Unify the method of extracting tags (#3138)

This commit is contained in:
abcang 2017-05-20 03:19:14 +09:00 committed by Eugen Rochko
parent d2e0edd721
commit d22cec81fb
2 changed files with 27 additions and 1 deletions

View file

@ -30,4 +30,30 @@ module Extractor
end
possible_entries
end
def extract_hashtags_with_indices(text, _options = {})
return [] unless text =~ /#/
tags = []
text.scan(Tag::HASHTAG_RE) do |hash_text, _|
match_data = $LAST_MATCH_INFO
start_position = match_data.char_begin(1) - 1
end_position = match_data.char_end(1)
after = $'
if after =~ %r{\A://}
hash_text.match(/(.+)(https?\Z)/) do |matched|
hash_text = matched[1]
end_position -= matched[2].char_length
end
end
tags << {
hashtag: hash_text,
indices: [start_position, end_position],
}
end
tags.each { |tag| yield tag[:hashtag], tag[:indices].first, tag[:indices].last } if block_given?
tags
end
end

View file

@ -2,7 +2,7 @@
class ProcessHashtagsService < BaseService
def call(status, tags = [])
tags = status.text.scan(Tag::HASHTAG_RE).map(&:first) if status.local?
tags = Extractor.extract_hashtags(status.text) if status.local?
tags.map { |str| str.mb_chars.downcase }.uniq(&:to_s).each do |tag|
status.tags << Tag.where(name: tag).first_or_initialize(name: tag)