Question: cannot search anything by i18n zh.yaml

[hitzhangjie]

Could you help sovling the searching problem? I write a ebook here: https://hitzhangjie.pro/go-internals/, it uses hugo-book theme. I don't know why the searching not working. I use Chrome to debug the problem, and I see the document is indexed.

Maybe the problem is relevant with i18n zh booksearchConfig, I change the tokenize function: from str.replace(/[\x00-\x7F]/g, '').split(''); to str.split(/\W+/).concat(str.replace(/[\x00-\x7F]/g, '').split('')).filter(e => !!e);

And it worked.

[alex-shpak]

Hi! Nice to see theme used :) that's a lot of content. It is possible that search config tokenization needs update, as I don't have any idea how to search in chineese properly, and relied on google-help.

Although, when I naively put chineese 'Lorem ipsum' to page and search fractions of it, it works for me in zh locale. Can you send what content you have and what are you trying to search as example?

[hitzhangjie]

For example, I want to search both Chinese and English words, like '码农' or 'AST' which appears in markdown file.

Let me show an example here to reproduce the problem.

Case: tokenize function uses 'str.replace(/[\x00-\x7F]/g, '').split('');'

Let's search '码农', then we see the search result seems OK, it returns a document:

Then I search 'AST', then we see the search result is empty, it should returns 2 documents:

Case: tokenize function uses 'str.split(/\W+/).concat(str.replace(/[\x00-\x7F]/g, '').split('')).filter(e => !!e);'

Let's search both words again, it works.

search '码农':

search 'ast':

---

ps: Actually I don't know the internals about tokenize function, I came across this problem before and wrote down the right tokenizer settings from Google. Hope this could help 'hugo-book' theme.

Here is the ebook address, you may want to test it here: https://hitzhangjie.pro/go-internals/