Add spellchecker runner and initial generated dictionary (ru)

[razum2um]

I suggest using yaspeller to check md content.

• russian maintainers supposed to use it locally => Rakefile changed
• and need to sync exceptions list (dictionary, array or regexps) => new json file added

Usage:

npm i -g yaspeller
rake 'check:spelling[ru]'

Default output looks like this:

www.ruby-lang.org/ru/news/_posts/2013-02-24-ruby-2-0-0-p0-is-released.md 170 ms
-----
Typos: 2
1. эксперемент (73:32, suggest: эксперимент)
2. отправлии (171:1, suggest: отправили)

www.ruby-lang.org/ru/news/_posts/2013-12-21-ruby-version-policy-changes-with-2-1-0.md 191 ms
-----
Typos: 2
1. yдаление (42:3, en: y*******, ru: *даление, suggest: удаление)
2. oбратно (43:3, en: o******, ru: *братно, suggest: обратно, обратной)
-----

All those errors fixed in related PR: #2719

Note that it suggests replacements and finds locale-mismatched symbols as well.

Update dictionary from scratch

Use this repo

rm lib/spelling/ru/dictionary.json
rake 'check:spelling[ru,json]' # generates ./yaspeller_report.json which we git-ignore
cd ../yaspeller-dictionary-builder
python src/dictionary.py ../www.ruby-lang.org/yaspeller_report.json > ../www.ruby-lang.org/lib/spelling/ru/dictionary.json

Limitations:

• Unfortunately, it supports only ru, ukrainian, and en. Consider adding it for en?
• As the cli utility is based on free api and has rate limits, I don't put it into CI / standard linter flow. Maybe a git hook makes sense if a commit contains changes under ru

[lex111]

Thanks for the suggestion, but I would prefer to add this check to CI (GitHub actions) as a separate workflow.

As the cli utility is based on free api and has rate limits, I don't put it into CI

Actually it is not big deal, I use yaspeller on many projects and never experienced with its rate limits, moreover it is possible to make yaspeller run only when files related with Russian translation has been changed.

[Nakilon]

Personally I used hunspell like this:

hunspell -d dict-en-20210701/en_US,dict_ru_ru-aot-0.4.5/russian-aot -p my.txt -l ru/news/_posts/202* | sort | uniq

Their README told to download the dictionaries in .oxt format, rename to .zip and unpack.
And the my.txt in the example is a text file with skipped words, like:

Alexandr
CVE
Gemfile
RDoc
Savca
aycabta
bundler
lang
...