From 61e2e62326423da968e61c45a4ba23fdc1827d04 Mon Sep 17 00:00:00 2001
From: Mile Shi <shimile18@gmail.com>
Date: Fri, 2 May 2025 16:49:56 +0800
Subject: [PATCH] [fix] issue #105 tagging not works when enable ignore

---
 .../result/WordResultHandlerWordTags.java     | 18 ++++++--
 .../sensitive/word/bugs/b105/Bug105Test.java  | 45 +++++++++++++++++++
 2 files changed, 60 insertions(+), 3 deletions(-)
 create mode 100644 src/test/java/com/github/houbb/sensitive/word/bugs/b105/Bug105Test.java

diff --git a/src/main/java/com/github/houbb/sensitive/word/support/result/WordResultHandlerWordTags.java b/src/main/java/com/github/houbb/sensitive/word/support/result/WordResultHandlerWordTags.java
index cab8a62..3893f55 100644
--- a/src/main/java/com/github/houbb/sensitive/word/support/result/WordResultHandlerWordTags.java
+++ b/src/main/java/com/github/houbb/sensitive/word/support/result/WordResultHandlerWordTags.java
@@ -19,13 +19,25 @@ public class WordResultHandlerWordTags extends AbstractWordResultHandler<WordTag
     protected WordTagsDto doHandle(IWordResult wordResult, IWordContext wordContext, String originalText) {
         // 截取
         String word = InnerWordCharUtils.getString(originalText.toCharArray(), wordResult);
-        // 标签
 
+        // 创建 DTO 并设置原始单词
         WordTagsDto dto = new WordTagsDto();
         dto.setWord(word);
-        // 获取 tags
-        Set<String> wordTags = InnerWordTagUtils.tags(word, wordContext);
+
+        // 如果启用了字符忽略功能，清理单词后再查找标签
+        StringBuilder wordForTagLookup = new StringBuilder();
+        char[] chars = word.toCharArray();
+        for (int i = 0; i < word.length(); i++) {
+            // 如果字符不被忽略，则保留
+            // TODO: 此处innercontext 被设为null，是否合理？
+            if (!wordContext.charIgnore().ignore(i, chars, null)) {
+                wordForTagLookup.append(chars[i]);
+            }
+        }
+        // 获取 tags (使用清理后的单词查找标签)
+        Set<String> wordTags = InnerWordTagUtils.tags(wordForTagLookup.toString(), wordContext);
         dto.setTags(wordTags);
+
         return dto;
     }
 
diff --git a/src/test/java/com/github/houbb/sensitive/word/bugs/b105/Bug105Test.java b/src/test/java/com/github/houbb/sensitive/word/bugs/b105/Bug105Test.java
new file mode 100644
index 0000000..913d2ec
--- /dev/null
+++ b/src/test/java/com/github/houbb/sensitive/word/bugs/b105/Bug105Test.java
@@ -0,0 +1,45 @@
+package com.github.houbb.sensitive.word.bugs.b105;
+
+import java.util.Arrays;
+import java.util.Collections;
+import java.util.HashSet;
+import java.util.List;
+
+import org.junit.Assert;
+import org.junit.Test;
+
+import com.github.houbb.sensitive.word.bs.SensitiveWordBs;
+import com.github.houbb.sensitive.word.support.ignore.SensitiveWordCharIgnores;
+import com.github.houbb.sensitive.word.support.result.WordResultHandlers;
+import com.github.houbb.sensitive.word.support.result.WordTagsDto;
+import com.github.houbb.sensitive.word.support.tag.WordTags;
+
+/**
+ * 测试Issue #105: 处理带有噪音字符时的标签查找
+ * 验证在启用字符忽略功能时，敏感词标签仍能被正确查找
+ */
+public class Bug105Test {
+
+        @Test
+        public void testNoiseCharacterInTaggedWords() {
+                // 配置同时启用字符忽略和标签的实例
+                SensitiveWordBs ignoreAndTagWordBs = SensitiveWordBs.newInstance()
+                                .charIgnore(SensitiveWordCharIgnores.specialChars()) // 启用字符忽略
+                                .wordTag(WordTags.map(Collections.singletonMap("毛主席",
+                                                new HashSet<>(Arrays.asList("政治", "领导人")))))
+                                .init();
+
+                // 包含噪音字符的敏感词文本
+                final String noisyText = "你好毛---主---席";
+
+                // 测试同时启用字符忽略和标签的实例（修复前会失败）
+                List<WordTagsDto> fixedWord = ignoreAndTagWordBs.findAll(noisyText, WordResultHandlers.wordTags());
+                Assert.assertEquals(1, fixedWord.size());
+                Assert.assertEquals("毛---主---席", fixedWord.get(0).getWord());
+                Assert.assertNotNull("标签不应为空", fixedWord.get(0).getTags());
+                Assert.assertTrue("应包含'政治'标签", fixedWord.get(0).getTags().contains("政治"));
+                Assert.assertTrue("应包含'领导人'标签", fixedWord.get(0).getTags().contains("领导人"));
+
+                System.out.println("Fixed result: " + fixedWord);
+        }
+}