update 拆分时 getChinese改用正则表达式判断中文

This commit is contained in:
lijiarao
2024-05-22 10:18:17 +08:00
parent a5f0b725d1
commit edc1783b22
2 changed files with 8 additions and 3 deletions
@@ -1229,14 +1229,16 @@ public class FileSpiltService {
sarFileSplitItemsValEO.setItemContent(itemContent);
return sarFileSplitItemsValEO;
}
private static final Pattern CHINESE_PATTERN = Pattern.compile("[\u4e00-\u9fa5]");
public String getChinese(String source) {
char[] chars = source.toCharArray();
StringBuilder stringBuilder = new StringBuilder();
for (char aChar : chars) {
int length = String.valueOf(aChar).getBytes().length;
if (length == 3 && aChar != 8195) {
String s = String.valueOf(aChar);
Matcher matcher = CHINESE_PATTERN.matcher(s);
if (matcher.find()) {
stringBuilder.append(aChar);
}
}
@@ -2,7 +2,10 @@ package com.jero.modules.split.util;
import com.aspose.words.*;
import com.jero.modules.split.vo.TitleNumberVO;
import lombok.extern.slf4j.Slf4j;
import org.apache.commons.lang3.StringUtils;
import org.apache.commons.logging.Log;
import org.apache.poi.xwpf.usermodel.VerticalAlign;
import java.io.InputStream;
import java.util.ArrayList;
@@ -14,6 +17,7 @@ import java.util.Map;
* @author liJiaRao
* @date 2023-09-20 14:42
*/
@Slf4j
public class AsposeUtil {
/**
* 获得序号标题和无标题序号
@@ -57,7 +61,6 @@ public class AsposeUtil {
labelString = labelString.replaceAll("\t", "").replaceAll("\b", "")
.replaceAll(" ", "").replaceAll("\"", "").replaceAll("'", "").replaceAll(" ", "");
System.out.println("Text:" + labelString + text);
if (StringUtils.isNotBlank(text)) {
if (titleNumberMap.containsKey(text)) {
List<String> strings = titleNumberMap.get(text);