_create_env, armenian_UTF_8_close_env, armenian_UTF_8_stem},
+  {"hye", ENC_UTF_8, armenian_UTF_8_create_env, armenian_UTF_8_close_env, armenian_UTF_8_stem},
+  {"id", ENC_ISO_8859_1, indonesian_ISO_8859_1_create_env, indonesian_ISO_8859_1_close_env, indonesian_ISO_8859_1_stem},
+  {"id", ENC_UTF_8, indonesian_UTF_8_create_env, indonesian_UTF_8_close_env, indonesian_UTF_8_stem},
+  {"ind", ENC_ISO_8859_1, indonesian_ISO_8859_1_create_env, indonesian_ISO_8859_1_close_env, indonesian_ISO_8859_1_stem},
+  {"ind", ENC_UTF_8, indonesian_UTF_8_create_env, indonesian_UTF_8_close_env, indonesian_UTF_8_stem},
+  {"indonesian", ENC_ISO_8859_1, indonesian_ISO_8859_1_create_env, indonesian_ISO_8859_1_close_env, indonesian_ISO_8859_1_stem},
+  {"indonesian", ENC_UTF_8, indonesian_UTF_8_create_env, indonesian_UTF_8_close_env, indonesian_UTF_8_stem},
+  {"irish", ENC_ISO_8859_1, irish_ISO_8859_1_create_env, irish_ISO_8859_1_close_env, irish_ISO_8859_1_stem},
+  {"irish", ENC_UTF_8, irish_UTF_8_create_env, irish_UTF_8_close_env, irish_UTF_8_stem},
   {"it", ENC_ISO_8859_1, italian_ISO_8859_1_create_env, italian_ISO_8859_1_close_env, italian_ISO_8859_1_stem},
   {"it", ENC_UTF_8, italian_UTF_8_create_env, italian_UTF_8_close_env, italian_UTF_8_stem},
   {"ita", ENC_ISO_8859_1, italian_ISO_8859_1_create_env, italian_ISO_8859_1_close_env, italian_ISO_8859_1_stem},
   {"ita", ENC_UTF_8, italian_UTF_8_create_env, italian_UTF_8_close_env, italian_UTF_8_stem},
   {"italian", ENC_ISO_8859_1, italian_ISO_8859_1_create_env, italian_ISO_8859_1_close_env, italian_ISO_8859_1_stem},
   {"italian", ENC_UTF_8, italian_UTF_8_create_env, italian_UTF_8_close_env, italian_UTF_8_stem},
+  {"lit", ENC_UTF_8, lithuanian_UTF_8_create_env, lithuanian_UTF_8_close_env, lithuanian_UTF_8_stem},
+  {"lithuanian", ENC_UTF_8, lithuanian_UTF_8_create_env, lithuanian_UTF_8_close_env, lithuanian_UTF_8_stem},
+  {"lt", ENC_UTF_8, lithuanian_UTF_8_create_env, lithuanian_UTF_8_close_env, lithuanian_UTF_8_stem},
+  {"ne", ENC_UTF_8, nepali_UTF_8_create_env, nepali_UTF_8_close_env, nepali_UTF_8_stem},
+  {"nep", ENC_UTF_8, nepali_UTF_8_create_env, nepali_UTF_8_close_env, nepali_UTF_8_stem},
+  {"nepali", ENC_UTF_8, nepali_UTF_8_create_env, nepali_UTF_8_close_env, nepali_UTF_8_stem},
   {"nl", ENC_ISO_8859_1, dutch_ISO_8859_1_create_env, dutch_ISO_8859_1_close_env, dutch_ISO_8859_1_stem},
   {"nl", ENC_UTF_8, dutch_UTF_8_create_env, dutch_UTF_8_close_env, dutch_UTF_8_stem},
   {"nld", ENC_ISO_8859_1, dutch_ISO_8859_1_create_env, dutch_ISO_8859_1_close_env, dutch_ISO_8859_1_stem},
@@ -154,37 +219,59 @@ static struct stemmer_modules modules[] = {
   {"rus", ENC_UTF_8, russian_UTF_8_create_env, russian_UTF_8_close_env, russian_UTF_8_stem},
   {"russian", ENC_KOI8_R, russian_KOI8_R_create_env, russian_KOI8_R_close_env, russian_KOI8_R_stem},
   {"russian", ENC_UTF_8, russian_UTF_8_create_env, russian_UTF_8_close_env, russian_UTF_8_stem},
+  {"serbian", ENC_UTF_8, serbian_UTF_8_create_env, serbian_UTF_8_close_env, serbian_UTF_8_stem},
   {"spa", ENC_ISO_8859_1, spanish_ISO_8859_1_create_env, spanish_ISO_8859_1_close_env, spanish_ISO_8859_1_stem},
   {"spa", ENC_UTF_8, spanish_UTF_8_create_env, spanish_UTF_8_close_env, spanish_UTF_8_stem},
   {"spanish", ENC_ISO_8859_1, spanish_ISO_8859_1_create_env, spanish_ISO_8859_1_close_env, spanish_ISO_8859_1_stem},
   {"spanish", ENC_UTF_8, spanish_UTF_8_create_env, spanish_UTF_8_close_env, spanish_UTF_8_stem},
+  {"sr", ENC_UTF_8, serbian_UTF_8_create_env, serbian_UTF_8_close_env, serbian_UTF_8_stem},
+  {"srp", ENC_UTF_8, serbian_UTF_8_create_env, serbian_UTF_8_close_env, serbian_UTF_8_stem},
   {"sv", ENC_ISO_8859_1, swedish_ISO_8859_1_create_env, swedish_ISO_8859_1_close_env, swedish_ISO_8859_1_stem},
   {"sv", ENC_UTF_8, swedish_UTF_8_create_env, swedish_UTF_8_close_env, swedish_UTF_8_stem},
   {"swe", ENC_ISO_8859_1, swedish_ISO_8859_1_create_env, swedish_ISO_8859_1_close_env, swedish_ISO_8859_1_stem},
   {"swe", ENC_UTF_8, swedish_UTF_8_create_env, swedish_UTF_8_close_env, swedish_UTF_8_stem},
   {"swedish", ENC_ISO_8859_1, swedish_ISO_8859_1_create_env, swedish_ISO_8859_1_close_env, swedish_ISO_8859_1_stem},
   {"swedish", ENC_UTF_8, swedish_UTF_8_create_env, swedish_UTF_8_close_env, swedish_UTF_8_stem},
+  {"ta", ENC_UTF_8, tamil_UTF_8_create_env, tamil_UTF_8_close_env, tamil_UTF_8_stem},
+  {"tam", ENC_UTF_8, tamil_UTF_8_create_env, tamil_UTF_8_close_env, tamil_UTF_8_stem},
+  {"tamil", ENC_UTF_8, tamil_UTF_8_create_env, tamil_UTF_8_close_env, tamil_UTF_8_stem},
   {"tr", ENC_UTF_8, turkish_UTF_8_create_env, turkish_UTF_8_close_env, turkish_UTF_8_stem},
   {"tur", ENC_UTF_8, turkish_UTF_8_create_env, turkish_UTF_8_close_env, turkish_UTF_8_stem},
   {"turkish", ENC_UTF_8, turkish_UTF_8_create_env, turkish_UTF_8_close_env, turkish_UTF_8_stem},
+  {"yi", ENC_UTF_8, yiddish_UTF_8_create_env, yiddish_UTF_8_close_env, yiddish_UTF_8_stem},
+  {"yid", ENC_UTF_8, yiddish_UTF_8_create_env, yiddish_UTF_8_close_env, yiddish_UTF_8_stem},
+  {"yiddish", ENC_UTF_8, yiddish_UTF_8_create_env, yiddish_UTF_8_close_env, yiddish_UTF_8_stem},
   {0,ENC_UNKNOWN,0,0,0}
 };
 static const char * algorithm_names[] = {
-  "danish", 
-  "dutch", 
-  "english", 
-  "finnish", 
-  "french", 
-  "german", 
-  "hungarian", 
-  "italian", 
-  "norwegian", 
-  "porter", 
-  "portuguese", 
-  "romanian", 
-  "russian", 
-  "spanish", 
-  "swedish", 
-  "turkish", 
+  "arabic",
+  "armenian",
+  "basque",
+  "catalan",
+  "danish",
+  "dutch",
+  "english",
+  "finnish",
+  "french",
+  "german",
+  "greek",
+  "hindi",
+  "hungarian",
+  "indonesian",
+  "irish",
+  "italian",
+  "lithuanian",
+  "nepali",
+  "norwegian",
+  "porter",
+  "portuguese",
+  "romanian",
+  "russian",
+  "serbian",
+  "spanish",
+  "swedish",
+  "tamil",
+  "turkish",
+  "yiddish",
   0
 };

data/ext/isomorfeus_ferret_ext/stem_modules.txt

@@ -9,27 +9,40 @@
 # List all the main algorithms for each language, in UTF-8, and also with
 # the most commonly used encoding.
 
+arabic          UTF_8                   arabic,ar,ara
+armenian        UTF_8                   armenian,hy,hye,arm
+basque          UTF_8,ISO_8859_1        basque,eu,eus,baq
+catalan         UTF_8,ISO_8859_1        catalan,ca,cat
 danish          UTF_8,ISO_8859_1        danish,da,dan
 dutch           UTF_8,ISO_8859_1        dutch,nl,dut,nld
 english         UTF_8,ISO_8859_1        english,en,eng
 finnish         UTF_8,ISO_8859_1        finnish,fi,fin
 french          UTF_8,ISO_8859_1        french,fr,fre,fra
 german          UTF_8,ISO_8859_1        german,de,ger,deu
-hungarian       UTF_8,ISO_8859_1        hungarian,hu,hun
+greek           UTF_8                   greek,el,gre,ell
+hindi           UTF_8                   hindi,hi,hin
+hungarian       UTF_8,ISO_8859_2        hungarian,hu,hun
+indonesian      UTF_8,ISO_8859_1        indonesian,id,ind
+irish           UTF_8,ISO_8859_1        irish,ga,gle
 italian         UTF_8,ISO_8859_1        italian,it,ita
+lithuanian      UTF_8                   lithuanian,lt,lit
+nepali          UTF_8                   nepali,ne,nep
 norwegian       UTF_8,ISO_8859_1        norwegian,no,nor
 portuguese      UTF_8,ISO_8859_1        portuguese,pt,por
 romanian        UTF_8,ISO_8859_2        romanian,ro,rum,ron
 russian         UTF_8,KOI8_R            russian,ru,rus
+serbian         UTF_8                   serbian,sr,srp
 spanish         UTF_8,ISO_8859_1        spanish,es,esl,spa
 swedish         UTF_8,ISO_8859_1        swedish,sv,swe
+tamil           UTF_8                   tamil,ta,tam
 turkish         UTF_8                   turkish,tr,tur
+yiddish         UTF_8                   yiddish,yi,yid
 
 # Also include the traditional porter algorithm for english.
 # The porter algorithm is included in the libstemmer distribution to assist
 # with backwards compatibility, but for new systems the english algorithm
 # should be used in preference.
-porter          UTF_8,ISO_8859_1        porter
+porter          UTF_8,ISO_8859_1        porter			english
 
 # Some other stemmers in the snowball project are not included in the standard
 # distribution. To compile a libstemmer with them in, add them to this list,
@@ -39,12 +52,12 @@ porter          UTF_8,ISO_8859_1        porter
 # algorithms are:
 #
 # german2          - This is a slight modification of the german stemmer.
-#german2          UTF_8,ISO_8859_1        german2
+#german2          UTF_8,ISO_8859_1        german2		german
 #
 # kraaij_pohlmann  - This is a different dutch stemmer.
-#kraaij_pohlmann  UTF_8,ISO_8859_1        kraaij_pohlmann
+#kraaij_pohlmann  UTF_8,ISO_8859_1        kraaij_pohlmann	dutch
 #
 # lovins           - This is an english stemmer, but fairly outdated, and
 #                    only really applicable to a restricted type of input text
 #                    (keywords in academic publications).
-#lovins           UTF_8,ISO_8859_1        lovins
+#lovins           UTF_8,ISO_8859_1        lovins		english