github/sensitive-word
1
0
Fork 0
mirror of https://github.com/houbb/sensitive-word.git synced 2026-03-22 00:17:35 +08:00
Code Issues Packages Projects Releases Wiki Activity

issue110,并废除一些黑白名单一次遍历时的冗余逻辑

Browse Source
This commit is contained in:
yds
2025-05-02 22:05:10 +08:00
parent 2f5c0e72d4
commit bfe3a606c9
7 changed files with 103 additions and 146 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
src/main/java/com/github/houbb/sensitive/word/api/IWordContext.java
View File

@@ -8,6 +8,17 @@ import com.github.houbb.sensitive.word.bs.SensitiveWordContext;
*/
public interface IWordContext {
/**
* 为true时，遇到第一个敏感词词就返回
* 解决issue110
* @return
*/
boolean failFastWordPattern();
IWordContext failFastWordPattern(boolean failFastWordPattern);
/**
* 是否忽略大小写
* @return 是否

9
src/main/java/com/github/houbb/sensitive/word/bs/SensitiveWordBs.java
View File

@@ -2,7 +2,6 @@ package com.github.houbb.sensitive.word.bs;
import com.github.houbb.heaven.support.handler.IHandler;
import com.github.houbb.heaven.util.common.ArgUtil;
import com.github.houbb.heaven.util.lang.StringUtil;
import com.github.houbb.heaven.util.util.CollectionUtil;
import com.github.houbb.sensitive.word.api.*;
import com.github.houbb.sensitive.word.api.combine.IWordAllowDenyCombine;
@@ -70,6 +69,9 @@ public class SensitiveWordBs implements ISensitiveWordDestroy {
*/
private boolean ignoreRepeat = false;
private boolean failFastWordPattern = true;
// 开启校验
/**
* 启用数字检测
@@ -278,6 +280,7 @@ public class SensitiveWordBs implements ISensitiveWordDestroy {
context.ignoreChineseStyle(ignoreChineseStyle);
context.ignoreEnglishStyle(ignoreEnglishStyle);
context.ignoreRepeat(ignoreRepeat);
context.failFastWordPattern(failFastWordPattern);
// 开启校验
context.enableNumCheck(enableNumCheck);
@@ -579,6 +582,10 @@ public class SensitiveWordBs implements ISensitiveWordDestroy {
this.ignoreRepeat = ignoreRepeat;
return this;
}
public SensitiveWordBs failFastWordPattern(boolean failFastWordPattern) {
this.failFastWordPattern = failFastWordPattern;
return this;
}
//------------------------------------------------------------------------------------ 公开方法 START
/**

17
src/main/java/com/github/houbb/sensitive/word/bs/SensitiveWordContext.java
View File

@@ -10,6 +10,12 @@ import com.github.houbb.sensitive.word.support.check.WordChecks;
*/
public class SensitiveWordContext implements IWordContext {
/**
* issue110
*/
private boolean failFastWordPattern;
/**
* 忽略大小写
* @since 0.0.4
@@ -220,6 +226,17 @@ public class SensitiveWordContext implements IWordContext {
return new SensitiveWordContext();
}
@Override
public boolean failFastWordPattern() {
return failFastWordPattern;
}
public IWordContext failFastWordPattern(boolean failFastWordPattern){
this.failFastWordPattern=failFastWordPattern;
return this;
}
@Override
public boolean ignoreCase() {
return ignoreCase;

41
src/main/java/com/github/houbb/sensitive/word/support/check/WordCheckWord.java
View File

@@ -1,14 +1,12 @@
package com.github.houbb.sensitive.word.support.check;
import com.github.houbb.heaven.annotation.ThreadSafe;
import com.github.houbb.heaven.util.lang.StringUtil;
import com.github.houbb.sensitive.word.api.ISensitiveWordCharIgnore;
import com.github.houbb.sensitive.word.api.IWordCheck;
import com.github.houbb.sensitive.word.api.IWordContext;
import com.github.houbb.sensitive.word.api.IWordData;
import com.github.houbb.sensitive.word.api.context.InnerSensitiveWordContext;
import com.github.houbb.sensitive.word.constant.enums.WordTypeEnum;
import com.github.houbb.sensitive.word.constant.enums.WordValidModeEnum;
import com.github.houbb.sensitive.word.constant.enums.WordContainsTypeEnum;
import com.github.houbb.sensitive.word.support.result.WordLengthResult;
@@ -44,47 +42,48 @@ public class WordCheckWord extends AbstractWordCheck {
final IWordData wordData = context.wordData();
final IWordData wordDataAllow = context.wordDataAllow();
final ISensitiveWordCharIgnore wordCharIgnore = context.charIgnore();
final boolean failFast = context.failFastWordPattern();
// 前一个条件
StringBuilder stringBuilder = new StringBuilder();
char[] rawChars = txt.toCharArray();
int tempLen = 0;
int maxWhite = 0;
int maxBlack = 0;
boolean firstCheck = true;
WordContainsTypeEnum wordContainsTypeEnumAllow = wordDataAllow.contains(stringBuilder, innerContext);
WordContainsTypeEnum wordContainsTypeEnumDeny = wordData.contains(stringBuilder, innerContext);
for (int i = beginIndex; i < rawChars.length; i++) {
if (wordCharIgnore.ignore(i, rawChars, innerContext) && tempLen != 0) {
tempLen++;
continue;
}
char mappingChar = formatCharMapping.get(rawChars[i]);
stringBuilder.append(mappingChar);
tempLen++;
if (firstCheck || !WordContainsTypeEnum.NOT_FOUND.equals(wordContainsTypeEnumAllow)) {
wordContainsTypeEnumAllow = wordDataAllow.contains(stringBuilder, innerContext);
if (WordContainsTypeEnum.CONTAINS_END.equals(wordContainsTypeEnumAllow)) {
maxWhite += tempLen;
wordContainsTypeEnumAllow = WordContainsTypeEnum.NOT_FOUND;
WordContainsTypeEnum wordContainsTypeEnumAllow = wordDataAllow.contains(stringBuilder, innerContext);
WordContainsTypeEnum wordContainsTypeEnumDeny = wordData.contains(stringBuilder, innerContext);
if (WordContainsTypeEnum.CONTAINS_END.equals(wordContainsTypeEnumAllow)) {
maxWhite += tempLen;
if (!failFast) {
//此处将tempLen设为0，为了防止重复累加
tempLen = 0;
}else{
//为falFast模式，主动设为notFound退出循环
wordContainsTypeEnumAllow=WordContainsTypeEnum.NOT_FOUND;
}
}
if (firstCheck || !WordContainsTypeEnum.NOT_FOUND.equals(wordContainsTypeEnumDeny)) {
wordContainsTypeEnumDeny = wordData.contains(stringBuilder, innerContext);
if (WordContainsTypeEnum.CONTAINS_END.equals(wordContainsTypeEnumDeny)) {
maxBlack += tempLen;
wordContainsTypeEnumDeny = WordContainsTypeEnum.NOT_FOUND;
if (WordContainsTypeEnum.CONTAINS_END.equals(wordContainsTypeEnumDeny)) {
maxBlack += tempLen;
if (!failFast) {
//此处将tempLen设为0，为了防止重复累加
tempLen = 0;
}else{
//为falFast模式，主动设为notFound退出循环
wordContainsTypeEnumDeny=WordContainsTypeEnum.NOT_FOUND;
}
}
firstCheck = false;
if (WordContainsTypeEnum.NOT_FOUND.equals(wordContainsTypeEnumAllow) &&
WordContainsTypeEnum.NOT_FOUND.equals(wordContainsTypeEnumDeny)) {
break;

96
src/main/java/com/github/houbb/sensitive/word/support/check/WordCheckWordMaxLen.java
View File

@@ -1,96 +0,0 @@
package com.github.houbb.sensitive.word.support.check;
import com.github.houbb.heaven.annotation.ThreadSafe;
import com.github.houbb.sensitive.word.api.ISensitiveWordCharIgnore;
import com.github.houbb.sensitive.word.api.IWordCheck;
import com.github.houbb.sensitive.word.api.IWordContext;
import com.github.houbb.sensitive.word.api.IWordData;
import com.github.houbb.sensitive.word.api.context.InnerSensitiveWordContext;
import com.github.houbb.sensitive.word.constant.enums.WordContainsTypeEnum;
import com.github.houbb.sensitive.word.constant.enums.WordTypeEnum;
import com.github.houbb.sensitive.word.support.result.WordLengthResult;
import java.util.Map;
/**
* 敏感词监测实现
* @author binbin.hou
* @since 0.26.0
*/
@Deprecated
public class WordCheckWordMaxLen extends AbstractWordCheck {
@Override
protected Class<? extends IWordCheck> getSensitiveCheckClass() {
return WordCheckWordMaxLen.class;
}
@Override
protected WordLengthResult getActualLength(int beginIndex, InnerSensitiveWordContext innerContext) {
final String txt = innerContext.originalText();
final Map<Character, Character> formatCharMapping = innerContext.formatCharMapping();
final IWordContext context = innerContext.wordContext();
final IWordData wordData = context.wordData();
final IWordData wordDataAllow = context.wordDataAllow();
final ISensitiveWordCharIgnore wordCharIgnore = context.charIgnore();
// 前一个条件
StringBuilder stringBuilder = new StringBuilder();
char[] rawChars = txt.toCharArray();
int tempLen = 0;
int maxWhite = 0;
int maxBlack = 0;
boolean firstCheck = true;
WordContainsTypeEnum wordContainsTypeEnumAllow = wordDataAllow.contains(stringBuilder, innerContext);
WordContainsTypeEnum wordContainsTypeEnumDeny = wordData.contains(stringBuilder, innerContext);
for (int i = beginIndex; i < rawChars.length; i++) {
if (wordCharIgnore.ignore(i, rawChars, innerContext) && tempLen != 0) {
tempLen++;
continue;
}
char mappingChar = formatCharMapping.get(rawChars[i]);
stringBuilder.append(mappingChar);
tempLen++;
if (firstCheck || !WordContainsTypeEnum.NOT_FOUND.equals(wordContainsTypeEnumAllow)) {
wordContainsTypeEnumAllow = wordDataAllow.contains(stringBuilder, innerContext);
if (WordContainsTypeEnum.CONTAINS_END.equals(wordContainsTypeEnumAllow)) {
maxWhite += tempLen;
wordContainsTypeEnumAllow = WordContainsTypeEnum.NOT_FOUND;
}
}
// 黑名单命中
if (firstCheck || !WordContainsTypeEnum.NOT_FOUND.equals(wordContainsTypeEnumDeny)) {
wordContainsTypeEnumDeny = wordData.contains(stringBuilder, innerContext);
if (WordContainsTypeEnum.CONTAINS_END.equals(wordContainsTypeEnumDeny)) {
maxBlack += tempLen;
wordContainsTypeEnumDeny = WordContainsTypeEnum.NOT_FOUND;
}
}
// 不再是第一次检测
firstCheck = false;
// 黑白名单都未匹配
if (WordContainsTypeEnum.NOT_FOUND.equals(wordContainsTypeEnumAllow) &&
WordContainsTypeEnum.NOT_FOUND.equals(wordContainsTypeEnumDeny)) {
break;
}
}
return WordLengthResult.newInstance()
.wordAllowLen(maxWhite)
.wordDenyLen(maxBlack);
}
@Override
protected String getType() {
return WordTypeEnum.WORD.getCode();
}
}

47
src/test/java/com/github/houbb/sensitive/word/bs/SensitiveWordFailFastTest.java Normal file
View File

@@ -0,0 +1,47 @@
package com.github.houbb.sensitive.word.bs;
import com.github.houbb.sensitive.word.api.IWordDeny;
import org.junit.Assert;
import org.junit.Test;
import java.util.Arrays;
import java.util.List;
public class SensitiveWordFailFastTest {
@Test
public void failFastTest() {
SensitiveWordBs bs = SensitiveWordBs.newInstance()
.failFastWordPattern(true)
.wordDeny(new IWordDeny() {
@Override
public List<String> deny() {
return Arrays.asList("我的世界", "我的");
}
}).init();
String text = "我在我的家里玩我的世界";
List<String> textList = bs.findAll(text);
Assert.assertEquals(Arrays.asList("我的", "我的"), textList);
}
@Test
public void fallOverTest() {
SensitiveWordBs bs = SensitiveWordBs.newInstance()
.failFastWordPattern(false)
.wordDeny(new IWordDeny() {
@Override
public List<String> deny() {
return Arrays.asList("我的世界", "我的");
}
}).init();
String text = "我在我的家里玩我的世界";
List<String> textList = bs.findAll(text);
Assert.assertEquals(Arrays.asList("我的", "我的世界"), textList);
}
}

28
src/test/java/com/github/houbb/sensitive/word/bs/SensitiveWordMaxFirstTest.java
View File

@@ -1,28 +0,0 @@
package com.github.houbb.sensitive.word.bs;
import com.github.houbb.sensitive.word.api.IWordDeny;
import org.junit.Assert;
import org.junit.Test;
import java.util.Arrays;
import java.util.List;
public class SensitiveWordMaxFirstTest {
@Test
public void maxFirstTest() {
SensitiveWordBs bs = SensitiveWordBs.newInstance()
.wordDeny(new IWordDeny() {
@Override
public List<String> deny() {
return Arrays.asList("我的世界", "我的");
}
}).init();
String text = "我的世界我的好玩";
List<String> textList = bs.findAll(text);
// Assert.assertEquals("", textList.toString());
}
}