Relevanssi supports Chinese Character?

Relevanssi, by default, understands Chinese, Japanese and Korean characters like it understands all other UTF-8 characters. That’s not the problem.

The problem is that words are not separated by spaces in Chinese, Japanese and Korean, so when there’s a piece of text in Chinese, Relevanssi sees one long word instead of multiple words.

Since Relevanssi search weights are very much based on words, you essentially get partial, random results. Not all words are found if they are in the middle of a sentence, and the weights for whatever is found don’t make much sense.

Hi. I found there are few differences when using Relevanssi & WordPress default search engine for Chinese characters.
Like searching for the same keyword.
with Relevanssi, 17 items for the result
with WordPress, 18 items for the result

I want to try this phpjieba.
https://www.relevanssi.com/knowledge-base/relevanssi-and-languages/

Step 1: I installed the package on my server
Wondering the Step 2: php.ini

php.ini
extension=jieba.so
jieba.enable=1
jieba.dict_path=/data/softs/phpjieba/cjieba/dict

*jieba.dict_path Point to the corresponding directory where the dictionary is located, please replace and change according to your own compilation directory`

I am not sure what the path should be in this case.
Would you have some clues?

Thanks again!

I have no idea. All the jieba documentation is in Chinese, which I can’t read. I’d assume the make install step will create the dictionary file somewhere, and the path should point there.

My case was installed on the root. I think.
/phpjieba/cjieba/dict
After a couple of tries no successes… I gave up.
Relevanssi is already good enough at reading Chinese characters though
Thank you anyway.

If you’re not using the jieba, it’s best to ignore all relevancy ratings. The scores Relevanssi assigns to posts are all wrong without the jieba and make very little sense. I’m not sure whether you actually get many benefits from using Relevanssi compared to the default WP search.