D-Net

package eu.dnetlib.data.mapreduce.hbase.lodExport.utils;

import java.net.URLDecoder;

import java.util.ArrayList;

import java.util.Set;

import com.google.common.collect.Multimap;

public class Blocking {

	public static ArrayList<String> tokenBlocking(Multimap<String, String >fields){

		ArrayList<String> blockingKeys = new ArrayList<>();

		Set<String> keys = fields.keySet();

		for(String key : keys){

			if(key.equals("id")) continue;

			for(String value : fields.get(key) ){

				try{

                    value = URLDecoder.decode(value, "UTF-8").replaceAll("_", " ");

                } catch (Exception ex) {

                    value = value.replaceAll("_", " ");

                }

				String[] tokens = value.split("[\\W_]");

					for (int i = 0; i < tokens.length; i++) {

                   String currentToken = tokens[i].trim();

                   if (0 < currentToken.length()) {

                      blockingKeys.add(currentToken);

                   }

				}

			}

		}

		return blockingKeys;

	}

}