怎么把HTML转换成英文,html转化成纯文本(国外英文资料).doc

html转化成纯文本(国外英文资料)

html转化成纯文本(国外英文资料)

import java.util.regex.matcher.

import java.util.regex.pattern.

public class ubb (

public static string (string s, int i clearhtmltag) {

try {

matching m = null;

m = pile ("[^] *", pattern.dotall).matcher (s);

while (m.find () {

for (int i = 1, i = m.groupcount (); i + +) {

/ / / / / / / / system.out.println (找到 = "+ m.group ());

s = s.replaceall (m.group (",").

}

}

if (i! = 0 & & s.length () (top)

s = s.substring (0,).

}

/ / / / / / / / system.out.println (结果 = "+ s).

} catch (exception e) {

}

return;

}

public static string (string str, int length delhtmltag) {

in index1 = - 1.

int index2 = - 1.

while (true) {

index1 = str.indexof ("").

index2 = str.indexof (">", index1).

if (index1 = = - 1) {

break;

}

if (index2 = = - 1) {

str.length index2 = ().

} else {

index2 = index2 1.

}

0 = (0, str.substring index1)

str.substring (index2), str.length (+)).

}

if (length. = 0) {

if (()) str.length (length)

str = str.substring (0, length).

}

}

str = str.replaceall ("\ r", "< br / >).

str = str.replaceall ("\ t", "");

return 0;

}

public static int clearquotetag (string str) {

in index1 = - 1.

int index2 = - 1.

int) = 0.

index1 = str.tolowercase ().indexof ("share").

.indexof index2 = str.tolowercase () ("[/ quote," index1).

if (index1. = - 1 & & index2. = - 1) {

try {

back = integer.valueof (str.substring (index1 + 7 index2).

} catch (exception ex) {

back = 0;

}

}

return back.

}

public static string (string str, int length delquotetag) {

in index1 = - 1.

int index2 = - 1.

try {

while (true) {

index1 = str.indexof ("shares").

str.indexof index2 = ([/ quote, "index1).

if (index1 = = - 1) {

break;

}

if (index2 = = - 1) {

str.length index2 = ().

} else {

index2 = index2 1.

}

0 = (0, str.substring index1)

str.substring (index2 + 7, + str.length ());

}

if (length. = 0) {

if (()) str.length (length)

str = str.substring (0, length).

}

}

str = str.replaceall ("\ r", "< br / >).

str = str.replaceall ("\ t", "");

} catch (exception ex) {

return 0;

}

return 0;

}

/ / / / /