1 | /**********************************************************************
|
---|
2 | *
|
---|
3 | * ApplyXSLT.java
|
---|
4 | *
|
---|
5 | * Copyright 2006-2010 The New Zealand Digital Library Project
|
---|
6 | *
|
---|
7 | * A component of the Greenstone digital library software
|
---|
8 | * from the New Zealand Digital Library Project at the
|
---|
9 | * University of Waikato, New Zealand.
|
---|
10 | *
|
---|
11 | * This program is free software; you can redistribute it and/or modify
|
---|
12 | * it under the terms of the GNU General Public License as published by
|
---|
13 | * the Free Software Foundation; either version 2 of the License, or
|
---|
14 | * (at your option) any later version.
|
---|
15 | *
|
---|
16 | * This program is distributed in the hope that it will be useful,
|
---|
17 | * but WITHOUT ANY WARRANTY; without even the implied warranty of
|
---|
18 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
---|
19 | * GNU General Public License for more details.
|
---|
20 | *
|
---|
21 | * You should have received a copy of the GNU General Public License
|
---|
22 | * along with this program; if not, write to the Free Software
|
---|
23 | * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
|
---|
24 | *
|
---|
25 | *********************************************************************/
|
---|
26 |
|
---|
27 |
|
---|
28 | package org.nzdl.gsdl;
|
---|
29 |
|
---|
30 | import java.io.*;
|
---|
31 | import java.util.HashMap;
|
---|
32 | import java.util.Iterator;
|
---|
33 | import java.util.Map;
|
---|
34 | import java.util.Map.Entry;
|
---|
35 |
|
---|
36 | import javax.xml.transform.Transformer;
|
---|
37 | import javax.xml.transform.TransformerConfigurationException;
|
---|
38 | import javax.xml.transform.TransformerException;
|
---|
39 | import javax.xml.transform.TransformerFactory;
|
---|
40 | import javax.xml.transform.stream.StreamResult;
|
---|
41 | import javax.xml.transform.stream.StreamSource;
|
---|
42 |
|
---|
43 | import javax.xml.parsers.*;
|
---|
44 | import javax.xml.transform.dom.*;
|
---|
45 | import org.w3c.dom.*;
|
---|
46 |
|
---|
47 |
|
---|
48 |
|
---|
49 | /**
|
---|
50 | * Use the TraX interface to perform a transformation in the simplest manner possible
|
---|
51 | * (3 statements).
|
---|
52 | */
|
---|
53 | public class ApplyXSLT
|
---|
54 | {
|
---|
55 |
|
---|
56 | public static final String DOC_START = new String ("<?DocStart?>");
|
---|
57 | public static final String DOC_END = new String ("<?DocEnd?>");
|
---|
58 | public static final String INPUT_END = new String ("<?Done?>");
|
---|
59 |
|
---|
60 | private static final String RECORD_ELEMENT = "record";
|
---|
61 | private static final String CONTROLFIELD_ELEMENT = "controlfield";
|
---|
62 | private static final String SUBFIELD_ELEMENT = "subfield";
|
---|
63 | private static final String LEADER_ELEMENT = "leader";
|
---|
64 |
|
---|
65 | private final int BEFORE_READING = 0;
|
---|
66 | private final int IS_READING = 1;
|
---|
67 | private String xsl_file;
|
---|
68 | private String mapping_file;
|
---|
69 |
|
---|
70 | private String sourcelang;
|
---|
71 | private String targetlang;
|
---|
72 | private HashMap paramMap;
|
---|
73 |
|
---|
74 | public ApplyXSLT(String sourcelang, String targetlang, HashMap param_map){
|
---|
75 | initParams(sourcelang, targetlang, param_map);
|
---|
76 | }
|
---|
77 |
|
---|
78 | public ApplyXSLT(String xsl_file, String sourcelang, String targetlang, HashMap param_map)
|
---|
79 | {
|
---|
80 | this.xsl_file = xsl_file;
|
---|
81 | initParams(sourcelang, targetlang, param_map);
|
---|
82 | }
|
---|
83 |
|
---|
84 | public ApplyXSLT(String xsl_file, String mapping_file, String sourcelang, String targetlang, HashMap param_map) {
|
---|
85 | this.xsl_file = xsl_file;
|
---|
86 | this.mapping_file = mapping_file;
|
---|
87 | initParams(sourcelang, targetlang, param_map);
|
---|
88 | }
|
---|
89 |
|
---|
90 | private void initParams(String sourcelang, String targetlang, HashMap param_map)
|
---|
91 | {
|
---|
92 | this.sourcelang = sourcelang;
|
---|
93 | this.targetlang = targetlang;
|
---|
94 | // if only target language is provided, assume source language is English
|
---|
95 | if(sourcelang.equals("") && !targetlang.equals("")) {
|
---|
96 | this.sourcelang = "en";
|
---|
97 | }
|
---|
98 |
|
---|
99 | // any custom parameters to be passed into the XSLT would be in the map by now
|
---|
100 | paramMap = param_map;
|
---|
101 | }
|
---|
102 |
|
---|
103 | /*
|
---|
104 | private boolean isOpen(BufferedReader br) {
|
---|
105 | // The code used to rely on br.ready() to determine whether it should
|
---|
106 | // continue looping to read input (from standard-in). However on
|
---|
107 | // closer read of the JavaDoc, br.ready() can return false even
|
---|
108 | // when the input stream is still open -- it's just that there isn't
|
---|
109 | // any input lines to read in at the moment.
|
---|
110 |
|
---|
111 | // If the input stream has been closed (from the external calling Perl code),
|
---|
112 | // then the only sure way to determine that this condition has arise is
|
---|
113 | // to ask br.ready() and catch the Exception that occurs ... because the
|
---|
114 | // input stream is not closed!
|
---|
115 |
|
---|
116 | boolean is_open = true;
|
---|
117 |
|
---|
118 | try {
|
---|
119 | boolean is_ready = br.ready();
|
---|
120 | }
|
---|
121 | catch (Exception e) { // Could be more selective and go for IOException ??
|
---|
122 | //System.err.println("ApplyXSLT::isOpen() encountered an exception => so input is closed");
|
---|
123 | is_open = false;
|
---|
124 | }
|
---|
125 |
|
---|
126 | return is_open;
|
---|
127 | }
|
---|
128 | */
|
---|
129 |
|
---|
130 | private boolean process()
|
---|
131 | {
|
---|
132 | try{
|
---|
133 |
|
---|
134 | // Use System InputStream to receive piped data from the perl program
|
---|
135 | InputStreamReader ir = new InputStreamReader(System.in, "UTF8");
|
---|
136 | BufferedReader br = new BufferedReader(ir);
|
---|
137 |
|
---|
138 | int system_status = BEFORE_READING;
|
---|
139 | StringBuffer a_doc = new StringBuffer();
|
---|
140 | String output_file = new String();
|
---|
141 |
|
---|
142 |
|
---|
143 | String this_line;
|
---|
144 | while ((this_line = br.readLine()) != null) {
|
---|
145 |
|
---|
146 | //System.err.println("Read in from pipe, line: " + this_line);
|
---|
147 |
|
---|
148 | if(system_status == BEFORE_READING){
|
---|
149 | if(this_line.compareTo(DOC_START) == 0){
|
---|
150 | // If this_line is DOC_START then we require the next line of input to be
|
---|
151 | // the filename
|
---|
152 | output_file = br.readLine(); // read the next line as the output file name
|
---|
153 | if (output_file == null) {
|
---|
154 | // A problem of some form occurred
|
---|
155 | return false;
|
---|
156 | }
|
---|
157 | //System.err.println("Read in from pipe, next line: " + output_file);
|
---|
158 | system_status = IS_READING;
|
---|
159 | a_doc = new StringBuffer();
|
---|
160 | }
|
---|
161 | else if(this_line.compareTo(INPUT_END) == 0){
|
---|
162 | return true;
|
---|
163 | }
|
---|
164 | else{
|
---|
165 | system_status = BEFORE_READING;
|
---|
166 | }
|
---|
167 |
|
---|
168 | }
|
---|
169 | else if(system_status == IS_READING){
|
---|
170 | if(this_line.compareTo(DOC_END) == 0){
|
---|
171 | boolean result = false;
|
---|
172 | if (mapping_file !=null && !mapping_file.equals("")){
|
---|
173 | result = translateXMLWithMapping(a_doc.toString(), output_file);
|
---|
174 | }
|
---|
175 | else{
|
---|
176 | result = translateXML(a_doc.toString(), output_file);
|
---|
177 | }
|
---|
178 |
|
---|
179 | if (!result){
|
---|
180 | return false;
|
---|
181 | }
|
---|
182 |
|
---|
183 | system_status = BEFORE_READING;
|
---|
184 |
|
---|
185 | }
|
---|
186 | else{
|
---|
187 | a_doc.append(this_line + "\n");
|
---|
188 | }
|
---|
189 | }
|
---|
190 | else{
|
---|
191 | System.err.println ("Undefined system status in ApplyXSLT.java main().");
|
---|
192 | System.exit(-1);
|
---|
193 | }
|
---|
194 |
|
---|
195 | }
|
---|
196 |
|
---|
197 | return true;
|
---|
198 |
|
---|
199 | //if(br != null) {
|
---|
200 | // br.close();
|
---|
201 | // br = null;
|
---|
202 | //}
|
---|
203 | }
|
---|
204 | catch (Exception e) {
|
---|
205 | System.err.println("Receiving piped data error!" + e.toString());
|
---|
206 | e.printStackTrace();
|
---|
207 | }
|
---|
208 |
|
---|
209 | return false;
|
---|
210 | }
|
---|
211 |
|
---|
212 | // reads xml from stdin, but no <?DocStart?><?DocEnd?> markers, and sends output to STDOUT
|
---|
213 | private boolean processPipedFromStdIn()
|
---|
214 | {
|
---|
215 | try{
|
---|
216 | //System.err.println("Received nothing\n");
|
---|
217 |
|
---|
218 | ReadStreamGobbler readInStreamGobbler = new ReadStreamGobbler(System.in, true);
|
---|
219 | readInStreamGobbler.start();
|
---|
220 | readInStreamGobbler.join();
|
---|
221 | String outputstr = readInStreamGobbler.getOutput();
|
---|
222 |
|
---|
223 | // Using join() above, even though we use only one streamGobbler thread, and even
|
---|
224 | // though we're not dealing with the input/output/error streams of a Process object.
|
---|
225 | // But the join() call here didn't break things.
|
---|
226 | // http://www.javaworld.com/article/2071275/core-java/when-runtime-exec---won-t.html?page=2
|
---|
227 |
|
---|
228 | boolean result = false;
|
---|
229 | if (mapping_file !=null && !mapping_file.equals("")){
|
---|
230 | result = translateXMLWithMapping(outputstr, null); // null: no outputfile, send to STDOUT
|
---|
231 | }
|
---|
232 | else{
|
---|
233 | result = translateXML(outputstr, null); // null: no outputfile, send to STDOUT
|
---|
234 | }
|
---|
235 |
|
---|
236 | if (!result){
|
---|
237 | System.err.println("Translation Failed!!");
|
---|
238 | return false;
|
---|
239 | } else {
|
---|
240 | return true;
|
---|
241 | }
|
---|
242 | }catch (Exception e) {
|
---|
243 | System.err.println("Receiving piped data error!" + e.toString());
|
---|
244 | e.printStackTrace();
|
---|
245 | }
|
---|
246 | return false;
|
---|
247 | }
|
---|
248 |
|
---|
249 |
|
---|
250 | private boolean translateXML(String full_doc, String output_file)
|
---|
251 | throws IOException,TransformerException, TransformerConfigurationException, FileNotFoundException
|
---|
252 | {
|
---|
253 |
|
---|
254 | StringReader str = new StringReader(full_doc) ;
|
---|
255 |
|
---|
256 | TransformerFactory tFactory = TransformerFactory.newInstance();
|
---|
257 | Transformer transformer = tFactory.newTransformer(new StreamSource(xsl_file));
|
---|
258 |
|
---|
259 | setTransformerParams(transformer); // sourcelang and targetlang and any further custom parameters to be passed into the XSLT
|
---|
260 |
|
---|
261 | if(output_file != null) {
|
---|
262 | transformer.transform(new StreamSource(str), new StreamResult(new FileOutputStream(output_file)));
|
---|
263 | } else {
|
---|
264 | transformer.transform(new StreamSource(str), new StreamResult(System.out));
|
---|
265 | }
|
---|
266 | return true;
|
---|
267 | }
|
---|
268 |
|
---|
269 | private boolean translateXMLWithMapping(String full_doc, String output_file)
|
---|
270 | throws IOException,TransformerException, TransformerConfigurationException, FileNotFoundException
|
---|
271 | {
|
---|
272 | StringReader str = new StringReader(full_doc) ;
|
---|
273 |
|
---|
274 | try{
|
---|
275 | TransformerFactory tFactory = TransformerFactory.newInstance();
|
---|
276 | Transformer transformer = tFactory.newTransformer(new StreamSource(xsl_file));
|
---|
277 |
|
---|
278 | Document mapping_doc = DocumentBuilderFactory.newInstance().newDocumentBuilder().parse(mapping_file);
|
---|
279 | Element mapping =mapping_doc.getDocumentElement();
|
---|
280 |
|
---|
281 | transformer.setParameter("mapping",mapping);
|
---|
282 | setTransformerParams(transformer); // sourcelang and targetlang and any further custom parameters to be passed into the XSLT
|
---|
283 |
|
---|
284 | Document output_doc = DocumentBuilderFactory.newInstance().newDocumentBuilder().newDocument();
|
---|
285 |
|
---|
286 | transformer.transform(new StreamSource(str), new DOMResult(output_doc));
|
---|
287 |
|
---|
288 | calculateRecordsLength(output_doc);
|
---|
289 |
|
---|
290 | transformer = tFactory.newTransformer();
|
---|
291 |
|
---|
292 | transformer.transform(new DOMSource(output_doc), new StreamResult(new FileOutputStream(output_file)));
|
---|
293 |
|
---|
294 | }
|
---|
295 | catch(Exception e){
|
---|
296 | e.printStackTrace();
|
---|
297 | return false;
|
---|
298 | }
|
---|
299 |
|
---|
300 | return true;
|
---|
301 | }
|
---|
302 |
|
---|
303 | private void calculateRecordsLength(Document output_doc){
|
---|
304 | NodeList records = output_doc.getDocumentElement().getElementsByTagName(RECORD_ELEMENT);
|
---|
305 |
|
---|
306 | for(int i=0;i<records.getLength();i++){
|
---|
307 | Element record = (Element)records.item(i);
|
---|
308 | calculateRecordLength(record);
|
---|
309 | }
|
---|
310 | }
|
---|
311 |
|
---|
312 | private void calculateRecordLength(Element record){
|
---|
313 | int total_length =0;
|
---|
314 | NodeList controlfileds = record.getElementsByTagName(CONTROLFIELD_ELEMENT);
|
---|
315 | for(int i=0;i<controlfileds.getLength();i++){
|
---|
316 | Element controlfiled = (Element)controlfileds.item(i);
|
---|
317 | total_length +=getElementTextValue(controlfiled).length();
|
---|
318 | }
|
---|
319 |
|
---|
320 | NodeList subfileds = record.getElementsByTagName(SUBFIELD_ELEMENT);
|
---|
321 | for(int i=0;i<subfileds.getLength();i++){
|
---|
322 | Element subfiled = (Element)subfileds.item(i);
|
---|
323 | total_length +=getElementTextValue(subfiled).length();
|
---|
324 | }
|
---|
325 |
|
---|
326 | String record_length = total_length+"";
|
---|
327 | //fill in a extra digit as record length needs to be five characters long
|
---|
328 | if (total_length < 10000){
|
---|
329 | record_length = "0"+record_length;
|
---|
330 | if (total_length < 1000){
|
---|
331 | record_length = "0"+record_length;
|
---|
332 | }
|
---|
333 | if (total_length < 100){
|
---|
334 | record_length = "0"+record_length;
|
---|
335 | }
|
---|
336 | if (total_length < 10){
|
---|
337 | record_length = "0"+record_length;
|
---|
338 | }
|
---|
339 |
|
---|
340 | }
|
---|
341 |
|
---|
342 | NodeList leaders = record.getElementsByTagName(LEADER_ELEMENT);
|
---|
343 |
|
---|
344 | //only one leader element
|
---|
345 | if (leaders.getLength() >0){
|
---|
346 | Element leader_element = (Element)leaders.item(0);
|
---|
347 | removeFirstTextNode(leader_element);
|
---|
348 | leader_element.insertBefore(leader_element.getOwnerDocument().createTextNode(record_length),leader_element.getFirstChild());
|
---|
349 | }
|
---|
350 |
|
---|
351 | }
|
---|
352 |
|
---|
353 | private void removeFirstTextNode(Element element){
|
---|
354 | //remove the first text node
|
---|
355 | NodeList children_nodelist = element.getChildNodes();
|
---|
356 | for (int i = 0; i < children_nodelist.getLength(); i++) {
|
---|
357 | Node child_node = children_nodelist.item(i);
|
---|
358 | if (child_node.getNodeType() == Node.TEXT_NODE) {
|
---|
359 | element.removeChild(child_node);
|
---|
360 | return;
|
---|
361 | }
|
---|
362 | }
|
---|
363 |
|
---|
364 | }
|
---|
365 |
|
---|
366 | private String getElementTextValue(Element element)
|
---|
367 | {
|
---|
368 | String text ="";
|
---|
369 |
|
---|
370 | // Find the node child
|
---|
371 | NodeList children_nodelist = element.getChildNodes();
|
---|
372 | for (int i = 0; i < children_nodelist.getLength(); i++) {
|
---|
373 | Node child_node = children_nodelist.item(i);
|
---|
374 | if (child_node.getNodeType() == Node.TEXT_NODE) {
|
---|
375 | text +=child_node.getNodeValue();
|
---|
376 | }
|
---|
377 | }
|
---|
378 |
|
---|
379 | return text;
|
---|
380 | }
|
---|
381 |
|
---|
382 |
|
---|
383 | private void setMappingVariable(Document style_doc){
|
---|
384 | Node child = style_doc.getDocumentElement().getFirstChild();
|
---|
385 | while(child != null) {
|
---|
386 | String name = child.getNodeName();
|
---|
387 | if (name.equals("xsl:variable")) {
|
---|
388 | Element variable_element = (Element)child;
|
---|
389 | if ( variable_element.getAttribute("name").trim().equals("mapping")){
|
---|
390 | variable_element.setAttribute("select","document('"+mapping_file+"')/Mapping");
|
---|
391 | }
|
---|
392 | }
|
---|
393 | child = child.getNextSibling();
|
---|
394 | }
|
---|
395 |
|
---|
396 | }
|
---|
397 |
|
---|
398 | private void setTransformerParams(Transformer transformer)
|
---|
399 | {
|
---|
400 | if(targetlang != "") {
|
---|
401 | transformer.setParameter("sourcelang",sourcelang);
|
---|
402 | transformer.setParameter("targetlang",targetlang);
|
---|
403 | }
|
---|
404 |
|
---|
405 | // handle any custom parameters that are also to be passed into the XSLT
|
---|
406 | Iterator i = paramMap.entrySet().iterator();
|
---|
407 | while(i.hasNext()) {
|
---|
408 | Map.Entry entry = (Map.Entry)i.next();
|
---|
409 | String paramName = (String)entry.getKey();
|
---|
410 | String paramValue = (String)entry.getValue();
|
---|
411 |
|
---|
412 | transformer.setParameter(paramName, paramValue);
|
---|
413 | }
|
---|
414 |
|
---|
415 | }
|
---|
416 |
|
---|
417 | private void translate(String xml_file, String xsl_file, String output_file)throws IOException,TransformerException, TransformerConfigurationException, FileNotFoundException, IOException{
|
---|
418 |
|
---|
419 | TransformerFactory tFactory = TransformerFactory.newInstance();
|
---|
420 | Transformer transformer = tFactory.newTransformer(new StreamSource(xsl_file));
|
---|
421 |
|
---|
422 | OutputStreamWriter output = null;
|
---|
423 | if (output_file.equals("")) {
|
---|
424 | output = new OutputStreamWriter(System.out, "UTF-8");
|
---|
425 | }
|
---|
426 | else{
|
---|
427 | output = new OutputStreamWriter(new FileOutputStream(output_file), "UTF-8");
|
---|
428 | }
|
---|
429 |
|
---|
430 | setTransformerParams(transformer); // sourcelang and targetlang and any further custom parameters to be passed into the XSLT
|
---|
431 | transformer.transform(new StreamSource(new File(xml_file)),new StreamResult(output));
|
---|
432 |
|
---|
433 | }
|
---|
434 | private void translateWithMapping(String xml_file, String xsl_file, String mapping_file, String output_file)throws IOException,TransformerException, TransformerConfigurationException, FileNotFoundException {
|
---|
435 |
|
---|
436 | TransformerFactory tFactory = TransformerFactory.newInstance();
|
---|
437 | Transformer transformer = tFactory.newTransformer(new StreamSource(xsl_file));
|
---|
438 |
|
---|
439 | OutputStreamWriter output = null;
|
---|
440 | if (output_file.equals("")) {
|
---|
441 | output = new OutputStreamWriter(System.out, "UTF-8");
|
---|
442 | }
|
---|
443 | else{
|
---|
444 | output = new OutputStreamWriter(new FileOutputStream(output_file), "UTF-8");
|
---|
445 | }
|
---|
446 | try {
|
---|
447 | Document mapping_doc = DocumentBuilderFactory.newInstance().newDocumentBuilder().parse(mapping_file);
|
---|
448 | Element mapping =mapping_doc.getDocumentElement();
|
---|
449 |
|
---|
450 | transformer.setParameter("mapping",mapping);
|
---|
451 | } catch (Exception e) {
|
---|
452 | System.err.println("Couldn't load in mapping file");
|
---|
453 | e.printStackTrace();
|
---|
454 | }
|
---|
455 | setTransformerParams(transformer); // sourcelang and targetlang and any further custom parameters to be passed into the XSLT
|
---|
456 | transformer.transform(new StreamSource(new File(xml_file)),new StreamResult(output));
|
---|
457 |
|
---|
458 | }
|
---|
459 |
|
---|
460 | static public String replaceAll(String source_string, String match_regexp, String replace_string)
|
---|
461 | {
|
---|
462 | return source_string.replaceAll(match_regexp, replace_string);
|
---|
463 | }
|
---|
464 |
|
---|
465 | // Necessary for paperspast.dm, but can be used generally.
|
---|
466 | // The get-chunks cmd of gti.pl perl script when run over paperspast.dm returns XML with source and target lines
|
---|
467 | // like: [c=paperspast] {All newspapers} for source and [c=paperspast,l=mi] {Niupepa katoa} for target
|
---|
468 | // This function returns just the 'string' portion of the chunk of data: e.g 'All newspapers' and 'Niupepa katoa'
|
---|
469 | static public String getChunkString(String target_file_text)
|
---|
470 | {
|
---|
471 | int startindex = target_file_text.indexOf("[");
|
---|
472 | if(startindex != 0) {
|
---|
473 | return target_file_text;
|
---|
474 | } // to test that the input requires processing
|
---|
475 |
|
---|
476 | // else
|
---|
477 | startindex = target_file_text.indexOf("{");
|
---|
478 | int endindex = target_file_text.lastIndexOf("}");
|
---|
479 | if(startindex != -1 && endindex != -1) {
|
---|
480 | return target_file_text.substring(startindex+1, endindex); // skips { and }
|
---|
481 | } else {
|
---|
482 | return target_file_text;
|
---|
483 | }
|
---|
484 |
|
---|
485 | }
|
---|
486 |
|
---|
487 | // Necessary for paperspast.dm, but can be used generally.
|
---|
488 | // The get-chunks cmd of gti.pl perl script when run over paperspast.dm returns XML with source and target lines
|
---|
489 | // like: [c=paperspast] {All newspapers} for source and [c=paperspast,l=mi] {Niupepa katoa} for target
|
---|
490 | // This function returns just the 'attribute' portion of the chunk of data: e.g 'c=paperspast' and 'c=paperspast,l=mi'
|
---|
491 | static public String getChunkAttr(String target_file_text)
|
---|
492 | {
|
---|
493 | int startindex = target_file_text.indexOf("[");
|
---|
494 | if(startindex != 0) {
|
---|
495 | return target_file_text;
|
---|
496 | } // to test that the input requires processing
|
---|
497 |
|
---|
498 | // else
|
---|
499 | startindex = target_file_text.indexOf("{");
|
---|
500 | int endindex = target_file_text.lastIndexOf("}");
|
---|
501 | if(startindex != -1 && endindex != -1) {
|
---|
502 | endindex = target_file_text.lastIndexOf("]", startindex); // look for ] preceding the {
|
---|
503 | if(endindex > 1) { //if(endindex != -1) {
|
---|
504 | // so there's something to substring between [ and ]
|
---|
505 | return target_file_text.substring(1, endindex).trim(); // skips [ and ]
|
---|
506 | }
|
---|
507 | }
|
---|
508 | return target_file_text;
|
---|
509 | }
|
---|
510 |
|
---|
511 | public static void main(String[] args)
|
---|
512 | {
|
---|
513 | String xml_file="";
|
---|
514 | String xsl_file="";
|
---|
515 | String mapping_file="";
|
---|
516 | String output_file="";
|
---|
517 |
|
---|
518 | String sourcelang="";
|
---|
519 | String targetlang="";
|
---|
520 |
|
---|
521 | boolean readFromStdInFlag = false;
|
---|
522 |
|
---|
523 | HashMap paramMap = new HashMap();
|
---|
524 | int index = -1; // index of the '=' sign in cmdline argument specifying custom parameters to be passed into the XSLT
|
---|
525 |
|
---|
526 | // Checking Arguments
|
---|
527 | if(args.length < 1)
|
---|
528 | {
|
---|
529 | printUsage();
|
---|
530 | }
|
---|
531 |
|
---|
532 | for (int i=0;i<args.length;i++){
|
---|
533 | if (args[i].equals("-m") && i+1 < args.length && !args[i+1].startsWith("-")){
|
---|
534 | mapping_file = args[++i];
|
---|
535 | checkFile(mapping_file.replaceAll("file:///",""));
|
---|
536 | }
|
---|
537 | else if (args[i].equals("-x") && i+1 < args.length && !args[i+1].startsWith("-")){
|
---|
538 | xml_file = args[++i];
|
---|
539 | checkFile(xml_file.replaceAll("file:///",""));
|
---|
540 | }
|
---|
541 | else if(args[i].equals("-t") && i+1 < args.length && !args[i+1].startsWith("-")){
|
---|
542 | xsl_file = args[++i];
|
---|
543 | checkFile( xsl_file.replaceAll("file:///",""));
|
---|
544 | }
|
---|
545 | else if(args[i].equals("-o") && i+1 < args.length && !args[i+1].startsWith("-")){
|
---|
546 | output_file = args[++i];
|
---|
547 |
|
---|
548 | }
|
---|
549 | // The two language parameters (-s and -l) are for the gti-generate-tmx-xml file
|
---|
550 | // which requires the target lang (code), and will accept the optional source lang (code)
|
---|
551 | else if(args[i].equals("-s") && i+1 < args.length && !args[i+1].startsWith("-")){
|
---|
552 | sourcelang = args[++i];
|
---|
553 | }
|
---|
554 | else if(args[i].equals("-l") && i+1 < args.length && !args[i+1].startsWith("-")){
|
---|
555 | targetlang = args[++i];
|
---|
556 | }
|
---|
557 | else if(args[i].equals("-c")){
|
---|
558 | readFromStdInFlag = true;
|
---|
559 | }
|
---|
560 | else if(args[i].equals("-h")){
|
---|
561 | printUsage();
|
---|
562 | }
|
---|
563 | else if ((index = args[i].indexOf("=")) != -1) { // custom parameters provided on the cmdline in the form paramName1=paramValue1 paramName2=paramValue2 etc
|
---|
564 | // that are to be passed into the XSLT
|
---|
565 | String paramName = args[i].substring(0, index);
|
---|
566 | String paramValue = args[i].substring(index+1); // skip the = sign
|
---|
567 | paramMap.put(paramName, paramValue);
|
---|
568 | index = -1;
|
---|
569 | }
|
---|
570 | else{
|
---|
571 | printUsage();
|
---|
572 | }
|
---|
573 |
|
---|
574 | }
|
---|
575 |
|
---|
576 |
|
---|
577 | ApplyXSLT core = null;
|
---|
578 |
|
---|
579 | if (xml_file.equals("") && !xsl_file.equals("")){//read from pipe line
|
---|
580 | if (mapping_file.equals("")){
|
---|
581 | core = new ApplyXSLT(xsl_file, sourcelang, targetlang, paramMap);
|
---|
582 | }
|
---|
583 | else{
|
---|
584 | core = new ApplyXSLT(xsl_file, mapping_file, sourcelang, targetlang, paramMap);
|
---|
585 | }
|
---|
586 |
|
---|
587 | if (core != null){
|
---|
588 | if(readFromStdInFlag) { // ApplyXSLT was run with -c: read from pipe but no <?DocStart?><?DocEnd?> markers
|
---|
589 | core.processPipedFromStdIn();
|
---|
590 | }
|
---|
591 | core.process(); //read from pipe line, but expecting <?DocStart?><?DocEnd?> embedding markers
|
---|
592 | }
|
---|
593 | else{
|
---|
594 | printUsage();
|
---|
595 | }
|
---|
596 | }
|
---|
597 | else if(!xml_file.equals("") && !xsl_file.equals("")){
|
---|
598 | core = new ApplyXSLT(sourcelang, targetlang, paramMap);
|
---|
599 | try {
|
---|
600 | if (mapping_file.equals("")) {
|
---|
601 | core.translate(xml_file,xsl_file,output_file);
|
---|
602 | } else {
|
---|
603 | core.translateWithMapping(xml_file,xsl_file,mapping_file, output_file);
|
---|
604 | }
|
---|
605 | }
|
---|
606 | catch(Exception e){e.printStackTrace();}
|
---|
607 | }
|
---|
608 | else{
|
---|
609 | printUsage();
|
---|
610 | }
|
---|
611 |
|
---|
612 | }
|
---|
613 |
|
---|
614 | private static void checkFile(String filename){
|
---|
615 | File file = new File(filename);
|
---|
616 | if (!file.exists()){
|
---|
617 | System.out.println("Error: "+filename+" doesn't exist!");
|
---|
618 | System.exit(-1);
|
---|
619 | }
|
---|
620 | }
|
---|
621 |
|
---|
622 | private static void printUsage(){
|
---|
623 | System.out.println("Usage: ApplyXSLT -x File -t File [-m File] [-o File] [-s sourcelang] [-l targetlang] [param-name=param-value]");
|
---|
624 | System.out.println("\t-x specifies the xml file (Note: optional for piped xml data)");
|
---|
625 | System.out.println("\t-c read xml file piped from stdin but without DocStart/DocEnd markers. Writes to stdout");
|
---|
626 | System.out.println("\t-t specifies the xsl file");
|
---|
627 | System.out.println("\t-m specifies the mapping file (for MARCXMLPlugout.pm only)");
|
---|
628 | System.out.println("\t-o specifies the output file name (output to screen if this option is absent)");
|
---|
629 | System.out.println("\t-s specifies the input language code for generating TMX file. Defaults to 'en' if none is provided");
|
---|
630 | System.out.println("\t-l specifies the output language code. Required if generating a TMX file.");
|
---|
631 | System.out.println("\tFor general transformations of an XML by an XSLT, you can pass in parameter name=value pairs if any need to passed on into the XSLT as xsl params.");
|
---|
632 | System.exit(-1);
|
---|
633 | }
|
---|
634 |
|
---|
635 |
|
---|
636 | // StreamGobblers used in reading/writing to Process' input and outputstreams can be re-used more generally.
|
---|
637 | // Here in ApplyXSTL.java we use it to read from a pipe line (stdin piped into this ApplyXSLT.java)
|
---|
638 | // Code based on http://www.javaworld.com/article/2071275/core-java/when-runtime-exec---won-t.html?page=2
|
---|
639 | class ReadStreamGobbler extends Thread
|
---|
640 | {
|
---|
641 | InputStream is = null;
|
---|
642 | StringBuffer outputstr = new StringBuffer();
|
---|
643 | boolean split_newlines = false;
|
---|
644 |
|
---|
645 |
|
---|
646 | public ReadStreamGobbler(InputStream is)
|
---|
647 | {
|
---|
648 | this.is = is;
|
---|
649 | split_newlines = false;
|
---|
650 | }
|
---|
651 |
|
---|
652 | public ReadStreamGobbler(InputStream is, boolean split_newlines)
|
---|
653 | {
|
---|
654 | this.is = is;
|
---|
655 | this.split_newlines = split_newlines;
|
---|
656 | }
|
---|
657 |
|
---|
658 | public void run()
|
---|
659 | {
|
---|
660 | BufferedReader br = null;
|
---|
661 | try {
|
---|
662 | br = new BufferedReader(new InputStreamReader(is, "UTF-8"));
|
---|
663 | String line=null;
|
---|
664 | while ( (line = br.readLine()) != null) {
|
---|
665 | //System.out.println("@@@ GOT LINE: " + line);
|
---|
666 | outputstr.append(line);
|
---|
667 | if(split_newlines) {
|
---|
668 | outputstr.append("\n");
|
---|
669 | }
|
---|
670 | }
|
---|
671 | } catch (IOException ioe) {
|
---|
672 | ioe.printStackTrace();
|
---|
673 | } finally {
|
---|
674 | System.err.println("ReadStreamGobbler:run() finished. Closing resource");
|
---|
675 | closeResource(br);
|
---|
676 | }
|
---|
677 | }
|
---|
678 |
|
---|
679 | public String getOutput() {
|
---|
680 | return outputstr.toString();
|
---|
681 | }
|
---|
682 |
|
---|
683 | // http://docs.oracle.com/javase/tutorial/essential/exceptions/finally.html
|
---|
684 | // http://stackoverflow.com/questions/481446/throws-exception-in-finally-blocks
|
---|
685 | public void closeResource(Closeable resourceHandle) {
|
---|
686 | try {
|
---|
687 | if(resourceHandle != null) {
|
---|
688 | resourceHandle.close();
|
---|
689 | resourceHandle = null;
|
---|
690 | }
|
---|
691 | } catch(Exception e) {
|
---|
692 | System.err.println("Exception closing resource: " + e.getMessage());
|
---|
693 | e.printStackTrace();
|
---|
694 | }
|
---|
695 | }
|
---|
696 | }
|
---|
697 |
|
---|
698 | }
|
---|
699 |
|
---|
700 |
|
---|