import java.util.ArrayList;
import java.util.Arrays;
import java.util.HashMap;
import java.util.Iterator;
import java.util.List;
import java.util.Map;
import java.util.Map.Entry;
/**
* 过滤敏感词,并把敏感词替换成*
*
*/
public class SensitiveWordUtils {
static ArrayList first = new ArrayList();
static String[] sortFirst;
static char[] charFirst;
static HashMap> map = new HashMap>();
static HashMap sortMap = new HashMap();
static HashMap charMap = new HashMap();
static ArrayList temp;
static String key, value;
int length;
/**
* 带参数的构造函数
*
* @param keys
* 敏感词
* @param tContent
* 需要过滤的内容
*/
public SensitiveWordUtils(List keys, String tContent) {
for (String k : keys) {
if (!first.contains(k.substring(0, 1))) {
first.add(k.substring(0, 1));
}
length = k.length();
for (int i = 1; i < length; i++) {
key = k.substring(0, i);
value = k.substring(i, i + 1);
if (i == 1 && !first.contains(key)) {
first.add(key);
}
// 有,添加
if (map.containsKey(key)) {
if (!map.get(key).contains(value)) {
map.get(key).add(value);
}
}
// 没有添加
else {
temp = new ArrayList();
temp.add(value);
map.put(key, temp);
}
}
}
sortFirst = first.toArray(new String[first.size()]);
Arrays.sort(sortFirst); // 排序
charFirst = new char[first.size()];
for (int i = 0; i < charFirst.length; i++) {
charFirst[i] = first.get(i).charAt(0);
}
Arrays.sort(charFirst); // 排序
String[] sortValue;
ArrayList v;
Map.Entry> entry;
Iterator>> iter = map.entrySet()
.iterator();
while (iter.hasNext()) {
entry = (Map.Entry>) iter.next();
v = (ArrayList) entry.getValue();
sortValue = v.toArray(new String[v.size()]);
Arrays.sort(sortValue); // 排序
sortMap.put(entry.getKey(), sortValue);
}
char[] charValue;
iter = map.entrySet().iterator();
while (iter.hasNext()) {
entry = (Map.Entry>) iter.next();
v = (ArrayList) entry.getValue();
charValue = new char[v.size()];
for (int i = 0; i < charValue.length; i++) {
charValue[i] = v.get(i).charAt(0);
}
Arrays.sort(charValue); // 排序
charMap.put(entry.getKey(), charValue);
}
}
/**
* 把敏感词替换成*
*
* @param content
* 需要过滤的内容
* @return 过滤完后的符合要求的内容
*/
public String replace(String content) {
String r = null, f, c = content;
String replacedword = content;
char g;
char[] temps;
int length = c.length();
for (int i = 0; i < length - 1; i++) {
g = c.charAt(i);
// 二分查找
if (Arrays.binarySearch(charFirst, g) > -1) {
tag : for (int j = i + 1; j < length; j++) {
f = c.substring(i, j);
g = c.charAt(j);
temps = charMap.get(f);
if (temps == null) { // 找到了
System.out.println("ok");
r = f;
String str = "";
for (int m = 1; m <= r.length(); m++) {
str = str + "*";
}
replacedword = c.replace(r, str);
c = replacedword;
break tag;
}
// 二分查找
if (Arrays.binarySearch(temps, g) > -1) {
if (j == length - 1) {
// print("find!");
System.out.println("find!");
r = c.substring(i, j + 1);
String str = "";
for (int m = 1; m <= r.length(); m++) {
str = str + "*";
}
replacedword = c.replace(r, str);
c = replacedword;
break tag;
}
} else { // 没有找到了
break;
}
}
}
}
return replacedword;
}
}