#ifndef FTS_FILTER_H
#define FTS_FILTER_H
struct fts_language;
struct fts_filter;
/*
Settings are given in the form of a const char * const *settings =
{"key, "value", "key2", "value2", NULL} array of string pairs.
The array has to be NULL terminated.
*/
/*
Settings: "stopwords_dir", path to the directory containing stopword files.
Stopword files are looked up in "<path>"/stopwords_<lang>.txt
*/
extern const struct fts_filter *fts_filter_stopwords;
/*
Settings: "lang", language of the stemmed language.
*/
extern const struct fts_filter *fts_filter_stemmer_snowball;
/*
Settings: "id", description of the normalizing/translitterating rules
to use. See
http://userguide.icu-project.org/transforms/general#TOC-Transliterator-Identifiers
for syntax. Defaults to "Any-Lower; NFKD; [: Nonspacing Mark :] Remove; NFC"
"maxlen", maximum length of tokens that ICU normalizer will output.
Defaults to 250.
*/
extern const struct fts_filter *fts_filter_normalizer_icu;
/* Lowercases the input. Supports UTF8, if libicu is available. */
extern const struct fts_filter *fts_filter_lowercase;
/* Removes <'s> suffix from words. */
extern const struct fts_filter *fts_filter_english_possessive;
/* Removes prefixing contractions from words. */
extern const struct fts_filter *fts_filter_contractions;
/* Register all built-in filters. */
void fts_filters_init(void);
void fts_filters_deinit(void);
/* Register a new class explicitly. Built-in classes are automatically
registered. */
/*
Filtering workflow, find --> create --> filter --> destroy.
*/
struct fts_filter *parent,
const struct fts_language *lang,
const char *const *settings,
struct fts_filter **filter_r,
const char **error_r);
/* Returns 1 if token is returned in *token, 0 if token was filtered
out (*token is also set to NULL) and -1 on error.
Input is also given via *token.
*/
const char **error_r);
#endif