Search results for polish are in spanish

[krzysztofzuraw]

Issue description

When user searches for term in polish language search results are presented in spanish:

Yet clicking on term results in polish entry 🤷‍♂

Language

Polish

Operating system

Tested on mac

[das-g]

I can reproduce this in Firefox on Linux and my locale is neither Spanish nor Polish:

LANG=de_CH.UTF-8
LC_CTYPE="de_CH.UTF-8"
LC_NUMERIC="de_CH.UTF-8"
LC_TIME="de_CH.UTF-8"
LC_COLLATE="de_CH.UTF-8"
LC_MONETARY="de_CH.UTF-8"
LC_MESSAGES="de_CH.UTF-8"
LC_PAPER="de_CH.UTF-8"
LC_NAME="de_CH.UTF-8"
LC_ADDRESS="de_CH.UTF-8"
LC_TELEPHONE="de_CH.UTF-8"
LC_MEASUREMENT="de_CH.UTF-8"
LC_IDENTIFICATION="de_CH.UTF-8"
LC_ALL=

To reproduce:

1. go to https://tutorial.djangogirls.org
2. choose "Polski"
3. enter something in the top left search field (where it says "Wpisz szukaną frazę")

expected:

You get results from the Polish translation of the Django Girls tutorial

observed:

• When you enter a search term that occurs in the Polish translation of the Django Girls tutorial, but not in the Spanish one, you don't get any results. E.g. "edytor"
• When you enter a search term that occurs in the Spanish translation, you get results from there. E.g. "editor"
  • The links for the search results go to the respective Polish chapters, though.

[das-g]

Is this a bug in GitBook legacy? I can't see any related settings on our side (i.e. in the content of this repo) that might be misconfigured.

[krzysztofzuraw]

It can be - I found that gitbook legacy is using https://github.com/GitbookIO/plugin-search which seems not maintained anymore

[das-g]

I've systematically tried in all languages listed on https://tutorial.djangogirls.org/

• means search results appear in that same language
• means search results appear in Spanish instead

There seems to be a clear pattern linked to the order in which the translations are listed on https://tutorial.djangogirls.org/ (which probably also is the order in which they were initially published?): Everything after Spanish has Spanish search results, while those listed before Spanish have results in the respective proper language:

• 🇺🇸 English
• 🇨🇿 Čeština (beta)
• 🇩🇪 Deutsch
• 🇬🇷 Ελληνικά (beta)
• 🇪🇸 Español (beta)
• 🇮🇷 فارسی (beta)
• 🇫🇷 Français (beta)
• 🇭🇺 Magyar (beta)
• 🇮🇹 Italiano (beta)
• 🇯🇵 日本語
• 🇰🇵/🇰🇷 한국어 (beta)
• 🇵🇱 Polski
• 🇧🇷 Português-brasileiro
• 🇷🇺 Русский (beta)
• 🇸🇰 Slovenčina (beta)
• 🇹🇷 Türkçe (beta)
• 🇺🇦 Українська
• 🇨🇳 简体中文

[das-g]

I wonder whether this is related to #177.

@SamyPesse, can you help us here?

[magul]

I have just bumped into probably root cause here. When I build tutorial locally I got

magul@bohr tutorial (crowdin-translation-de +) $ gitbook build
info: parsing multilingual book, with 18 languages 
info: 17 plugins are installed 
info: 15 explicitly listed 
info: loading plugin "heading-anchors"... OK 
info: loading plugin "ga"... OK 
info: loading plugin "richquotes"... OK 
info: loading plugin "github"... OK 
info: loading plugin "language-picker"... OK 
info: loading plugin "sidebar-ad"... OK 
info: loading plugin "codeblock-label"... OK 
info: loading plugin "sectionx-ex"... OK 
info: loading plugin "collapsible-menu"... OK 
info: loading plugin "highlight"... OK 
info: loading plugin "search"... OK 
info: loading plugin "lunr"... OK 
info: loading plugin "sharing"... OK 
info: loading plugin "fontsettings"... OK 
info: loading plugin "theme-default"... OK 
info: found 18 asset files 
warn: "options" property is deprecated, use config.get(key) instead 
info:  
info: generating language "en" 
info: found 26 pages 
info: found 57 asset files 
warn: "sections" property is deprecated, use page.content instead 
warn: "this.generator" property is deprecated, use "this.output.name" instead 
warn: "navigation" property is deprecated 
warn: "book" property is deprecated, use "this" directly instead 
info:  
info: generating language "cs" 
info: found 25 pages 
info: found 50 asset files 
info:  
info: generating language "de" 
info: found 26 pages 
info: found 56 asset files 
info:  
info: generating language "el" 
info: found 26 pages 
info: found 54 asset files 
info:  
info: generating language "es" 
info: found 26 pages 
info: found 55 asset files 
info:  
info: generating language "fa" 
info: found 26 pages 
info: found 56 asset files 
warn: search index is too big, indexing is now disabled 
info:  
info: generating language "fr" 
info: found 26 pages 
info: found 57 asset files 
info:  
info: generating language "hu" 
info: found 25 pages 
info: found 50 asset files 
info:  
info: generating language "it" 
info: found 25 pages 
info: found 50 asset files 
info:  
info: generating language "ja" 
info: found 26 pages 
info: found 55 asset files 
info:  
info: generating language "ko" 
info: found 26 pages 
info: found 75 asset files 
info:  
info: generating language "pl" 
info: found 26 pages 
info: found 54 asset files 
info:  
info: generating language "pt" 
info: found 26 pages 
info: found 61 asset files 
info:  
info: generating language "ru" 
info: found 26 pages 
info: found 74 asset files 
info:  
info: generating language "sk" 
info: found 26 pages 
info: found 53 asset files 
info:  
info: generating language "tr" 
info: found 26 pages 
info: found 53 asset files 
info:  
info: generating language "uk" 
info: found 25 pages 
info: found 49 asset files 
info:  
info: generating language "zh" 
info: found 25 pages 
info: found 49 asset files 
info: >> generation finished with success in 92.5s ! 

There's a warning during rendering Persian language:

warn: search index is too big, indexing is now disabled 

I would assume, that increasing index threshold should resolve that issue. (We shold alos check hot to do so on Gitbook service).

[IllicLanthresh]

I'm having the same issue but I see that the problem is not tied to Polish-Spanish, searching yields results in all languages, independently of the selected language. Probably better to modify the issue tittle