database/postgres
1
0
Fork 0
mirror of https://github.com/postgres/postgres.git synced 2025-11-10 17:42:29 +03:00
Code Activity

Update snowball

Browse Source 
Update to snowball tag v2.0.0.  Major changes are new stemmers for
Basque, Catalan, and Hindi.

Discussion: https://www.postgresql.org/message-id/flat/a8eeabd6-2be1-43fe-401e-a97594c38478%402ndquadrant.com
This commit is contained in:
Peter Eisentraut
2020-06-08 07:58:51 +02:00
parent 57cb806308
commit cbcc8726bb
97 changed files with 6965 additions and 2217 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
src/backend/snowball/dict_snowball.c
View File

@@ -26,6 +26,8 @@
/* Now we can include the original Snowball header.h */
#include "snowball/libstemmer/header.h"
#include "snowball/libstemmer/stem_ISO_8859_1_basque.h"
#include "snowball/libstemmer/stem_ISO_8859_1_catalan.h"
#include "snowball/libstemmer/stem_ISO_8859_1_danish.h"
#include "snowball/libstemmer/stem_ISO_8859_1_dutch.h"
#include "snowball/libstemmer/stem_ISO_8859_1_english.h"
@@ -44,6 +46,8 @@
#include "snowball/libstemmer/stem_ISO_8859_2_romanian.h"
#include "snowball/libstemmer/stem_KOI8_R_russian.h"
#include "snowball/libstemmer/stem_UTF_8_arabic.h"
#include "snowball/libstemmer/stem_UTF_8_basque.h"
#include "snowball/libstemmer/stem_UTF_8_catalan.h"
#include "snowball/libstemmer/stem_UTF_8_danish.h"
#include "snowball/libstemmer/stem_UTF_8_dutch.h"
#include "snowball/libstemmer/stem_UTF_8_english.h"
@@ -51,6 +55,7 @@
#include "snowball/libstemmer/stem_UTF_8_french.h"
#include "snowball/libstemmer/stem_UTF_8_german.h"
#include "snowball/libstemmer/stem_UTF_8_greek.h"
#include "snowball/libstemmer/stem_UTF_8_hindi.h"
#include "snowball/libstemmer/stem_UTF_8_hungarian.h"
#include "snowball/libstemmer/stem_UTF_8_indonesian.h"
#include "snowball/libstemmer/stem_UTF_8_irish.h"
@@ -92,6 +97,8 @@ static const stemmer_module stemmer_modules[] =
/*
* Stemmers list from Snowball distribution
*/
STEMMER_MODULE(basque, PG_LATIN1, ISO_8859_1),
STEMMER_MODULE(catalan, PG_LATIN1, ISO_8859_1),
STEMMER_MODULE(danish, PG_LATIN1, ISO_8859_1),
STEMMER_MODULE(dutch, PG_LATIN1, ISO_8859_1),
STEMMER_MODULE(english, PG_LATIN1, ISO_8859_1),
@@ -110,6 +117,8 @@ static const stemmer_module stemmer_modules[] =
STEMMER_MODULE(romanian, PG_LATIN2, ISO_8859_2),
STEMMER_MODULE(russian, PG_KOI8R, KOI8_R),
STEMMER_MODULE(arabic, PG_UTF8, UTF_8),
STEMMER_MODULE(basque, PG_UTF8, UTF_8),
STEMMER_MODULE(catalan, PG_UTF8, UTF_8),
STEMMER_MODULE(danish, PG_UTF8, UTF_8),
STEMMER_MODULE(dutch, PG_UTF8, UTF_8),
STEMMER_MODULE(english, PG_UTF8, UTF_8),
@@ -117,6 +126,7 @@ static const stemmer_module stemmer_modules[] =
STEMMER_MODULE(french, PG_UTF8, UTF_8),
STEMMER_MODULE(german, PG_UTF8, UTF_8),
STEMMER_MODULE(greek, PG_UTF8, UTF_8),
STEMMER_MODULE(hindi, PG_UTF8, UTF_8),
STEMMER_MODULE(hungarian, PG_UTF8, UTF_8),
STEMMER_MODULE(indonesian, PG_UTF8, UTF_8),
STEMMER_MODULE(irish, PG_UTF8, UTF_8),