Fix tag normalization and migration not removing duplicate tags (#11441)

Fix #11428
main
Eugen Rochko 5 years ago committed by GitHub
parent aefeb65656
commit e136112ab7
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23

@ -65,7 +65,7 @@ class Tag < ApplicationRecord
class << self
def find_or_create_by_names(name_or_names)
Array(name_or_names).map(&method(:normalize)).uniq.map do |normalized_name|
Array(name_or_names).map(&method(:normalize)).uniq { |str| str.mb_chars.downcase.to_s }.map do |normalized_name|
tag = matching_name(normalized_name).first || create(name: normalized_name)
yield tag if block_given?
@ -77,7 +77,7 @@ class Tag < ApplicationRecord
def search_for(term, limit = 5, offset = 0)
pattern = sanitize_sql_like(normalize(term.strip)) + '%'
Tag.where(arel_table[:name].lower.matches(pattern.downcase))
Tag.where(arel_table[:name].lower.matches(pattern.mb_chars.downcase.to_s))
.order(:name)
.limit(limit)
.offset(offset)
@ -92,7 +92,7 @@ class Tag < ApplicationRecord
end
def matching_name(name_or_names)
names = Array(name_or_names).map { |name| normalize(name).downcase }
names = Array(name_or_names).map { |name| normalize(name).mb_chars.downcase.to_s }
if names.size == 1
where(arel_table[:name].lower.eq(names.first))
@ -104,7 +104,7 @@ class Tag < ApplicationRecord
private
def normalize(str)
str.gsub(/\A#/, '').mb_chars.to_s
str.gsub(/\A#/, '')
end
end

@ -2,6 +2,19 @@ class AddCaseInsensitiveIndexToTags < ActiveRecord::Migration[5.2]
disable_ddl_transaction!
def up
Tag.connection.select_all('SELECT string_agg(id::text, \',\') AS ids FROM tags GROUP BY lower(name) HAVING count(*) > 1').to_hash.each do |row|
canonical_tag_id = row['ids'].split(',').first
redundant_tag_ids = row['ids'].split(',')[1..-1]
safety_assured do
execute "UPDATE accounts_tags AS t0 SET tag_id = #{canonical_tag_id} WHERE tag_id IN (#{redundant_tag_ids.join(', ')}) AND NOT EXISTS (SELECT t1.tag_id FROM accounts_tags AS t1 WHERE t1.tag_id = #{canonical_tag_id} AND t1.account_id = t0.account_id)"
execute "UPDATE statuses_tags AS t0 SET tag_id = #{canonical_tag_id} WHERE tag_id IN (#{redundant_tag_ids.join(', ')}) AND NOT EXISTS (SELECT t1.tag_id FROM statuses_tags AS t1 WHERE t1.tag_id = #{canonical_tag_id} AND t1.status_id = t0.status_id)"
execute "UPDATE featured_tags AS t0 SET tag_id = #{canonical_tag_id} WHERE tag_id IN (#{redundant_tag_ids.join(', ')}) AND NOT EXISTS (SELECT t1.tag_id FROM featured_tags AS t1 WHERE t1.tag_id = #{canonical_tag_id} AND t1.account_id = t0.account_id)"
end
Tag.where(id: redundant_tag_ids).in_batches.delete_all
end
safety_assured { execute 'CREATE UNIQUE INDEX CONCURRENTLY index_tags_on_name_lower ON tags (lower(name))' }
remove_index :tags, name: 'index_tags_on_name'
remove_index :tags, name: 'hashtag_search_index'

@ -82,6 +82,40 @@ RSpec.describe Tag, type: :model do
end
end
describe '.find_normalized' do
it 'returns tag for a multibyte case-insensitive name' do
upcase_string = 'abcABCやゆよ'
downcase_string = 'abcabcやゆよ';
tag = Fabricate(:tag, name: downcase_string)
expect(Tag.find_normalized(upcase_string)).to eq tag
end
end
describe '.matching_name' do
it 'returns tags for multibyte case-insensitive names' do
upcase_string = 'abcABCやゆよ'
downcase_string = 'abcabcやゆよ';
tag = Fabricate(:tag, name: downcase_string)
expect(Tag.matching_name(upcase_string)).to eq [tag]
end
end
describe '.find_or_create_by_names' do
it 'runs a passed block once per tag regardless of duplicates' do
upcase_string = 'abcABCやゆよ'
downcase_string = 'abcabcやゆよ';
count = 0
Tag.find_or_create_by_names([upcase_string, downcase_string]) do |tag|
count += 1
end
expect(count).to eq 1
end
end
describe '.search_for' do
it 'finds tag records with matching names' do
tag = Fabricate(:tag, name: "match")

Loading…
Cancel
Save