1 | /*
|
---|
2 | * GATEServices.java
|
---|
3 | * Copyright (C) 2002 New Zealand Digital Library, http://www.nzdl.org
|
---|
4 | *
|
---|
5 | * This program is free software; you can redistribute it and/or modify
|
---|
6 | * it under the terms of the GNU General Public License as published by
|
---|
7 | * the Free Software Foundation; either version 2 of the License, or
|
---|
8 | * (at your option) any later version.
|
---|
9 | *
|
---|
10 | * This program is distributed in the hope that it will be useful,
|
---|
11 | * but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
12 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
---|
13 | * GNU General Public License for more details.
|
---|
14 | *
|
---|
15 | * You should have received a copy of the GNU General Public License
|
---|
16 | * along with this program; if not, write to the Free Software
|
---|
17 | * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
---|
18 | */
|
---|
19 | package org.greenstone.gsdl3.service;
|
---|
20 |
|
---|
21 |
|
---|
22 | // Greenstone classes
|
---|
23 | import org.greenstone.gsdl3.util.*;
|
---|
24 |
|
---|
25 | // GATE classes
|
---|
26 | import gate.*;
|
---|
27 | import gate.creole.*;
|
---|
28 | import gate.gui.*;
|
---|
29 | import gate.util.persistence.PersistenceManager;
|
---|
30 |
|
---|
31 | // XML classes
|
---|
32 | import org.w3c.dom.Element;
|
---|
33 | import org.w3c.dom.Node;
|
---|
34 | import org.w3c.dom.NodeList;
|
---|
35 |
|
---|
36 | // General Java classes
|
---|
37 | import java.io.File;
|
---|
38 | import java.net.URL;
|
---|
39 | import java.util.Collection;
|
---|
40 | import java.util.HashSet;
|
---|
41 | import java.util.Iterator;
|
---|
42 | import java.util.Set;
|
---|
43 |
|
---|
44 |
|
---|
45 | /**
|
---|
46 | * <p>Title: GATE Greenstone3 integration</p>
|
---|
47 | * <p>Description: </p>
|
---|
48 | * <p>Copyright: Copyright (c) 2003</p>
|
---|
49 | * <p>Company: University of Waikato</p>
|
---|
50 | * @author unascribed
|
---|
51 | * @version 1.0
|
---|
52 | */
|
---|
53 |
|
---|
54 | public class GATEServices
|
---|
55 | extends ServiceRack {
|
---|
56 |
|
---|
57 | // the services on offer
|
---|
58 | // these strings must match what is found in the properties file
|
---|
59 | protected static final String GATE_POS_TAG_SERVICE = "GatePOSTag";
|
---|
60 |
|
---|
61 | protected static final String ANNOTATION_TYPE_PARAM = "annotationType";
|
---|
62 |
|
---|
63 | protected Element config_info = null;
|
---|
64 |
|
---|
65 | protected CorpusController application;
|
---|
66 |
|
---|
67 | protected Corpus corpus;
|
---|
68 |
|
---|
69 | protected String[] annotation_types = { "Date", "Location",
|
---|
70 | "Organization", "Person" };
|
---|
71 | // Address,
|
---|
72 |
|
---|
73 | /** constructor */
|
---|
74 | public GATEServices()
|
---|
75 | {
|
---|
76 | }
|
---|
77 |
|
---|
78 |
|
---|
79 | /** configure this service */
|
---|
80 | public boolean configure(Element info, Element extra_info)
|
---|
81 | {
|
---|
82 | System.out.println("Configuring GATEServices...");
|
---|
83 | this.config_info = info;
|
---|
84 |
|
---|
85 | // set up short_service_info_ - for now just has name and type
|
---|
86 | Element tag_service = this.doc.createElement(GSXML.SERVICE_ELEM);
|
---|
87 | tag_service.setAttribute(GSXML.TYPE_ATT, GSXML.SERVICE_TYPE_ENRICH);
|
---|
88 | tag_service.setAttribute(GSXML.NAME_ATT, GATE_POS_TAG_SERVICE);
|
---|
89 | this.short_service_info.appendChild(tag_service);
|
---|
90 |
|
---|
91 |
|
---|
92 | // Configure GATE for use
|
---|
93 | try {
|
---|
94 | Gate.init();
|
---|
95 | Gate.getUserConfig().put(GateConstants.DOCUMENT_ADD_SPACE_ON_UNPACK_FEATURE_NAME,
|
---|
96 | new Boolean(false));
|
---|
97 |
|
---|
98 | // MainFrame mainFrame = new MainFrame();
|
---|
99 | // mainFrame.setSize(new java.awt.Dimension(800, 600));
|
---|
100 | // mainFrame.setVisible(true);
|
---|
101 |
|
---|
102 | // Load the (pre-created) application
|
---|
103 | URL applicationFileURL = ClassLoader.getSystemResource("gate.app");
|
---|
104 | File applicationFile = new File(applicationFileURL.getFile());
|
---|
105 | this.application = (CorpusController) PersistenceManager.loadObjectFromFile(applicationFile);
|
---|
106 |
|
---|
107 | /* Collection processing_resources = application.getPRs();
|
---|
108 | Iterator pr_iterator = processing_resources.iterator();
|
---|
109 | while (pr_iterator.hasNext()) {
|
---|
110 | ProcessingResource pr = (ProcessingResource) pr_iterator.next();
|
---|
111 | String pr_name = pr.getName();
|
---|
112 | System.out.println("PR name: " + pr_name);
|
---|
113 | if (pr_name.startsWith("ANNIE POS Tagger")) {
|
---|
114 | String as_name = ((POSTagger) pr).getInputASName();
|
---|
115 | System.out.println("AS name: " + as_name);
|
---|
116 | }
|
---|
117 | } */
|
---|
118 |
|
---|
119 | // Create a new corpus
|
---|
120 | this.corpus = Factory.newCorpus("GSDL3 Corpus");
|
---|
121 | this.application.setCorpus(this.corpus);
|
---|
122 | }
|
---|
123 | catch (Exception e) {
|
---|
124 | e.printStackTrace();
|
---|
125 | }
|
---|
126 |
|
---|
127 | return true;
|
---|
128 | }
|
---|
129 |
|
---|
130 | protected Element getServiceDescription(String service, String lang, String subset) {
|
---|
131 |
|
---|
132 | if (!service.equals(GATE_POS_TAG_SERVICE)) {
|
---|
133 | return null;
|
---|
134 | }
|
---|
135 | Element tag_service = this.doc.createElement(GSXML.SERVICE_ELEM);
|
---|
136 | tag_service.setAttribute(GSXML.TYPE_ATT, GSXML.SERVICE_TYPE_ENRICH);
|
---|
137 | tag_service.setAttribute(GSXML.NAME_ATT, GATE_POS_TAG_SERVICE);
|
---|
138 | if (subset==null || subset.equals(GSXML.DISPLAY_TEXT_ELEM)) {
|
---|
139 | tag_service.appendChild(GSXML.createDisplayTextElement(this.doc, GSXML.DISPLAY_TEXT_NAME, getTextString(service+".name", lang)));
|
---|
140 | tag_service.appendChild(GSXML.createDisplayTextElement(this.doc, GSXML.DISPLAY_TEXT_DESCRIPTION, getTextString(service+".description", lang)));
|
---|
141 | tag_service.appendChild(GSXML.createDisplayTextElement(this.doc, GSXML.DISPLAY_TEXT_SUBMIT, getTextString(service+".submit", lang)));
|
---|
142 | }
|
---|
143 | if (subset==null || subset.equals(GSXML.PARAM_ELEM+GSXML.LIST_MODIFIER)) {
|
---|
144 | Element param_list = this.doc.createElement(GSXML.PARAM_ELEM+GSXML.LIST_MODIFIER);
|
---|
145 | createParameter(ANNOTATION_TYPE_PARAM, param_list, lang);
|
---|
146 | tag_service.appendChild(param_list);
|
---|
147 | }
|
---|
148 | return tag_service;
|
---|
149 | }
|
---|
150 |
|
---|
151 |
|
---|
152 | /** creates a new param element and adds it to the param list */
|
---|
153 | protected void createParameter(String name, Element param_list,
|
---|
154 | String lang)
|
---|
155 | {
|
---|
156 | Element param = null;
|
---|
157 |
|
---|
158 | if (name.equals(ANNOTATION_TYPE_PARAM)) {
|
---|
159 | int len = this.annotation_types.length;
|
---|
160 | String[] annotation_type_names = new String[len];
|
---|
161 | for (int i = 0; i < len; i++) {
|
---|
162 | annotation_type_names[i] = getTextString("param." + name + "." + this.annotation_types[i], lang);
|
---|
163 | }
|
---|
164 |
|
---|
165 | param = GSXML.createParameterDescription(this.doc, name, getTextString("param." + name, lang), GSXML.PARAM_TYPE_ENUM_MULTI, this.annotation_types[0], this.annotation_types, annotation_type_names);
|
---|
166 | param_list.appendChild(param);
|
---|
167 | }
|
---|
168 |
|
---|
169 | }
|
---|
170 |
|
---|
171 |
|
---|
172 | protected Element processGatePOSTag(Element request)
|
---|
173 | {
|
---|
174 | // System.out.println("(GatePOSTag) Request:\n" + converter_.getPrettyString(request));
|
---|
175 |
|
---|
176 | // Create a new (empty) result message
|
---|
177 | Element result = this.doc.createElement(GSXML.RESPONSE_ELEM);
|
---|
178 | result.setAttribute(GSXML.FROM_ATT, GATE_POS_TAG_SERVICE);
|
---|
179 | result.setAttribute(GSXML.TYPE_ATT, GSXML.REQUEST_TYPE_PROCESS);
|
---|
180 |
|
---|
181 | // Get the parameters of the request
|
---|
182 | Element param_list = (Element) GSXML.getChildByTagName(request, GSXML.PARAM_ELEM+GSXML.LIST_MODIFIER);
|
---|
183 | if (param_list == null) {
|
---|
184 | System.err.println("Error: GatePOSTag request had no paramList.");
|
---|
185 | return result; // Return the empty result
|
---|
186 | }
|
---|
187 |
|
---|
188 | // Process the request parameters
|
---|
189 | Set annTypes = new HashSet();
|
---|
190 | Element param = (Element) param_list.getFirstChild();
|
---|
191 | while (param != null) {
|
---|
192 | // Identify the annotation types desired
|
---|
193 | if (param.getAttribute(GSXML.NAME_ATT).equals("annotationType")) {
|
---|
194 | String annotation_type = GSXML.getValue(param);
|
---|
195 | String [] types = annotation_type.split(",");
|
---|
196 | for (int i=0; i<types.length; i++) {
|
---|
197 | annTypes.add(types[i]);
|
---|
198 | }
|
---|
199 | }
|
---|
200 |
|
---|
201 | param = (Element) param.getNextSibling();
|
---|
202 | }
|
---|
203 |
|
---|
204 | // Get the request content
|
---|
205 | Element doc_node_list = (Element) GSXML.getChildByTagName(request, GSXML.DOC_NODE_ELEM+GSXML.LIST_MODIFIER);
|
---|
206 | if (doc_node_list == null) {
|
---|
207 | System.err.println("Error: GatePOSTag request specified no doc nodes.");
|
---|
208 | return result; // Return the empty result
|
---|
209 | }
|
---|
210 |
|
---|
211 | // Process each document node in the list
|
---|
212 | NodeList doc_nodes = doc_node_list.getChildNodes();
|
---|
213 | for (int i = 0; i < doc_nodes.getLength(); i++) {
|
---|
214 | Element doc_node = (Element) doc_nodes.item(i);
|
---|
215 | Element content = (Element) GSXML.getChildByTagName(doc_node, "nodeContent");
|
---|
216 | Node content_text = (Node) GSXML.getNodeTextNode(content);
|
---|
217 | String text = content_text.getNodeValue();
|
---|
218 |
|
---|
219 | // GATE needs the text to be a valid HTML file
|
---|
220 | text = "<html><head></head><body>" + text + "</body></html>";
|
---|
221 | String annotated_text = processText(text, annTypes);
|
---|
222 |
|
---|
223 | // Remove the surrounding HTML tags
|
---|
224 | annotated_text = annotated_text.substring(49, annotated_text.length() - 13);
|
---|
225 | // System.out.println("GATE result:\n" + annotated_text);
|
---|
226 |
|
---|
227 | annotated_text = "<nodeContent>" + annotated_text + "</nodeContent>";
|
---|
228 | Element annotated_content = this.converter.getDOM(annotated_text).getDocumentElement();
|
---|
229 | doc_node.replaceChild(doc_node.getOwnerDocument().importNode(annotated_content, true), content);
|
---|
230 | }
|
---|
231 |
|
---|
232 | result.appendChild(this.doc.importNode(doc_node_list, true));
|
---|
233 | // System.out.println("GatePOSTag result:\n" + converter_.getPrettyString(result));
|
---|
234 | return result;
|
---|
235 | }
|
---|
236 |
|
---|
237 |
|
---|
238 | public String processText(String text, Set annotationTypesToExport)
|
---|
239 | {
|
---|
240 | try {
|
---|
241 | // Create a new document containing the text
|
---|
242 | FeatureMap parameters = Factory.newFeatureMap();
|
---|
243 | parameters.put("stringContent", text);
|
---|
244 | parameters.put("markupAware", new Boolean(true));
|
---|
245 | parameters.put("preserveOriginalContent", new Boolean(true));
|
---|
246 | parameters.put("collectRepositioningInfo", new Boolean(true));
|
---|
247 | Document doc = (Document) Factory.createResource("gate.corpora.DocumentImpl",
|
---|
248 | parameters);
|
---|
249 |
|
---|
250 | // Add it to the corpus
|
---|
251 | this.corpus.clear();
|
---|
252 | this.corpus.add(doc);
|
---|
253 |
|
---|
254 | // Process the corpus
|
---|
255 | this.application.execute();
|
---|
256 |
|
---|
257 | // Extract all the annotations
|
---|
258 | AnnotationSet annSet = doc.getAnnotations();
|
---|
259 |
|
---|
260 | // Return the desired annotations
|
---|
261 | AnnotationSet outputAnnotations = annSet.get(annotationTypesToExport);
|
---|
262 | String result = doc.toXml(outputAnnotations, false);
|
---|
263 | Factory.deleteResource(doc);
|
---|
264 |
|
---|
265 | // 1. Escape the GATE result
|
---|
266 | result = GSXML.xmlSafe(result);
|
---|
267 |
|
---|
268 | // 2. Unescape the annotation tags
|
---|
269 | Iterator setIterator = annotationTypesToExport.iterator();
|
---|
270 | while (setIterator.hasNext()) {
|
---|
271 | String annotationType = (String) setIterator.next();
|
---|
272 | result = result.replaceAll("<" + annotationType + ">",
|
---|
273 | "<annotation type=\"" + annotationType + "\">");
|
---|
274 | result = result.replaceAll("</" + annotationType + ">",
|
---|
275 | "</annotation>");
|
---|
276 | }
|
---|
277 |
|
---|
278 | return result;
|
---|
279 | }
|
---|
280 | catch (Exception e) {
|
---|
281 | e.printStackTrace();
|
---|
282 | return null;
|
---|
283 | }
|
---|
284 | }
|
---|
285 | }
|
---|