1
|
package eu.dnetlib.data.collector.plugins;
|
2
|
|
3
|
import java.io.BufferedReader;
|
4
|
import java.io.FileInputStream;
|
5
|
import java.io.IOException;
|
6
|
import java.io.InputStreamReader;
|
7
|
import java.net.MalformedURLException;
|
8
|
import java.net.URL;
|
9
|
import java.util.Iterator;
|
10
|
|
11
|
import org.apache.commons.io.input.BOMInputStream;
|
12
|
import org.apache.commons.lang3.StringEscapeUtils;
|
13
|
import org.apache.commons.lang3.StringUtils;
|
14
|
import org.apache.commons.logging.Log;
|
15
|
import org.apache.commons.logging.LogFactory;
|
16
|
import org.dom4j.Document;
|
17
|
import org.dom4j.DocumentHelper;
|
18
|
import org.dom4j.Element;
|
19
|
|
20
|
import eu.dnetlib.data.collector.plugin.AbstractCollectorPlugin;
|
21
|
import eu.dnetlib.data.collector.rmi.CollectorServiceException;
|
22
|
import eu.dnetlib.data.collector.rmi.InterfaceDescriptor;
|
23
|
|
24
|
/**
|
25
|
* Please use eu.dnetlib.data.collector.plugins.HttpCSVCollectorPlugin instead
|
26
|
*/
|
27
|
@Deprecated
|
28
|
public class FileCSVCollectorPlugin extends AbstractCollectorPlugin {
|
29
|
|
30
|
private static final Log log = LogFactory.getLog(FileCSVCollectorPlugin.class);
|
31
|
|
32
|
class FileCSVIterator implements Iterator<String> {
|
33
|
|
34
|
private String next;
|
35
|
|
36
|
private BufferedReader reader;
|
37
|
|
38
|
private String separator;
|
39
|
private String quote;
|
40
|
|
41
|
public FileCSVIterator(final BufferedReader reader, final String separator, final String quote) {
|
42
|
this.reader = reader;
|
43
|
this.separator = separator;
|
44
|
this.quote = quote;
|
45
|
next = calculateNext();
|
46
|
}
|
47
|
|
48
|
@Override
|
49
|
public boolean hasNext() {
|
50
|
return next != null;
|
51
|
}
|
52
|
|
53
|
@Override
|
54
|
public String next() {
|
55
|
final String s = next;
|
56
|
next = calculateNext();
|
57
|
return s;
|
58
|
}
|
59
|
|
60
|
private String calculateNext() {
|
61
|
try {
|
62
|
final Document document = DocumentHelper.createDocument();
|
63
|
final Element root = document.addElement("csvRecord");
|
64
|
|
65
|
String newLine = reader.readLine();
|
66
|
|
67
|
// FOR SOME FILES IT RETURN NULL ALSO IF THE FILE IS NOT READY DONE
|
68
|
if (newLine == null) {
|
69
|
newLine = reader.readLine();
|
70
|
}
|
71
|
if (newLine == null) {
|
72
|
log.info("there is no line, closing RESULT SET");
|
73
|
|
74
|
reader.close();
|
75
|
return null;
|
76
|
}
|
77
|
final String[] currentRow = newLine.split(separator);
|
78
|
|
79
|
if (currentRow != null) {
|
80
|
|
81
|
for (int i = 0; i < currentRow.length; i++) {
|
82
|
final String hAttribute = (headers != null) && (i < headers.length) ? headers[i] : "column" + i;
|
83
|
|
84
|
final Element row = root.addElement("column");
|
85
|
if (i == identifierNumber) {
|
86
|
row.addAttribute("isID", "true");
|
87
|
}
|
88
|
final String value = StringUtils.isBlank(quote) ? currentRow[i] : StringUtils.strip(currentRow[i], quote);
|
89
|
|
90
|
row.addAttribute("name", hAttribute).addText(value);
|
91
|
}
|
92
|
return document.asXML();
|
93
|
}
|
94
|
} catch (final IOException e) {
|
95
|
log.error("Error calculating next csv element", e);
|
96
|
}
|
97
|
return null;
|
98
|
}
|
99
|
|
100
|
@Override
|
101
|
public void remove() {
|
102
|
throw new UnsupportedOperationException();
|
103
|
}
|
104
|
|
105
|
}
|
106
|
|
107
|
private String[] headers = null;
|
108
|
private int identifierNumber;
|
109
|
|
110
|
@Override
|
111
|
public Iterable<String> collect(final InterfaceDescriptor interfaceDescriptor, final String fromDate, final String untilDate)
|
112
|
throws CollectorServiceException {
|
113
|
final String header = interfaceDescriptor.getParams().get("header");
|
114
|
final String separator = StringEscapeUtils.unescapeJava(interfaceDescriptor.getParams().get("separator"));
|
115
|
final String quote = interfaceDescriptor.getParams().get("quote");
|
116
|
|
117
|
identifierNumber = Integer.parseInt(interfaceDescriptor.getParams().get("identifier"));
|
118
|
URL u = null;
|
119
|
try {
|
120
|
u = new URL(interfaceDescriptor.getBaseUrl());
|
121
|
} catch (final MalformedURLException e1) {
|
122
|
throw new CollectorServiceException(e1);
|
123
|
}
|
124
|
final String baseUrl = u.getPath();
|
125
|
|
126
|
log.info("base URL = " + baseUrl);
|
127
|
|
128
|
try {
|
129
|
|
130
|
final BufferedReader br = new BufferedReader(new InputStreamReader(new BOMInputStream(new FileInputStream(baseUrl))));
|
131
|
|
132
|
if ((header != null) && "true".equals(header.toLowerCase())) {
|
133
|
final String[] tmpHeader = br.readLine().split(separator);
|
134
|
if (StringUtils.isNotBlank(quote)) {
|
135
|
int i = 0;
|
136
|
headers = new String[tmpHeader.length];
|
137
|
for (final String h : tmpHeader) {
|
138
|
headers[i] = StringUtils.strip(h, quote);
|
139
|
i++;
|
140
|
}
|
141
|
} else headers = tmpHeader;
|
142
|
}
|
143
|
return () -> new FileCSVIterator(br, separator, quote);
|
144
|
} catch (final Exception e) {
|
145
|
throw new CollectorServiceException(e);
|
146
|
}
|
147
|
}
|
148
|
|
149
|
}
|