release branch 0.0.13

src/main/java/com/github/houbb/sensitive/word/api/IWordAllow.java

@@ -0,0 +1,19 @@
+package com.github.houbb.sensitive.word.api;
+
+import java.util.List;
+
+/**
+ * 允许的内容-返回的内容不被当做敏感词
+ * @author binbin.hou
+ * @since 0.0.13
+ */
+public interface IWordAllow {
+
+    /**
+     * 获取结果
+     * @return 结果
+     * @since 0.0.13
+     */
+    List<String> allow();
+
+}

src/main/java/com/github/houbb/sensitive/word/api/IWordData.java

@@ -7,6 +7,7 @@ import java.util.List;
  * @author binbin.hou
  * @since 0.0.1
  */
+@Deprecated
 public interface IWordData {
 
     /**

src/main/java/com/github/houbb/sensitive/word/api/IWordDeny.java

@@ -0,0 +1,19 @@
+package com.github.houbb.sensitive.word.api;
+
+import java.util.List;
+
+/**
+ * 拒绝出现的数据-返回的内容被当做是敏感词
+ * @author binbin.hou
+ * @since 0.0.13
+ */
+public interface IWordDeny {
+
+    /**
+     * 获取结果
+     * @return 结果
+     * @since 0.0.13
+     */
+    List<String> deny();
+
+}

src/main/java/com/github/houbb/sensitive/word/bs/SensitiveWordBs.java

@@ -1,10 +1,13 @@
 package com.github.houbb.sensitive.word.bs;
 
 import com.github.houbb.heaven.constant.CharConst;
-import com.github.houbb.sensitive.word.api.IWordContext;
-import com.github.houbb.sensitive.word.api.IWordData;
-import com.github.houbb.sensitive.word.api.IWordMap;
+import com.github.houbb.heaven.util.common.ArgUtil;
+import com.github.houbb.heaven.util.util.CollectionUtil;
+import com.github.houbb.sensitive.word.api.*;
+import com.github.houbb.sensitive.word.exception.SensitiveWordException;
+import com.github.houbb.sensitive.word.support.allow.WordAllows;
 import com.github.houbb.sensitive.word.support.data.SensitiveWordData;
+import com.github.houbb.sensitive.word.support.deny.WordDenys;
 import com.github.houbb.sensitive.word.support.map.SensitiveWordMap;
 
 import java.util.List;
@@ -30,37 +33,42 @@ public class SensitiveWordBs {
      *
      * @since 0.0.1
      */
-    private static volatile IWordMap sensitiveWordMap;
+    private IWordMap sensitiveWordMap;
 
     /**
      * 默认的执行上下文
      *
      * @since 0.0.4
      */
-    private volatile IWordContext context;
+    private final IWordContext context = buildDefaultContext();
+
+    /**
+     * 禁止的单词
+     * @since 0.0.13
+     */
+    private IWordDeny wordDeny = WordDenys.system();
+
+    /**
+     * 允许的单词
+     * @since 0.0.13
+     */
+    private IWordAllow wordAllow = WordAllows.system();
 
     /**
      * DCL 初始化 wordMap 信息
      *
-     * @return 初始化后的结果
+     * 注意：map 的构建是一个比较耗时的动作
      * @since 0.0.4
      */
-    private static IWordMap initWordMap() {
-        if (sensitiveWordMap == null) {
-            synchronized (IWordMap.class) {
-                if (sensitiveWordMap == null) {
-                    // 加载配置信息
-                    IWordData wordData = new SensitiveWordData();
-                    List<String> lines = wordData.getWordData();
+    private synchronized void initWordMap() {
+        // 加载配置信息
+        List<String> denyList = wordDeny.deny();
+        List<String> allowList = wordAllow.allow();
+        List<String> results = CollectionUtil.difference(denyList, allowList);
 
-                    // 初始化 DFA 信息
-                    sensitiveWordMap = new SensitiveWordMap();
-                    sensitiveWordMap.initWordMap(lines);
-                }
-            }
-        }
-
-        return sensitiveWordMap;
+        // 初始化 DFA 信息
+        sensitiveWordMap = new SensitiveWordMap();
+        sensitiveWordMap.initWordMap(results);
     }
 
     /**
@@ -72,12 +80,44 @@ public class SensitiveWordBs {
      * @since 0.0.1
      */
     public static SensitiveWordBs newInstance() {
-        initWordMap();
+        return new SensitiveWordBs();
+    }
 
-        SensitiveWordBs bs = new SensitiveWordBs();
-        bs.context = buildDefaultContext();
+    /**
+     * 初始化
+     *
+     * 1. 根据配置，初始化对应的 map。比较消耗性能。
+     * @since 0.0.13
+     * @return this
+     */
+    public SensitiveWordBs init() {
+        this.initWordMap();
 
-        return bs;
+        return this;
+    }
+
+    /**
+     * 设置禁止的实现
+     * @param wordDeny 禁止的实现
+     * @return this
+     * @since 0.0.13
+     */
+    public SensitiveWordBs wordDeny(IWordDeny wordDeny) {
+        ArgUtil.notNull(wordDeny, "wordDeny");
+        this.wordDeny = wordDeny;
+        return this;
+    }
+
+    /**
+     * 设置允许的实现
+     * @param wordAllow 允许的实现
+     * @return this
+     * @since 0.0.13
+     */
+    public SensitiveWordBs wordAllow(IWordAllow wordAllow) {
+        ArgUtil.notNull(wordAllow, "wordAllow");
+        this.wordAllow = wordAllow;
+        return this;
     }
 
     /**
@@ -122,7 +162,7 @@ public class SensitiveWordBs {
      * @return 结果
      * @since 0.0.4
      */
-    private static IWordContext buildDefaultContext() {
+    private IWordContext buildDefaultContext() {
         IWordContext wordContext = SensitiveWordContext.newInstance();
         // 格式统一化
         wordContext.ignoreCase(true);
@@ -148,6 +188,8 @@ public class SensitiveWordBs {
      * @since 0.0.1
      */
     public boolean contains(final String target) {
+        statusCheck();
+
         return sensitiveWordMap.contains(target, context);
     }
 
@@ -161,6 +203,8 @@ public class SensitiveWordBs {
      * @since 0.0.1
      */
     public List<String> findAll(final String target) {
+        statusCheck();
+
         return sensitiveWordMap.findAll(target, context);
     }
 
@@ -173,6 +217,8 @@ public class SensitiveWordBs {
      * @since 0.0.1
      */
     public String findFirst(final String target) {
+        statusCheck();
+
         return sensitiveWordMap.findFirst(target, context);
     }
 
@@ -185,6 +231,8 @@ public class SensitiveWordBs {
      * @since 0.0.2
      */
     public String replace(final String target, final char replaceChar) {
+        statusCheck();
+
         return sensitiveWordMap.replace(target, replaceChar, context);
     }
 
@@ -200,4 +248,15 @@ public class SensitiveWordBs {
         return this.replace(target, CharConst.STAR);
     }
 
+
+    /**
+     * 状态校验
+     * @since 0.0.13
+     */
+    private void statusCheck(){
+        if(sensitiveWordMap == null) {
+            this.init();
+        }
+    }
+
 }

src/main/java/com/github/houbb/sensitive/word/core/SensitiveWordHelper.java

@@ -0,0 +1,82 @@
+package com.github.houbb.sensitive.word.core;
+
+import com.github.houbb.sensitive.word.bs.SensitiveWordBs;
+
+import java.util.List;
+
+/**
+ * 敏感词工具类
+ * @author binbin.hou
+ * @since 0.0.13
+ */
+public final class SensitiveWordHelper {
+
+    private SensitiveWordHelper(){}
+
+    /**
+     * 默认的实现
+     * @since 0.0.13
+     */
+    private static final SensitiveWordBs WORD_BS = SensitiveWordBs.newInstance().init();
+
+    /**
+     * 是否包含敏感词
+     *
+     * @param target 目标字符串
+     * @return 是否
+     * @since 0.0.13
+     */
+    public static boolean contains(final String target) {
+        return WORD_BS.contains(target);
+    }
+
+    /**
+     * 返回所有的敏感词
+     * 1. 这里是默认去重的，且是有序的。
+     * 2. 如果不存在，返回空列表
+     *
+     * @param target 目标字符串
+     * @return 敏感词列表
+     * @since 0.0.1
+     */
+    public static List<String> findAll(final String target) {
+        return WORD_BS.findAll(target);
+    }
+
+    /**
+     * 返回第一个敏感词
+     * （1）如果不存在，则返回 {@code null}
+     *
+     * @param target 目标字符串
+     * @return 敏感词
+     * @since 0.0.13
+     */
+    public static String findFirst(final String target) {
+        return WORD_BS.findFirst(target);
+    }
+
+    /**
+     * 替换所有内容
+     *
+     * @param target      目标字符串
+     * @param replaceChar 替换为的 char
+     * @return 替换后结果
+     * @since 0.0.13
+     */
+    public static String replace(final String target, final char replaceChar) {
+        return WORD_BS.replace(target, replaceChar);
+    }
+
+    /**
+     * 替换所有内容
+     * 1. 默认使用空格替换，避免星号改变 md 的格式。
+     *
+     * @param target 目标字符串
+     * @return 替换后结果
+     * @since 0.0.13
+     */
+    public static String replace(final String target) {
+        return WORD_BS.replace(target);
+    }
+
+}

src/main/java/com/github/houbb/sensitive/word/support/allow/WordAllowInit.java

@@ -0,0 +1,43 @@
+package com.github.houbb.sensitive.word.support.allow;
+
+import com.github.houbb.heaven.annotation.ThreadSafe;
+import com.github.houbb.heaven.support.pipeline.Pipeline;
+import com.github.houbb.heaven.support.pipeline.impl.DefaultPipeline;
+import com.github.houbb.sensitive.word.api.IWordAllow;
+
+import java.util.ArrayList;
+import java.util.List;
+
+/**
+ * 初始化类
+ *
+ * @author binbin.hou
+ * @since 0.0.13
+ */
+@ThreadSafe
+public abstract class WordAllowInit implements IWordAllow {
+
+    /**
+     * 初始化列表
+     *
+     * @param pipeline 当前列表泳道
+     * @since 0.0.13
+     */
+    protected abstract void init(final Pipeline<IWordAllow> pipeline);
+
+    @Override
+    public List<String> allow() {
+        Pipeline<IWordAllow> pipeline = new DefaultPipeline<>();
+        this.init(pipeline);
+
+        List<String> results = new ArrayList<>();
+        List<IWordAllow> wordAllows = pipeline.list();
+        for (IWordAllow wordAllow : wordAllows) {
+            List<String> allowList = wordAllow.allow();
+            results.addAll(allowList);
+        }
+
+        return results;
+    }
+
+}

src/main/java/com/github/houbb/sensitive/word/support/allow/WordAllowSystem.java

@@ -0,0 +1,23 @@
+package com.github.houbb.sensitive.word.support.allow;
+
+import com.github.houbb.heaven.annotation.ThreadSafe;
+import com.github.houbb.heaven.util.io.StreamUtil;
+import com.github.houbb.sensitive.word.api.IWordAllow;
+import com.github.houbb.sensitive.word.api.IWordDeny;
+
+import java.util.List;
+
+/**
+ * 系统默认的信息
+ * @author binbin.hou
+ * @since 0.0.13
+ */
+@ThreadSafe
+public class WordAllowSystem implements IWordAllow {
+
+    @Override
+    public List<String> allow() {
+        return StreamUtil.readAllLines("/sensitive_word_allow.txt");
+    }
+
+}

src/main/java/com/github/houbb/sensitive/word/support/allow/WordAllows.java

@@ -0,0 +1,49 @@
+package com.github.houbb.sensitive.word.support.allow;
+
+import com.github.houbb.heaven.support.instance.impl.Instances;
+import com.github.houbb.heaven.support.pipeline.Pipeline;
+import com.github.houbb.heaven.util.util.ArrayUtil;
+import com.github.houbb.sensitive.word.api.IWordAllow;
+
+/**
+ * 所有允许的结果
+ * @author binbin.hou
+ * @since 0.0.13
+ */
+public final class WordAllows {
+
+    private WordAllows(){}
+
+    /**
+     * 责任链
+     * @param wordAllow 允许
+     * @param others 其他
+     * @return 结果
+     * @since 0.0.13
+     */
+    public static IWordAllow chains(final IWordAllow wordAllow,
+                                    final IWordAllow... others) {
+        return new WordAllowInit() {
+            @Override
+            protected void init(Pipeline<IWordAllow> pipeline) {
+                pipeline.addLast(wordAllow);
+
+                if(ArrayUtil.isNotEmpty(others)) {
+                    for(IWordAllow other : others) {
+                        pipeline.addLast(other);
+                    }
+                }
+            }
+        };
+    }
+
+    /**
+     * 系统实现
+     * @return 结果
+     * @since 0.0.13
+     */
+    public static IWordAllow system() {
+        return Instances.singleton(WordAllowSystem.class);
+    }
+
+}

src/main/java/com/github/houbb/sensitive/word/support/deny/WordDenyInit.java

@@ -0,0 +1,44 @@
+package com.github.houbb.sensitive.word.support.deny;
+
+import com.github.houbb.heaven.annotation.ThreadSafe;
+import com.github.houbb.heaven.support.pipeline.Pipeline;
+import com.github.houbb.heaven.support.pipeline.impl.DefaultPipeline;
+import com.github.houbb.heaven.util.io.StreamUtil;
+import com.github.houbb.sensitive.word.api.IWordDeny;
+
+import java.util.ArrayList;
+import java.util.List;
+
+/**
+ * 初始化类
+ *
+ * @author binbin.hou
+ * @since 0.0.13
+ */
+@ThreadSafe
+public abstract class WordDenyInit implements IWordDeny {
+
+    /**
+     * 初始化列表
+     *
+     * @param pipeline 当前列表泳道
+     * @since 0.0.13
+     */
+    protected abstract void init(final Pipeline<IWordDeny> pipeline);
+
+    @Override
+    public List<String> deny() {
+        Pipeline<IWordDeny> pipeline = new DefaultPipeline<>();
+        this.init(pipeline);
+
+        List<String> results = new ArrayList<>();
+        List<IWordDeny> wordDenies = pipeline.list();
+        for (IWordDeny wordDeny : wordDenies) {
+            List<String> denyList = wordDeny.deny();
+            results.addAll(denyList);
+        }
+
+        return results;
+    }
+
+}

src/main/java/com/github/houbb/sensitive/word/support/deny/WordDenySystem.java

@@ -0,0 +1,27 @@
+package com.github.houbb.sensitive.word.support.deny;
+
+import com.github.houbb.heaven.annotation.ThreadSafe;
+import com.github.houbb.heaven.util.guava.Guavas;
+import com.github.houbb.heaven.util.io.StreamUtil;
+import com.github.houbb.sensitive.word.api.IWordDeny;
+import com.github.houbb.sensitive.word.constant.AppConst;
+
+import java.util.List;
+
+/**
+ * 系统默认的信息
+ * @author binbin.hou
+ * @since 0.0.13
+ */
+@ThreadSafe
+public class WordDenySystem implements IWordDeny {
+
+    @Override
+    public List<String> deny() {
+        List<String> results = StreamUtil.readAllLines("/dict.txt");
+        results.addAll(StreamUtil.readAllLines("/dict_en.txt"));
+        results.addAll(StreamUtil.readAllLines("/sensitive_word_deny.txt"));
+        return results;
+    }
+
+}

src/main/java/com/github/houbb/sensitive/word/support/deny/WordDenys.java

@@ -0,0 +1,49 @@
+package com.github.houbb.sensitive.word.support.deny;
+
+import com.github.houbb.heaven.support.instance.impl.Instances;
+import com.github.houbb.heaven.support.pipeline.Pipeline;
+import com.github.houbb.heaven.util.util.ArrayUtil;
+import com.github.houbb.sensitive.word.api.IWordDeny;
+
+/**
+ * 所有拒绝的结果
+ * @author binbin.hou
+ * @since 0.0.13
+ */
+public final class WordDenys {
+
+    private WordDenys(){}
+
+    /**
+     * 责任链
+     * @param wordDeny 拒绝
+     * @param others 其他
+     * @return 结果
+     * @since 0.0.13
+     */
+    public static IWordDeny chains(final IWordDeny wordDeny,
+                                   final IWordDeny... others) {
+        return new WordDenyInit() {
+            @Override
+            protected void init(Pipeline<IWordDeny> pipeline) {
+                pipeline.addLast(wordDeny);
+
+                if(ArrayUtil.isNotEmpty(others)) {
+                    for(IWordDeny other : others) {
+                        pipeline.addLast(other);
+                    }
+                }
+            }
+        };
+    }
+
+    /**
+     * 系统实现
+     * @return 结果
+     * @since 0.0.13
+     */
+    public static IWordDeny system() {
+        return Instances.singleton(WordDenySystem.class);
+    }
+
+}

src/main/resources/dict.txt

@@ -1,4 +1,3 @@
-
 001工程
 007手机防盗软件任意显软件
 007间谍专业版

src/test/java/com/github/houbb/sensitive/word/bs/SensitiveWordBsTest.java

@@ -1,5 +1,7 @@
 package com.github.houbb.sensitive.word.bs;
 
+import com.github.houbb.sensitive.word.support.allow.WordAllows;
+import com.github.houbb.sensitive.word.support.deny.WordDenys;
 import org.junit.Assert;
 import org.junit.Test;
 
@@ -97,4 +99,15 @@ public class SensitiveWordBsTest {
         Assert.assertEquals("ｆｕｃｋ", word);
     }
 
+    @Test
+    public void configTest() {
+        SensitiveWordBs wordBs = SensitiveWordBs.newInstance()
+                .wordDeny(WordDenys.system())
+                .wordAllow(WordAllows.system())
+                .init();
+
+        final String text = "五星红旗迎风飘扬，毛主席的画像屹立在天安门前。";
+        Assert.assertTrue(wordBs.contains(text));
+    }
+
 }

src/test/java/com/github/houbb/sensitive/word/core/SensitiveWordHelperTest.java

@@ -0,0 +1,100 @@
+package com.github.houbb.sensitive.word.core;
+
+import org.junit.Assert;
+import org.junit.Test;
+
+import java.util.List;
+
+/**
+ * <p> project: sensitive-word-SensitiveWordBsTest </p>
+ * <p> create on 2020/1/7 23:43 </p>
+ *
+ * @author Administrator
+ * @since 0.0.13
+ */
+public class SensitiveWordHelperTest {
+
+    /**
+     * 是否包含
+     * @since 0.0.1
+     */
+    @Test
+    public void containsTest() {
+        final String text = "五星红旗迎风飘扬，毛主席的画像屹立在天安门前。";
+
+        Assert.assertTrue(SensitiveWordHelper.contains(text));
+    }
+
+    /**
+     * 返回所有敏感词
+     * @since 0.0.1
+     */
+    @Test
+    public void findAllTest() {
+        final String text = "五星红旗迎风飘扬，毛主席的画像屹立在天安门前。";
+
+        List<String> wordList = SensitiveWordHelper.findAll(text);
+        Assert.assertEquals("[五星红旗, 毛主席, 天安门]", wordList.toString());
+    }
+
+    /**
+     * 返回所有第一个匹配的敏感词
+     * @since 0.0.1
+     */
+    @Test
+    public void findFirstTest() {
+        final String text = "五星红旗迎风飘扬，毛主席的画像屹立在天安门前。";
+
+        String word = SensitiveWordHelper.findFirst(text);
+        Assert.assertEquals("五星红旗", word);
+    }
+
+    /**
+     * 默认的替换策略
+     * @since 0.0.2
+     */
+    @Test
+    public void replaceTest() {
+        final String text = "五星红旗迎风飘扬，毛主席的画像屹立在天安门前。";
+
+        String result = SensitiveWordHelper.replace(text);
+        Assert.assertEquals("****迎风飘扬，***的画像屹立在***前。", result);
+    }
+
+    /**
+     * 自定义字符的替换策略
+     * @since 0.0.2
+     */
+    @Test
+    public void replaceCharTest() {
+        final String text = "五星红旗迎风飘扬，毛主席的画像屹立在天安门前。";
+
+        String result = SensitiveWordHelper.replace(text, '0');
+        Assert.assertEquals("0000迎风飘扬，000的画像屹立在000前。", result);
+    }
+
+    /**
+     * 忽略大小写
+     * @since 0.0.4
+     */
+    @Test
+    public void ignoreCaseTest() {
+        final String text = "fuCK the bad words.";
+
+        String word = SensitiveWordHelper.findFirst(text);
+        Assert.assertEquals("fuCK", word);
+    }
+
+    /**
+     * 忽略半角圆角
+     * @since 0.0.4
+     */
+    @Test
+    public void ignoreWidthTest() {
+        final String text = "ｆｕｃｋ the bad words.";
+
+        String word = SensitiveWordHelper.findFirst(text);
+        Assert.assertEquals("ｆｕｃｋ", word);
+    }
+
+}