adirh3 / Fluent-Search

Official repository for Fluent Search, use to report issues or ask for a new feature
1.42k stars 32 forks source link

Content search cannot find Chinese characters #690

Open ligq opened 1 year ago

ligq commented 1 year ago

Content search cannot find Chinese characters My windows is Chinese version win11 22H2 22621.755

u-Sir commented 1 year ago

可以试下多输入几个字😉

u-Sir commented 1 year ago

试了下,中文要至少3个字才能搜到

ligq commented 1 year ago

试了下,中文要至少3个字才能搜到

You are right. And I found for English we also have to input at least 3 letters to start the search. But for the Chinese, it's a bug. A Chinese character is equivalent to an English word, not a letter. The search should start after we input one Chinese character.

u-Sir commented 1 year ago

试了下,中文要至少3个字才能搜到

You are right. And I found for English we also have to input at least 3 letters to start the search. But for the Chinese, it's a bug. A Chinese character is equivalent to an English word, not a letter. The search should start after we input one Chinese character.

developer said "it's not only Chinese it's all languages"

adirh3 commented 1 year ago

I might add option to use Pinyin for content search, but it would double the index size. Would you be OK with that? (Double since it will account for both English a pinyin and Chinese)

ligq commented 1 year ago

I might add option to use Pinyin for content search, but it would double the index size. Would you be OK with that? (Double since it will account for both English a pinyin and Chinese)

Thank you for your quick response. Pinyin could be helpful, but it's not a perfect solution, since one pinyin generally corresponds to several Chinese characters.

adirh3 commented 1 year ago

I might add option to use Pinyin for content search, but it would double the index size. Would you be OK with that? (Double since it will account for both English a pinyin and Chinese)

Thank you for your quick response. Pinyin could be helpful, but it's not a perfect solution, since one pinyin generally corresponds to several Chinese characters.

Yes I am aware of that thanks to @u-Sir :) The thing is, searching with less than 3 letters would be extremely slow, probably like more than 6 seconds, and it will also be CPU intensive. I might enable that for Chinese specifically in the next nightly update and see how it goes.