add threshold for check_chinese_space_spam to avoid false positive

Author: ramsayleung

app/services/rule_based_classifier.rb

@@ -17,6 +17,7 @@ def classify
   private
 
   def check_chinese_spacing_spam
+    min_chinese_chars = 5
     # This pattern specifically looks for a Chinese character, followed by a space,
     # and then another Chinese character like this
     # 跟 单 像 捡 钱 ！ 再 不 进 群 是 傻 狗 ！
@@ -31,7 +32,7 @@ def check_chinese_spacing_spam
       threshold = Rails.application.config.chinese_space_spam_threshold
       ratio = chinese_chars > 0 ? spaced_chinese_words_count.to_f / chinese_chars : 0.0
 
-      if ratio > threshold
+      if ratio > threshold && chinese_chars >= min_chinese_chars
         Rails.logger.info "Classified as spam due to high Chinese character spacing ratio: #{ratio}"
         return Shared::ClassificationResult.new(is_spam: true, target: "message_content")
       end

vendor/dictionaries/user.dict.utf8

@@ -299,4 +299,13 @@ K 线
 冲鸭
 不良林
 约啪
-舔狗
+舔狗
+氪金号
+淘汰榜
+港澳通行证
+黑花
+川普
+学生党
+分享群
+冷静期
+大魔王