Browse Source

Change language detection to include hashtags as words (#11341)

tags/v2.9.3
Eugen Rochko 2 months ago
parent
commit
212848b66e
2 changed files with 4 additions and 4 deletions
  1. 1
    1
      app/lib/language_detector.rb
  2. 3
    3
      spec/lib/language_detector_spec.rb

+ 1
- 1
app/lib/language_detector.rb View File

@@ -69,7 +69,7 @@ class LanguageDetector
new_text = remove_html(text)
new_text.gsub!(FetchLinkCardService::URL_PATTERN, '')
new_text.gsub!(Account::MENTION_RE, '')
new_text.gsub!(Tag::HASHTAG_RE, '')
new_text.gsub!(Tag::HASHTAG_RE) { |string| string.gsub(/[#_]/, '#' => '', '_' => ' ').gsub(/[a-z][A-Z]|[a-zA-Z][\d]/) { |s| s.insert(1, ' ') }.downcase }
new_text.gsub!(/:#{CustomEmoji::SHORTCODE_RE_FRAGMENT}:/, '')
new_text.gsub!(/\s+/, ' ')
new_text

+ 3
- 3
spec/lib/language_detector_spec.rb View File

@@ -32,11 +32,11 @@ describe LanguageDetector do
expect(result).to eq 'Our website is and also'
end

it 'strips #hashtags from strings before detection' do
string = 'Hey look at all the #animals and #fish'
it 'converts #hashtags back to normal text before detection' do
string = 'Hey look at all the #animals and #FishAndChips'

result = described_class.instance.send(:prepare_text, string)
expect(result).to eq 'Hey look at all the and'
expect(result).to eq 'Hey look at all the animals and fish and chips'
end
end


Loading…
Cancel
Save