不使用NLP或者地址解析工具的情况下，主要根据正则表达式来实现地址的解析和手机号、姓名的提取。
DeliveryInfoMapParser
package y2023.m10.d09.parse;
import lombok.extern.slf4j.Slf4j;
import y2023.m10.d09.entity.DeliveryInfo;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.List;
/**
* 快递信息解析工具包：淘宝、当当、京东、微信、
*/
@Slf4j
public class DeliveryInfoMapParser {
private static final String[] nameKeys = new String[]{"收货人", "联系人", "姓名"};
private static final String[] phoneKeys = new String[]{"手机号码", "手机", "号码", "联系电话", "电话"};
private static final String[] addressKeys = new String[]{"所在地区", "地区"};
private static final String[] detailKeys = new String[]{"详细地址", "收货地址", "地址"};
public static DeliveryInfo mapParser(String text) {
// 半角符号专为全角符号
if (text.contains(":")) {
text = text.replace(":", "：");
}
// 如果不包含冒号，直接调用单行数据解析逻辑
if (!text.contains("：")) {
DeliveryInfoLineParser.lineParser(text);
}
// 分组
String[] temp = text.split("\n");
// 检验分组数据是否规范：规范数据的话，一行一个冒号
ArrayList<String> subTextList = new ArrayList<>();
if (temp.length == ParseUtil.appearNumber(text, "：")) {
subTextList.addAll(Arrays.asList(temp));
}
// 不规范数据：冒号前面的第一个空格符号就是分隔符
else {
for (String s : temp) {
split(subTextList, s);
}
}
// 数据分行准备完毕，下面进行数据解析
DeliveryInfo info = new DeliveryInfo();
info.setText(text);
// 解析姓名
info.setName(parseFromKV(nameKeys, subTextList));
// 解析手机
info.setPhone(parseFromKV(phoneKeys, subTextList));
// 解析地区
info.setTempText(parseFromKV(addressKeys, subTextList));
// 解析详细地址
info.setTempText(info.getTempText() + parseFromKV(detailKeys, subTextList));
// 解析省市区
ParseUtil.addressResolution(info);
ParseUtil.cleanValue(info);
return info;
}
private static String parseFromKV(String[] nameKeys, ArrayList<String> subTextList) {
for (String key : nameKeys) {
for (String subText : subTextList) {
if (subText.contains(key)) {
return getValueByKey(subText);
}
}
}
return "";
}
private static String getValueByKey(String subText) {
if (subText.contains("：")) {
return subText.substring(subText.indexOf("：") + 1);
}
return subText;
}
private static void split(List<String> temp, String subText) {
if (subText.indexOf("：") != subText.lastIndexOf("：")) {
int splitIndex = subText.substring(0, subText.lastIndexOf("：")).lastIndexOf(" ");
if (splitIndex < 0) {
splitIndex = subText.substring(0, subText.lastIndexOf("：")).lastIndexOf("，");
}
if (splitIndex < 0) {
splitIndex = subText.substring(0, subText.lastIndexOf("：")).lastIndexOf(",");
}
// 后面的加入队列
temp.add(subText.substring(splitIndex));
// 前面的继续递归解析
split(temp, subText.substring(0, splitIndex));
} else {
temp.add(subText);
}
}
}
DeliveryInfoLineParser
package y2023.m10.d09.parse;
import lombok.extern.slf4j.Slf4j;
import org.apache.commons.lang3.StringUtils;
import y2023.m10.d09.entity.DeliveryInfo;
import java.util.ArrayList;
import java.util.regex.Matcher;
import java.util.regex.Pattern;
/**
* 快递信息解析工具包
*/
@Slf4j
public class DeliveryInfoLineParser {
public static DeliveryInfo lineParser(String text) {
DeliveryInfo info = new DeliveryInfo();
info.setText(text);
// 如果有手机号，那先提取手机号：1开头的11位数字就是手机号，不做其他验证
ArrayList<ParseNode> phones = getPhone(text);
if (phones.size() == 1) {
ParseNode phone = phones.get(0);
info.setPhone(phone.getValue());
// 以手机号位基准解析信息，手机号左右是可能的姓名或者地址信息
splitNameAndAddress(info, phone);
// 尝试解析姓名信息
String nameStr = info.getName();
if (StringUtils.isNotEmpty(nameStr)) {
info.setName(nameStr.trim());
}
} else {
info.setTempText(text);
}
// 正则匹配省市区
ParseUtil.addressResolution(info);
// 再次验证并解析姓名
if (StringUtils.isEmpty(info.getName())) {
parseNameFromAddressFront(info);
}
if (StringUtils.isEmpty(info.getName())) {
parseNameFromAddressEnd(info);
}
ParseUtil.cleanValue(info);
return info;
}
private static void splitNameAndAddress(DeliveryInfo info, ParseNode phone) {
String text = info.getText();
// 包含省或者市的那一端是地址
String start = text.substring(0, phone.getStart()).trim();
String end = text.substring(phone.getEnd()).trim();
if ((start.contains("省") || start.contains("市")) && (!end.contains("省") && !end.contains("市"))) {
info.setTempText(start);
info.setName(end);
} else if ((end.contains("省") || end.contains("市")) && (!start.contains("省") && !start.contains("市"))) {
info.setTempText(end);
info.setName(start);
} else if ((start.contains("省") && start.contains("市")) && (!end.contains("省") || !end.contains("市"))) {
info.setTempText(start);
info.setName(end);
} else if ((end.contains("省") && end.contains("市")) && (!start.contains("省") || !start.contains("市"))) {
info.setTempText(end);
info.setName(start);
} else {
info.setTempText(start + " " + end);
}
}
private static final String[] separators = new String[]{",", "，", " "};
// demo: 坚果同学 南京市栖霞区栖霞街道人民广场399号 13851764116
private static void parseNameFromAddressFront(DeliveryInfo info) {
String addressStr;
if (StringUtils.isNotEmpty(info.getProvince())) {
addressStr = info.getProvince();
} else if (StringUtils.isNotEmpty(info.getCity())) {
addressStr = info.getCity();
} else {
return;
}
// 解析 name
String addressVal = null;
for (String s : separators) {
if (addressStr.contains(s)) {
String nameVal = ParseUtil.getValue(addressStr.substring(0, addressStr.indexOf(s)));
info.setName(nameVal);
addressVal = ParseUtil.getValue(addressStr.substring(addressStr.indexOf(s)));
}
}
if (StringUtils.isEmpty(addressVal)) {
addressVal = addressStr;
}
if (StringUtils.isNotEmpty(info.getProvince())) {
info.setProvince(addressVal);
} else if (StringUtils.isNotEmpty(info.getCity())) {
info.setCity(addressVal);
}
}
private static void parseNameFromAddressEnd(DeliveryInfo info) {
String addressStr = info.getDetail();
if (StringUtils.isEmpty(addressStr)) {
return;
}
for (String s : separators) {
if (addressStr.contains(s)) {
String nameStr = addressStr.substring(addressStr.indexOf(s));
info.setName(ParseUtil.getValue(nameStr));
String addressVal = ParseUtil.getValue(addressStr.substring(0, addressStr.indexOf(s)));
info.setDetail(addressVal);
}
}
}
// 提取文本手机号
private static ArrayList<ParseNode> getPhone(String text) {
String regex = "[1][\\d]{10}";
Pattern compile = Pattern.compile(regex);
Matcher matcher = compile.matcher(text);
ArrayList<ParseNode> phones = new ArrayList<>();
while (matcher.find()) {
String phone = matcher.group();
int start = matcher.start();
int end = matcher.end();
// 手机号在最后的场景
if (text.endsWith(phone)) {
ParseNode node = new ParseNode(phone, start, end);
phones.add(node);
}
// 11位数字后面不是数字的才是手机号
else if (!Character.isDigit(text.charAt(end))) {
ParseNode node = new ParseNode(phone, start, end);
phones.add(node);
}
}
return phones;
}
}
ParseNode
package y2023.m10.d09.parse;
import lombok.AllArgsConstructor;
import lombok.Data;
@Data
@AllArgsConstructor
public class ParseNode {
private String value;
private Integer start;
private Integer end;
}
ParseUtil
package y2023.m10.d09.parse;
import lombok.extern.slf4j.Slf4j;
import org.apache.commons.lang3.StringUtils;
import y2023.m10.d09.entity.DeliveryInfo;
import java.util.ArrayList;
import java.util.List;
import java.util.regex.Matcher;
import java.util.regex.Pattern;
@Slf4j
public class ParseUtil {
private static final List<String> directlyCities;
static {
directlyCities = new ArrayList<>();
directlyCities.add("北京");
directlyCities.add("上海");
directlyCities.add("天津");
directlyCities.add("重庆");
}
public static void addressResolution(DeliveryInfo info) {
String text = info.getTempText().trim();
// 直辖市特殊情况处理 => fix：上海徐汇区柳州路399弄G层 毛千
for (String directCity : directlyCities) {
if (text.indexOf(directCity) != -1 && text.indexOf(directCity) == text.lastIndexOf(directCity)) {
text = directCity + text;
}
}
// 不要随意调动正则顺序，避免特殊行政区划识别异常；fix：京东拷贝的地址没有省关键字，所以需要枚举所有省份信息用于匹配省份
String regex = "(?<province>北京市|北京|天津市|天津|上海市|上海|重庆市|重庆|香港|澳门|[^省]+省|[^自治区]+自治区|内蒙古|广西|西藏|宁夏|新疆|河北|山西|辽宁|吉林|黑龙江|江苏|浙江|安徽|福建|江西|山东|河南|湖北|湖南|广东|海南|四川|贵州|云南|陕西|甘肃|青海)?" +
"(?<city>上海市|上海|北京市|北京|天津市|天津|重庆市|重庆|香港|澳门|[^市]+市|[^州]+州|[^特别行政区]+特别行政区|[^县]+县|.*?地区|.*?行政单位|.+盟|市辖区)" +
"(?<area>[^区]+区|[^县]+县|[^镇]+镇|[^乡]+乡|.+场|.+旗|.+海域|.+岛)?" +
"(?<town>[^街道]+街道|[^镇]+镇|[^乡]+乡)?" +
// fix：详细地址匹配包括换行符在内的任何字符："[\s\S]*" 而不是 ".*"
"(?<village>[\\s\\S]*)?";
Matcher m = Pattern.compile(regex).matcher(text);
String provinceVal, cityVal, areaVal, townVal, villageVal;
while (m.find()) {
provinceVal = m.group("province");
info.setProvince(provinceVal == null ? "" : provinceVal.trim());
cityVal = m.group("city");
info.setCity(cityVal == null ? "" : cityVal.trim());
areaVal = m.group("area");
info.setArea(areaVal == null ? "" : areaVal.trim());
townVal = m.group("town");
info.setTown(townVal == null ? "" : townVal.trim());
villageVal = m.group("village");
info.setDetail(villageVal == null ? "" : villageVal.trim());
}
}
public static void cleanValue(DeliveryInfo info) {
// 最终数据清洗
info.setName(ParseUtil.getValue(info.getName()));
info.setPhone(ParseUtil.getValue(info.getPhone()));
info.setProvince(ParseUtil.getValue(info.getProvince()));
info.setCity(ParseUtil.getValue(info.getCity()));
info.setArea(ParseUtil.getValue(info.getArea()));
info.setTown(ParseUtil.getValue(info.getTown()));
info.setDetail(ParseUtil.getValue(info.getDetail()));
info.setTempText(null);
log.info("地址解析：{} => 姓名=[{}],手机=[{}],省=[{}],市=[{}],区县=[{}],街道=[{}],详细地址=[{}]",
info.getText(),info.getName(),info.getPhone(),
info.getProvince(), info.getCity(), info.getArea(),
info.getTown(), info.getDetail());
}
public static String getValue(String nodeText) {
return StringUtils.isEmpty(nodeText) ? nodeText :
nodeText.replace("\n", "")
.replace("\t", "")
.replace(",", "")
.replace("，", "")
.replace(" ", "")
.trim();
}
public static int appearNumber(String srcText, String findText) {
int count = 0;
Pattern p = Pattern.compile(findText);
Matcher m = p.matcher(srcText);
while (m.find()) {
count++;
}
return count;
}
}
参考资料
开始写作吧
文档信息
- 本文作者：Bob.Zhu
- 本文链接：https://adolphor.github.io/2023/10/09/01-address-parser/
- 版权声明：自由转载-非商用-非衍生-保持署名（创意共享3.0许可证）