Search in Chinese locale | Voters

Search in Chinese locale

added in v3

Charlton

In Chinese sentences, there are no spaces among phrases, or all hieraglyphs are lined up with no spaces in between.
So search in Chinese should do partial match, or match by each hieraglyph.

March 4, 2020

Nicolas Giard

marked this post as

added in v3

Lesca

Please add PostgreSQL dictionary - Chinese to the list. 
Currently the search only takes the whole sentence or phases separated by punctuation marks, but not word by word. 
As a global project, hopefully this can be improved.

Nicolas Giard

Lesca: Dictionaries are provided by PostgreSQL, not Wiki.js.

Lesca

Nicolas Giard: Hello Nicolas, you are right. I added a Chinese word Seperator extension "jieba" to postgres image: 
https://github.com/lesca/postgres-jieba
And set the default_text_search to jieba:
shared_preload_libraries = 'pg_jieba.so'  # (change requires restart)
# default_text_search_config='pg_catalog.simple'; default value
default_text_search_config='jiebacfg'; uncomment to make 'jiebacfg' as default
But the search doesn't work, as the option in "Wiki2" is "simple" and may override the "jieba" option in config. I tried basic as well, no luck. 
Any ideas to let "wiki2" use the custom extension?
Regards,
Lesca Fang

Nicolas Giard

Lesca: You can set your custom dictionnary directly in the DB by editing the "searchEngines" table and change the config column for the row with key "postgres". Then restart your wiki to apply the changes.

Lesca

Nicolas Giard: Thanks Nicolas. I changed the

postgres

{"dictLanguage":"jiebacfg"}

. Now everything works like a charm.

Hopefully it can be configured in web UI in future.