svn-gvsig-desktop / trunk / org.gvsig.desktop / org.gvsig.desktop.compat.cdc / org.gvsig.fmap.dal / org.gvsig.fmap.dal.file / org.gvsig.fmap.dal.file.csv / src / main / java / org / gvsig / fmap / dal / store / gml / virtualrows / XmlCommons.java @ 47638
History | View | Annotate | Download (5.02 KB)
1 |
package org.gvsig.fmap.dal.store.gml.virtualrows; |
---|---|
2 |
|
3 |
import java.io.BufferedReader; |
4 |
import java.io.File; |
5 |
import java.io.FileInputStream; |
6 |
import java.io.InputStream; |
7 |
import java.io.InputStreamReader; |
8 |
import java.nio.charset.Charset; |
9 |
import org.apache.commons.io.IOUtils; |
10 |
import org.apache.commons.io.input.BOMInputStream; |
11 |
import org.apache.commons.lang3.StringUtils; |
12 |
import org.apache.tika.detect.AutoDetectReader; |
13 |
import org.gvsig.tools.ToolsLocator; |
14 |
import org.gvsig.tools.i18n.I18nManager; |
15 |
import org.gvsig.tools.task.SimpleTaskStatus; |
16 |
import org.xml.sax.InputSource; |
17 |
|
18 |
/**
|
19 |
*
|
20 |
* @author jjdelcerro
|
21 |
*/
|
22 |
@SuppressWarnings("UseSpecificCatch") |
23 |
public class XmlCommons { |
24 |
|
25 |
public static Charset detectCharset(InputStream is) { |
26 |
try {
|
27 |
AutoDetectReader reader = new AutoDetectReader(is);
|
28 |
return reader.getCharset();
|
29 |
} catch(Throwable t) { |
30 |
return null; |
31 |
} |
32 |
} |
33 |
|
34 |
public static String detectCharsetName(InputStream is) { |
35 |
Charset charset = detectCharset(is);
|
36 |
if( charset==null ) { |
37 |
return null; |
38 |
} |
39 |
return charset.name();
|
40 |
} |
41 |
|
42 |
public static InputSource openReader(File xmlfile, Charset charset) { |
43 |
try {
|
44 |
FileInputStream fis = new FileInputStream(xmlfile); |
45 |
|
46 |
InputSource is = new InputSource();
|
47 |
is.setPublicId(xmlfile.getAbsolutePath()); |
48 |
is.setByteStream(fis); |
49 |
if( charset!=null ) { |
50 |
is.setEncoding(charset.name()); |
51 |
} |
52 |
return openReader(is);
|
53 |
} catch(Throwable t) { |
54 |
throw new RuntimeException("Can't open xml input stream.",t); |
55 |
} |
56 |
} |
57 |
|
58 |
public static InputSource openReader(InputStream xml, Charset charset) { |
59 |
InputSource is = new InputSource();
|
60 |
is.setByteStream(xml); |
61 |
if( charset!=null ) { |
62 |
is.setEncoding(charset.name()); |
63 |
} |
64 |
return openReader(is);
|
65 |
} |
66 |
|
67 |
public static InputSource openReader(InputSource is) { |
68 |
try {
|
69 |
if(StringUtils.isBlank(is.getEncoding())){
|
70 |
// EncodingDetector encodingDetector = TikaConfig.getDefaultConfig().getEncodingDetector();
|
71 |
// BufferedInputStream bis = new BufferedInputStream(is.getByteStream());
|
72 |
// Charset charset = encodingDetector.detect(bis, new Metadata());
|
73 |
// is.setEncoding(charset.name());
|
74 |
// is.setByteStream(bis);
|
75 |
AutoDetectReader reader = new AutoDetectReader(is.getByteStream());
|
76 |
is.setCharacterStream(reader); |
77 |
is.setEncoding(reader.getCharset().name()); |
78 |
} else {
|
79 |
BOMInputStream bomIs = new BOMInputStream(is.getByteStream());
|
80 |
is.setByteStream(bomIs); |
81 |
InputStreamReader reader = new InputStreamReader( |
82 |
is.getByteStream(), |
83 |
is.getEncoding() |
84 |
); |
85 |
is.setCharacterStream(reader); |
86 |
} |
87 |
return is;
|
88 |
} catch(Throwable t) { |
89 |
throw new RuntimeException("Can't open xml input stream.",t); |
90 |
} |
91 |
} |
92 |
|
93 |
public static long countLines(File xml, Charset charset, SimpleTaskStatus status) { |
94 |
try {
|
95 |
FileInputStream fis = new FileInputStream(xml); |
96 |
return countLines(fis, charset, status);
|
97 |
} catch(Throwable t) { |
98 |
throw new RuntimeException("Can't count lines.",t); |
99 |
} |
100 |
} |
101 |
|
102 |
public static long countLines(InputStream xml, Charset charset, SimpleTaskStatus status) { |
103 |
try {
|
104 |
long count = 0; |
105 |
// Reader reader = null;
|
106 |
BufferedReader br = null; |
107 |
status.setIndeterminate(); |
108 |
status.setCurValue(0);
|
109 |
try {
|
110 |
InputSource is = new InputSource(xml);
|
111 |
if( charset!=null ) { |
112 |
is.setEncoding(charset.name()); |
113 |
} |
114 |
is = openReader(is); |
115 |
br = new BufferedReader(is.getCharacterStream()); |
116 |
int n = 1; |
117 |
I18nManager i18n = ToolsLocator.getI18nManager(); |
118 |
while( br.readLine()!=null ) { |
119 |
if(count%n == 0){ |
120 |
status.message(i18n.getTranslation("_Calculating_lines"));
|
121 |
status.setCurValue(count); |
122 |
} |
123 |
// status.incrementCurrentValue();
|
124 |
|
125 |
count++; |
126 |
if(count > 100000){ |
127 |
n = 10000;
|
128 |
} else if(count > 10000){ |
129 |
n = 1000;
|
130 |
} else if(count > 1000){ |
131 |
n = 100;
|
132 |
} else if(count > 100){ |
133 |
n = 10;
|
134 |
} |
135 |
} |
136 |
} finally {
|
137 |
IOUtils.closeQuietly(br); |
138 |
} |
139 |
return count;
|
140 |
} catch(Throwable t) { |
141 |
throw new RuntimeException("Can't count lines.",t); |
142 |
} |
143 |
} |
144 |
} |