| 1 |
40723
|
giorgos.al
|
package eu.dnetlib.data.mapreduce.hbase.lodImport.utils;
|
| 2 |
|
|
import java.io.BufferedReader;
|
| 3 |
|
|
import java.io.IOException;
|
| 4 |
41002
|
giorgos.al
|
import java.io.InputStream;
|
| 5 |
|
|
import java.io.InputStreamReader;
|
| 6 |
40723
|
giorgos.al
|
|
| 7 |
|
|
/**
|
| 8 |
|
|
* Method: getLangURI
|
| 9 |
|
|
*
|
| 10 |
|
|
* @author Giorgos Alexiou
|
| 11 |
|
|
*
|
| 12 |
|
|
*/
|
| 13 |
|
|
|
| 14 |
|
|
public class MapLanguages {
|
| 15 |
|
|
|
| 16 |
|
|
/**
|
| 17 |
|
|
* This Method takes ISO iso639_3 codes for languages as input and returns the respective URI from lexvo.org as output.
|
| 18 |
|
|
* For example iso639_3 for Greek is "ell" and the respective URI from lexvo is <http://lexvo.org/id/iso639-3/ell>
|
| 19 |
|
|
*
|
| 20 |
|
|
* @param iso639_3_Code
|
| 21 |
|
|
* @return
|
| 22 |
|
|
* @throws IOException
|
| 23 |
|
|
*/
|
| 24 |
|
|
|
| 25 |
41007
|
giorgos.al
|
public MapLanguages(){
|
| 26 |
41002
|
giorgos.al
|
|
| 27 |
41007
|
giorgos.al
|
}
|
| 28 |
|
|
|
| 29 |
|
|
// @SuppressWarnings("resource")
|
| 30 |
41019
|
giorgos.al
|
public String getLangURI(String iso639_3_Code) throws IOException{
|
| 31 |
41007
|
giorgos.al
|
|
| 32 |
41002
|
giorgos.al
|
// File inputFile = new File(MapCountries.class.getClassLoader().getResource("rdfData/lexvo-iso639-3.tsv").getFile());
|
| 33 |
|
|
// InputStream is = new FileInputStream(path);
|
| 34 |
41008
|
eri.katsar
|
//InputStream is = this.getClass().getResourceAsStream("lodImport/lexvo-iso639-3.tsv");
|
| 35 |
|
|
InputStream is = ClassLoader.getSystemResourceAsStream("eu/dnetlib/data/mapreduce/hbase/lodImport/lexvo-iso639-3.tsv");
|
| 36 |
|
|
|
| 37 |
41002
|
giorgos.al
|
// File inputFile = new File(path);
|
| 38 |
|
|
BufferedReader reader = new BufferedReader(new InputStreamReader(is));
|
| 39 |
40723
|
giorgos.al
|
String currentLine;
|
| 40 |
|
|
while((currentLine = reader.readLine()) != null) {
|
| 41 |
|
|
String[] line = currentLine.split("\t");
|
| 42 |
|
|
// System.out.println(line[0].trim()+"--------------"+line[1].trim());
|
| 43 |
|
|
|
| 44 |
|
|
if(line[0].toString().trim().equals(iso639_3_Code)){
|
| 45 |
|
|
return line[1].toString().trim();
|
| 46 |
|
|
}
|
| 47 |
|
|
|
| 48 |
|
|
}
|
| 49 |
|
|
reader.close();
|
| 50 |
|
|
return "";
|
| 51 |
|
|
}
|
| 52 |
|
|
|
| 53 |
|
|
|
| 54 |
|
|
|
| 55 |
|
|
// public static void main(String[] args) throws IOException{
|
| 56 |
|
|
//
|
| 57 |
|
|
// System.out.println(MapLanguages.getLangURI("ell"));
|
| 58 |
|
|
//
|
| 59 |
|
|
// }
|
| 60 |
|
|
|
| 61 |
|
|
|
| 62 |
|
|
}
|