Indexing in different languages

[artycorp]

Good afternoon!

I wanted to index the document in Russian:

• connected the package https://github.com/MihaiValentin/lunr-languages

  #package.json
  ...
  "dependencies": {
  ...
  "lunr-languages": "^1.4.0"
  },
  ...

  -In the source code, index.js added the following:

  var lunr = require("lunr");
  require('lunr-languages/lunr.stemmer.support')(lunr);
  require('lunr-languages/lunr.ru')(lunr);
  require('lunr-languages/lunr.multi')(lunr);
  ...
    const index = lunr(function () {
      this.use(lunr.multiLanguage('en', 'ru'));
      this.ref("id");
      this.field("title");
      this.field("content");
  ...

• yarn build
• copy build directory to nginx
• open in browser
• enter some text in search bar

result:

Uncaught (in promise) Error: Cannot load unregistered function: stemmer-ru
    at 2.6c5301c6.js:2
    at Array.forEach (<anonymous>)
    at Function.P.Pipeline.load (2.6c5301c6.js:2)
    at Function.P.Index.load (2.6c5301c6.js:2)
    at 2.6c5301c6.js:2
    at async Promise.all (/index 0)

i am not strong in js can you help me?

[yoichiwo7]

I suspect you only modified src/index.js. This modification will let you generate multi-language index file but you can't load the index file unless you also modify the SearchBar side.

You'll also need to add lunr-languages statements into src/theme/SearchBar/index.js so that SearchBar can load the multi-language index file. Just add the statements after the following line. (adding require statements works but I'm not sure mixing import/require statements is appropriate)

https://github.com/cmfcmf/docusaurus-search-local/blob/4dfd5c6870f629554ad3949ea71347a290a0bb4c/src/theme/SearchBar/index.js#L8

[cmfcmf]

Interesting, I hadn't thought about stemming non-English languages. I am wondering whether there really is the usecase of using multiple languages on the same page/document or if we can assume that every page is written in a single language? @artycorp are your pages a mix of Russian and English?

[cmfcmf]

I added support for multiple languages in 187b6d2 - took some time to wrap my head around how lunr-languages works, but it should now work fine for single and multiple languages :tada: