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 / StructureExtractorImpl.java @ 47638
History | View | Annotate | Download (9.67 KB)
1 |
/*
|
---|---|
2 |
* To change this license header, choose License Headers in Project Properties.
|
3 |
* To change this template file, choose Tools | Templates
|
4 |
* and open the template in the editor.
|
5 |
*/
|
6 |
package org.gvsig.fmap.dal.store.gml.virtualrows; |
7 |
|
8 |
import java.io.File; |
9 |
import java.io.FileNotFoundException; |
10 |
import java.io.IOException; |
11 |
import java.io.InputStream; |
12 |
import java.io.Reader; |
13 |
import java.nio.charset.Charset; |
14 |
import java.util.ArrayList; |
15 |
import java.util.List; |
16 |
import java.util.Locale; |
17 |
import javax.xml.parsers.SAXParser; |
18 |
import javax.xml.parsers.SAXParserFactory; |
19 |
import org.apache.commons.lang3.StringUtils; |
20 |
import org.apache.tika.utils.CharsetUtils; |
21 |
import org.gvsig.fmap.dal.DALLocator; |
22 |
import org.gvsig.fmap.dal.DataManager; |
23 |
import org.gvsig.fmap.dal.store.gml.virtualrows.xmlinfo.XMLAttributeInfoImpl; |
24 |
import org.gvsig.fmap.dal.store.gml.virtualrows.xmlinfo.XMLInfoImpl; |
25 |
import org.gvsig.tools.ToolsLocator; |
26 |
import org.gvsig.tools.i18n.I18nManager; |
27 |
import org.gvsig.tools.task.SimpleTaskStatus; |
28 |
import org.slf4j.Logger; |
29 |
import org.slf4j.LoggerFactory; |
30 |
import org.xml.sax.Attributes; |
31 |
import org.xml.sax.InputSource; |
32 |
import org.xml.sax.Locator; |
33 |
import org.xml.sax.SAXException; |
34 |
import org.xml.sax.helpers.DefaultHandler; |
35 |
|
36 |
/**
|
37 |
*
|
38 |
* @author jjdelcerro
|
39 |
*/
|
40 |
public class StructureExtractorImpl { //implements StructureExtractor { |
41 |
|
42 |
private static final Logger LOGGER = LoggerFactory.getLogger(StructureExtractorImpl.class); |
43 |
|
44 |
@SuppressWarnings("UseSpecificCatch") |
45 |
private void extractTags(XMLInfoImpl xmlinfo, Reader reader, SimpleTaskStatus status) { |
46 |
if( reader == null ) { |
47 |
throw new IllegalArgumentException("reader is null"); |
48 |
} |
49 |
try {
|
50 |
final DataManager dataManager = DALLocator.getDataManager();
|
51 |
I18nManager i18n = ToolsLocator.getI18nManager(); |
52 |
status.message(i18n.getTranslation("_Reading_xml")+" 1/5"); |
53 |
status.setRangeOfValues(0, xmlinfo.getCountLines());
|
54 |
|
55 |
SAXParserFactory spf = SAXParserFactory.newInstance(); |
56 |
spf.setNamespaceAware(true);
|
57 |
SAXParser saxParser = spf.newSAXParser();
|
58 |
InputSource is = new InputSource(reader);
|
59 |
|
60 |
List<String> path = new ArrayList<>(); |
61 |
|
62 |
saxParser.parse(is, new DefaultHandler() {
|
63 |
private Locator locator;
|
64 |
int size;
|
65 |
int refreshInterval = 1; |
66 |
StringBuilder chars = new StringBuilder(); |
67 |
|
68 |
@Override
|
69 |
public void setDocumentLocator(Locator locator) { |
70 |
this.locator = locator;
|
71 |
} |
72 |
|
73 |
@Override
|
74 |
public void startElement(String uri, String localName, String qName, Attributes attributes) throws SAXException { |
75 |
int line = this.locator.getLineNumber(); |
76 |
int column = this.locator.getColumnNumber()-2-localName.length(); |
77 |
|
78 |
if(line % refreshInterval == 0) { |
79 |
status.setCurValue(line); |
80 |
} |
81 |
|
82 |
if(line > 100000){ |
83 |
refreshInterval = 10000;
|
84 |
} else if(line > 10000){ |
85 |
refreshInterval = 1000;
|
86 |
} else if(line > 1000){ |
87 |
refreshInterval = 100;
|
88 |
} else if(line > 100){ |
89 |
refreshInterval = 10;
|
90 |
} |
91 |
|
92 |
|
93 |
String idvalue = dataManager.createUniqueID();
|
94 |
|
95 |
path.add(localName); |
96 |
String path_s = StringUtils.join(path, "/"); |
97 |
XMLAttributeInfoImpl info = xmlinfo.getTag(path_s); |
98 |
if( info == null ) { |
99 |
info = new XMLAttributeInfoImpl(xmlinfo.getLocale(), path_s, getNs(qName));
|
100 |
xmlinfo.addTag(info); |
101 |
|
102 |
} |
103 |
if( path.size()>1 ) { |
104 |
List<String> parentpath = path.subList(0, path.size()-1); |
105 |
String parentpath_s = StringUtils.join(parentpath, "/"); |
106 |
XMLAttributeInfoImpl parentinfo = xmlinfo.getTag(parentpath_s); |
107 |
parentinfo.incrChildCount(localName); |
108 |
parentinfo.setLastChildID(localName, idvalue); |
109 |
} |
110 |
|
111 |
for (int i = 0; i < attributes.getLength(); i++) { |
112 |
String name = attributes.getLocalName(i);
|
113 |
String value = attributes.getValue(i);
|
114 |
String idvalueChild = dataManager.createUniqueID();
|
115 |
XMLAttributeInfoImpl infoChild = xmlinfo.getTag(path_s+"/"+name);
|
116 |
if( infoChild == null ) { |
117 |
infoChild = new XMLAttributeInfoImpl(xmlinfo.getLocale(), path_s+"/"+name, getNs(qName)); |
118 |
infoChild.setIsAttr(true);
|
119 |
xmlinfo.addTag(infoChild); |
120 |
} |
121 |
info.incrChildCount(infoChild.getName()); |
122 |
info.setLastChildID(infoChild.getName(), idvalueChild); |
123 |
infoChild.addValue(value); |
124 |
} |
125 |
chars.setLength(0);
|
126 |
} |
127 |
|
128 |
@Override
|
129 |
public void endElement(String uri, String localName, String qName) throws SAXException { |
130 |
int line = this.locator.getLineNumber(); |
131 |
|
132 |
// status.setCurValue(line);
|
133 |
|
134 |
String path_s = StringUtils.join(path, "/"); |
135 |
XMLAttributeInfoImpl info = xmlinfo.getTag(path_s); |
136 |
if( info == null ) { |
137 |
info = new XMLAttributeInfoImpl(xmlinfo.getLocale(), path_s, getNs(qName));
|
138 |
xmlinfo.addTag(info); |
139 |
} |
140 |
|
141 |
XMLAttributeInfoImpl parentinfo = null;
|
142 |
if( path.size()>1 ) { |
143 |
List<String> parentpath = path.subList(0, path.size()-1); |
144 |
String parentpath_s = StringUtils.join(parentpath, "/"); |
145 |
parentinfo = xmlinfo.getTag(parentpath_s); |
146 |
} |
147 |
|
148 |
if( info.hasChilds() || (parentinfo != null && (parentinfo.getChildsCount(localName)>1))) { |
149 |
String value = this.chars.toString(); |
150 |
if( StringUtils.isNotBlank(value) ) {
|
151 |
String name = info.getName()+"$v"; |
152 |
String idvalueChild = dataManager.createUniqueID();
|
153 |
XMLAttributeInfoImpl infoChild = xmlinfo.getTag(path_s+"/"+name);
|
154 |
if( infoChild == null ) { |
155 |
infoChild = new XMLAttributeInfoImpl(xmlinfo.getLocale(), path_s+"/"+name, getNs(qName)); |
156 |
xmlinfo.addTag(infoChild); |
157 |
} |
158 |
info.incrChildCount(infoChild.getName()); |
159 |
info.setLastChildID(infoChild.getName(), idvalueChild); |
160 |
infoChild.addValue(value); |
161 |
} |
162 |
} else {
|
163 |
String value = this.chars.toString(); |
164 |
info.addValue(value); |
165 |
} |
166 |
info.consolidateChildCounters(); |
167 |
path.remove(path.size()-1);
|
168 |
chars.setLength(0);
|
169 |
} |
170 |
|
171 |
@Override
|
172 |
public void characters(char[] ch, int start, int length) throws SAXException { |
173 |
int line = this.locator.getLineNumber(); |
174 |
|
175 |
// status.setCurValue(line);
|
176 |
this.chars.append(ch, start, length);
|
177 |
} |
178 |
|
179 |
|
180 |
}); |
181 |
} catch (Exception ex) { |
182 |
throw new RuntimeException("Can't extract tags.", ex); |
183 |
} |
184 |
} |
185 |
|
186 |
public XMLInfoImpl extractStructure(File xml, Charset charset, Locale locale, SimpleTaskStatus status) throws FileNotFoundException, IOException { |
187 |
XMLInfoImpl xmlinfo = new XMLInfoImpl();
|
188 |
xmlinfo.setLocale(locale); |
189 |
long count = XmlCommons.countLines(xml, charset, status);
|
190 |
xmlinfo.setCountLines(count); |
191 |
InputSource is = XmlCommons.openReader(xml,charset); |
192 |
return extractStructure(is, xmlinfo, status);
|
193 |
} |
194 |
|
195 |
public XMLInfoImpl extractStructure(InputStream xml, Charset charset, Locale locale, SimpleTaskStatus status) throws IOException { |
196 |
XMLInfoImpl xmlinfo = new XMLInfoImpl();
|
197 |
xmlinfo.setLocale(locale); |
198 |
xmlinfo.setCountLines(-1);
|
199 |
InputSource is = XmlCommons.openReader(xml, charset); |
200 |
return extractStructure(is, xmlinfo, status);
|
201 |
} |
202 |
|
203 |
public XMLInfoImpl extractStructure(Reader reader, Locale locale, SimpleTaskStatus status) { |
204 |
XMLInfoImpl xmlinfo = new XMLInfoImpl();
|
205 |
xmlinfo.setLocale(locale); |
206 |
InputSource is = new InputSource(reader);
|
207 |
return extractStructure(is, xmlinfo, status);
|
208 |
} |
209 |
|
210 |
public XMLInfoImpl extractStructure(InputSource is, XMLInfoImpl xmlinfo, SimpleTaskStatus status) {
|
211 |
|
212 |
if( xmlinfo.getCharset()==null ) { |
213 |
xmlinfo.setCharset(CharsetUtils.forName(is.getEncoding())); |
214 |
} |
215 |
|
216 |
extractTags(xmlinfo, is.getCharacterStream(), status); |
217 |
|
218 |
return xmlinfo;
|
219 |
} |
220 |
|
221 |
private String getNs(String s){ |
222 |
String[] ss = s.split(":"); |
223 |
if(ss.length == 1){ |
224 |
return null; |
225 |
} |
226 |
return ss[0]; |
227 |
} |
228 |
|
229 |
} |