Technical question about Flarum's searching function of non-latin languages

自在の偶

I am curious about it. These more than 10 years, I used many PHP based forums, all of which support global searching in Chinese, by default.

But in Flarum, Chinese searching is not supported by default. Additional searching drivers must be installed to achieve Chinese searching. I didn't test, but I guess Japanese, Korean and even most other non-latin languages, neither.

Searching is a basic need for most all users. But installing and configuring those drivers is not easy.

I know there must be a reason for Flarum not supporting Chinese searching by default. The detail is?

UPDATE: Thanks to GreXXL I tested on this site, input more than 3 (thus, at least 4) Chinese characters, the searching function will work, too. The remaining point is, most time, searching in Chinese will start in only 2 characters, because 2 Chinese characters could mean really a lot of information, like 90% city names and maybe 40% of people names, and so on.

Hope this could be heard and optimized in the future.

GreXXL

自在の偶 the only issue is that three letters are required. Other then that search should just work fine.

自在の偶

GreXXL I tried on this site. I input 4 Chinese characters - 青出于蓝, the searching result is nothing. Actually, there should be at least 1 result.

Karl

自在の偶 I offer Typesense for my hosting clients... Granted I don't have any foreign language communities in my stack, but maybe you'll yield better results integrating your Flarum with it since Typesense does a great job indexing based on characters and probability.

Edit, I do see you mentioned other search drivers.
I'm half Korean and curious about this. I hadn't known this was actually an issue so I'll try and dive into it either today or tomorrow and will publish my findings via GitHub Issue.

GreXXL

自在の偶 works just fine for me?

自在の偶

GreXXL my apologies 😅 I tested again, it really has the expected result. I am lost ...

自在の偶

Karl thanks for sharing. grexxl told the reason, at least 4 letters (in Chinese and other languages written in characters, 4 characters), then the default searching function will work.

I tested, and that's right.

The maintaining problem is 4 Chinese characters mean really a lot of information. Most time, we need to searching with 2 characters(at least). In fact, 2 characters mean a lot too. 90% of city names are in 2 characters.

Hope to see optimization 😅

GreXXL

自在の偶 this is not really Flarum related but a MySQL setting. You can change the setting on your own server there are some threads about this here. Alternatively you could use an extension and other search drivers like elastic search or Agolia.