|
1
|
package eu.dnetlib.data.collector.plugins.excel;
|
|
2
|
|
|
3
|
import java.util.HashMap;
|
|
4
|
import java.util.Iterator;
|
|
5
|
|
|
6
|
import eu.dnetlib.data.collector.plugins.HttpCSVCollectorPlugin;
|
|
7
|
import eu.dnetlib.data.collector.rmi.InterfaceDescriptor;
|
|
8
|
import org.junit.Assert;
|
|
9
|
import org.junit.Before;
|
|
10
|
import org.junit.Ignore;
|
|
11
|
import org.junit.Test;
|
|
12
|
/**
|
|
13
|
* Created by miriam on 10/05/2017.
|
|
14
|
*/
|
|
15
|
@Ignore
|
|
16
|
public class ReadExcelTest {
|
|
17
|
private InterfaceDescriptor descr;
|
|
18
|
private Read r;
|
|
19
|
|
|
20
|
private String excelPath = "eu/dnetlib/data/collector/plugins/excel/fwf_projects_100.xlsx";
|
|
21
|
|
|
22
|
@Before
|
|
23
|
public void setUp() throws Exception {
|
|
24
|
descr = new InterfaceDescriptor();
|
|
25
|
descr.setBaseUrl("https://pf.fwf.ac.at/en/research-in-practice/project-finder.xlsx?&&&search%5Bcall%5D=&search%5Bdecision_board_ids%5D=&search%5Bend_date%5D=&search%5Binstitute_name%5D=&search%5Blead_firstname%5D=&search%5Blead_lastname%5D=&search%5Bper_page%5D=10&search%5Bproject_number%5D=&search%5Bproject_title%5D=&search%5Bscience_discipline_id%5D=&search%5Bstart_date%5D=&search%5Bstatus_id%5D=&search%5Bwhat%5D=&action=index&controller=projects&locale=en&per_page=10" );
|
|
26
|
HashMap<String, String> params = new HashMap<String, String>();
|
|
27
|
|
|
28
|
params.put("argument", "{\"replace\":{\"header\":[{\"from\":\"&\",\"to\":\"and\"}],\"body\":[{\"from\":\"\\n\",\"to\":\" \"}]}}");
|
|
29
|
|
|
30
|
params.put("header_row","4");
|
|
31
|
params.put("tmp_file","//tmp//fwf.xslx");
|
|
32
|
params.put("remove_empty_lines","yes");
|
|
33
|
params.put("remove_lines_with_id"," – ");
|
|
34
|
params.put("col_id","1");
|
|
35
|
params.put("remove_tmp_file","no");
|
|
36
|
params.put("sheet_number","0");
|
|
37
|
params.put("file_to_save","/tmp/project_search.2017.05.10.csv");
|
|
38
|
params.put("separator", ",");
|
|
39
|
params.put("quote","\"");
|
|
40
|
descr.setParams(params);
|
|
41
|
// descr.setBaseUrl("file:///tmp/gsrt_whole.xlsx");
|
|
42
|
// HashMap<String, String> params = new HashMap<String, String>();
|
|
43
|
//
|
|
44
|
// params.put("header_row","0");
|
|
45
|
// params.put("tmp_file","/tmp/ERC.xslx");
|
|
46
|
// params.put("remove_empty_lines","yes");
|
|
47
|
// params.put("remove_lines_with_id"," – ");
|
|
48
|
// params.put("col_id","2");
|
|
49
|
// params.put("remove_tmp_file","no");
|
|
50
|
// params.put("sheet_number","0");
|
|
51
|
// params.put("file_to_save","/tmp/ERC.csv");
|
|
52
|
// params.put("separator", ",");
|
|
53
|
// params.put("quote","\"");
|
|
54
|
// descr.setParams(params);
|
|
55
|
r = new Read(descr);
|
|
56
|
r.setCollector(new HttpCSVCollectorPlugin());
|
|
57
|
}
|
|
58
|
|
|
59
|
@Test
|
|
60
|
//@Ignore
|
|
61
|
public void readExcelFromUrl()throws Exception{
|
|
62
|
Iterator<String> it = r.parseFile().iterator();
|
|
63
|
int i = 0;
|
|
64
|
String st = null;
|
|
65
|
try {
|
|
66
|
while (it.hasNext()) {
|
|
67
|
st = it.next();
|
|
68
|
Assert.assertNotNull(st);
|
|
69
|
i++;
|
|
70
|
//System.out.println(it.next());
|
|
71
|
}
|
|
72
|
}catch(Exception e){
|
|
73
|
System.out.println("linea " + i);
|
|
74
|
e.printStackTrace();
|
|
75
|
System.out.println(st);
|
|
76
|
}
|
|
77
|
|
|
78
|
}
|
|
79
|
|
|
80
|
// @Test
|
|
81
|
// public void readExcelFromFile() throws Exception{
|
|
82
|
// Iterator<String> it = r.parseFile().iterator();
|
|
83
|
// int i =1;
|
|
84
|
// String st = null;
|
|
85
|
// try {
|
|
86
|
// while (it.hasNext()) {
|
|
87
|
// st = it.next();
|
|
88
|
// Assert.assertNotNull(st);
|
|
89
|
// i++;
|
|
90
|
// //System.out.println(it.next());
|
|
91
|
// }
|
|
92
|
// }catch(Exception e){
|
|
93
|
// System.out.println("linea " + i);
|
|
94
|
// e.printStackTrace();
|
|
95
|
// System.out.println(st);
|
|
96
|
// }
|
|
97
|
//
|
|
98
|
//
|
|
99
|
// }
|
|
100
|
}
|