diff options
author | haturatu <taro@eyes4you.org> | 2024-10-09 01:36:19 +0900 |
---|---|---|
committer | haturatu <taro@eyes4you.org> | 2024-10-09 01:36:19 +0900 |
commit | 031b8014901bd101d92af148a841e9b9c8fb59f5 (patch) | |
tree | e284928681323b7bebf08f34bb15db8a11a20f5c /geturl.rb | |
parent | 736e6f490c6a9555141188c78c6aa48782d3b5b6 (diff) |
fi
Diffstat (limited to 'geturl.rb')
-rwxr-xr-x | geturl.rb | 4 |
1 files changed, 2 insertions, 2 deletions
@@ -7,7 +7,7 @@ require 'nokogiri' require 'charlock_holmes' require 'unicode_utils' -$FILE_PATH = './ok' +$FILE_PATH = '/Your/URLs/list/file' $RESULT_FILE = 'Result' $OTHER_ERROR_FILE = 'Other' $CONCURRENCY = 10 @@ -17,7 +17,7 @@ def is_garbled?(text) end def clean_title(title) - title = title.chars.reject { |ch| UnicodeUtils.general_category(ch).start_with?('C') && !['(', ')', '[', ']', '{', '}', '【', '】', '「', '」', '(' ,')' ].include?(ch) }.join + title = title.chars.reject { |ch| UnicodeUtils.general_category(ch).start_with?('C') && !['(', ')', '[', ']', '{', '}', '【', '】', '【', '】', '「', '」', '(' ,')' ].include?(ch) }.join title = UnicodeUtils.nfkc(title) title = title.chars.select(&:valid_encoding?).join title.strip |