1 | // Copyright 2012 Georg-August-Universität Göttingen, Germany |
---|
2 | // |
---|
3 | // Licensed under the Apache License, Version 2.0 (the "License"); |
---|
4 | // you may not use this file except in compliance with the License. |
---|
5 | // You may obtain a copy of the License at |
---|
6 | // |
---|
7 | // http://www.apache.org/licenses/LICENSE-2.0 |
---|
8 | // |
---|
9 | // Unless required by applicable law or agreed to in writing, software |
---|
10 | // distributed under the License is distributed on an "AS IS" BASIS, |
---|
11 | // WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
---|
12 | // See the License for the specific language governing permissions and |
---|
13 | // limitations under the License. |
---|
14 | |
---|
15 | package de.ugoe.cs.autoquest.plugin.html; |
---|
16 | |
---|
17 | import java.util.Arrays; |
---|
18 | import java.util.List; |
---|
19 | import java.util.Map; |
---|
20 | import java.util.regex.Matcher; |
---|
21 | import java.util.regex.Pattern; |
---|
22 | |
---|
23 | import org.xml.sax.SAXException; |
---|
24 | |
---|
25 | import de.ugoe.cs.autoquest.eventcore.Event; |
---|
26 | import de.ugoe.cs.autoquest.eventcore.IEventType; |
---|
27 | import de.ugoe.cs.autoquest.eventcore.guimodel.GUIModel; |
---|
28 | import de.ugoe.cs.autoquest.eventcore.guimodel.GUIModelException; |
---|
29 | import de.ugoe.cs.autoquest.eventcore.guimodel.IGUIElement; |
---|
30 | import de.ugoe.cs.autoquest.plugin.html.eventcore.HTMLEventTypeFactory; |
---|
31 | import de.ugoe.cs.autoquest.plugin.html.guimodel.HTMLDocumentSpec; |
---|
32 | import de.ugoe.cs.autoquest.plugin.html.guimodel.HTMLGUIElement; |
---|
33 | import de.ugoe.cs.autoquest.plugin.html.guimodel.HTMLGUIElementSpec; |
---|
34 | import de.ugoe.cs.autoquest.plugin.html.guimodel.HTMLPageElement; |
---|
35 | import de.ugoe.cs.autoquest.plugin.html.guimodel.HTMLPageElementSpec; |
---|
36 | import de.ugoe.cs.autoquest.plugin.html.guimodel.HTMLServerSpec; |
---|
37 | |
---|
38 | /** |
---|
39 | * <p> |
---|
40 | * This class provides the functionality to parse XML log files generated by the HTMLMonitor of |
---|
41 | * AutoQUEST. The result of parsing a file is a collection of event sequences and a GUI model |
---|
42 | * </p> |
---|
43 | * |
---|
44 | * @author Fabian Glaser, Patrick Harms |
---|
45 | * @version 1.0 |
---|
46 | * |
---|
47 | */ |
---|
48 | public class HTMLLogParser extends AbstractDefaultLogParser { |
---|
49 | |
---|
50 | /** |
---|
51 | * <p> |
---|
52 | * the pattern used for parsing HTML GUI element paths |
---|
53 | * </p> |
---|
54 | */ |
---|
55 | private Pattern htmlElementPattern = |
---|
56 | Pattern.compile("(\\w+)(\\[(\\d+)\\]|\\(htmlId=([\\w-]+)\\))"); |
---|
57 | |
---|
58 | /** |
---|
59 | * <p> |
---|
60 | * the pattern used for parsing parsing parameters |
---|
61 | * </p> |
---|
62 | */ |
---|
63 | private Pattern htmlElementSpecPattern = |
---|
64 | Pattern.compile("(\\w+)(\\[(\\d+)\\]|\\(htmlId=([\\w-]+)\\))?"); |
---|
65 | |
---|
66 | /** |
---|
67 | * <p> |
---|
68 | * parameters to influence parsing |
---|
69 | * </p> |
---|
70 | */ |
---|
71 | private Map<String, List<String>> parseParams; |
---|
72 | |
---|
73 | /** |
---|
74 | * <p> |
---|
75 | * TODO: comment |
---|
76 | * </p> |
---|
77 | * |
---|
78 | * @param parseParams |
---|
79 | */ |
---|
80 | public HTMLLogParser(Map<String, List<String>> parseParams) { |
---|
81 | this.parseParams = parseParams; |
---|
82 | |
---|
83 | for (String paramKey : parseParams.keySet()) { |
---|
84 | if (!"clearId".equals(paramKey) && !"clearIndex".equals(paramKey)) { |
---|
85 | throw new IllegalArgumentException("unknown parse parameter key " + paramKey); |
---|
86 | } |
---|
87 | } |
---|
88 | } |
---|
89 | |
---|
90 | /* (non-Javadoc) |
---|
91 | * @see de.ugoe.cs.autoquest.plugin.html.AbstractDefaultLogParser#handleGUIElement(String, Map) |
---|
92 | */ |
---|
93 | @Override |
---|
94 | protected boolean handleGUIElement(String id, Map<String, String> parameters) |
---|
95 | throws SAXException |
---|
96 | { |
---|
97 | HTMLGUIElementSpec specification = null; |
---|
98 | |
---|
99 | String parentId = parameters.get("parent"); |
---|
100 | HTMLGUIElement parent = (HTMLGUIElement) super.getGUIElementTree().find(parentId); |
---|
101 | |
---|
102 | if (parameters.containsKey("host")) { |
---|
103 | // this is a server specification |
---|
104 | int port = 80; |
---|
105 | String portStr = parameters.get("port"); |
---|
106 | |
---|
107 | if (portStr != null) { |
---|
108 | port = Integer.parseInt(portStr); |
---|
109 | } |
---|
110 | |
---|
111 | specification = new HTMLServerSpec(parameters.get("host"), port); |
---|
112 | } |
---|
113 | else if (parameters.containsKey("path")) { |
---|
114 | // this is a document specification |
---|
115 | |
---|
116 | if (parent != null) { |
---|
117 | if (!(parent.getSpecification() instanceof HTMLServerSpec)) { |
---|
118 | throw new SAXException |
---|
119 | ("invalid log: parent GUI element of a document is not of type server"); |
---|
120 | } |
---|
121 | |
---|
122 | specification = new HTMLDocumentSpec |
---|
123 | ((HTMLServerSpec) parent.getSpecification(), parameters.get("path"), |
---|
124 | parameters.get("query"), parameters.get("title")); |
---|
125 | } |
---|
126 | else if (parentId == null) { |
---|
127 | throw new SAXException("invalid log: a document has no parent id"); |
---|
128 | } |
---|
129 | } |
---|
130 | else if (parameters.containsKey("tagname")) { |
---|
131 | String tagName = parameters.get("tagname"); |
---|
132 | |
---|
133 | if (!tagNameMustBeConsidered(tagName)) { |
---|
134 | return true; |
---|
135 | } |
---|
136 | |
---|
137 | if (parent != null) { |
---|
138 | IGUIElement document = parent; |
---|
139 | |
---|
140 | while ((document != null) && |
---|
141 | (!(document.getSpecification() instanceof HTMLDocumentSpec))) |
---|
142 | { |
---|
143 | document = document.getParent(); |
---|
144 | } |
---|
145 | |
---|
146 | if (document == null) { |
---|
147 | throw new SAXException |
---|
148 | ("invalid log: parent hierarchy of a page element does not contain a " + |
---|
149 | "document"); |
---|
150 | } |
---|
151 | |
---|
152 | int index = -1; |
---|
153 | String indexStr = parameters.get("index"); |
---|
154 | |
---|
155 | if ((indexStr != null) && (!"".equals(indexStr))) { |
---|
156 | index = Integer.parseInt(indexStr); |
---|
157 | } |
---|
158 | |
---|
159 | String htmlId = parameters.get("htmlid"); |
---|
160 | |
---|
161 | if (clearIndex(tagName, index, htmlId, parent)) { |
---|
162 | index = -1; |
---|
163 | } |
---|
164 | |
---|
165 | if (clearHTMLId(tagName, index, htmlId, parent)) { |
---|
166 | htmlId = null; |
---|
167 | } |
---|
168 | |
---|
169 | if ((htmlId == null) && (index == -1)) { |
---|
170 | // set at least a default index, if all is to be ignored. |
---|
171 | index = 0; |
---|
172 | } |
---|
173 | |
---|
174 | specification = new HTMLPageElementSpec |
---|
175 | ((HTMLDocumentSpec) document.getSpecification(), tagName, htmlId, index); |
---|
176 | |
---|
177 | } |
---|
178 | else if (parentId == null) { |
---|
179 | throw new SAXException("invalid log: a page element has no parent id"); |
---|
180 | } |
---|
181 | } |
---|
182 | else { |
---|
183 | throw new SAXException("invalid log: unknown GUI element"); |
---|
184 | } |
---|
185 | |
---|
186 | if (specification != null) { |
---|
187 | try { |
---|
188 | super.getGUIElementTree().add(id, parentId, specification); |
---|
189 | } |
---|
190 | catch (GUIModelException e) { |
---|
191 | throw new SAXException("could not handle GUI element with id " + |
---|
192 | id + ": " + e.getMessage(), e); |
---|
193 | } |
---|
194 | return true; |
---|
195 | } |
---|
196 | else { |
---|
197 | return false; |
---|
198 | } |
---|
199 | } |
---|
200 | |
---|
201 | /** |
---|
202 | * <p> |
---|
203 | * TODO: comment |
---|
204 | * </p> |
---|
205 | * |
---|
206 | * @param tagName |
---|
207 | * @param parent |
---|
208 | * @return |
---|
209 | */ |
---|
210 | private boolean clearIndex(String tagName, int index, String id, HTMLGUIElement parent) { |
---|
211 | return clearSomething("clearIndex", tagName, index, id, parent); |
---|
212 | } |
---|
213 | |
---|
214 | /** |
---|
215 | * <p> |
---|
216 | * TODO: comment |
---|
217 | * </p> |
---|
218 | * |
---|
219 | * @param tagName |
---|
220 | * @param parent |
---|
221 | * @return |
---|
222 | */ |
---|
223 | private boolean clearHTMLId(String tagName, int index, String id, HTMLGUIElement parent) { |
---|
224 | return clearSomething("clearId", tagName, index, id, parent); |
---|
225 | } |
---|
226 | |
---|
227 | /** |
---|
228 | * <p> |
---|
229 | * TODO: comment |
---|
230 | * </p> |
---|
231 | * |
---|
232 | * @param tagName |
---|
233 | * @param parent |
---|
234 | * @return |
---|
235 | */ |
---|
236 | private boolean clearSomething(String parseParamId, |
---|
237 | String tagName, |
---|
238 | int index, |
---|
239 | String id, |
---|
240 | HTMLGUIElement parent) |
---|
241 | { |
---|
242 | if (parseParams.containsKey(parseParamId)) { |
---|
243 | for (String spec : parseParams.get(parseParamId)) { |
---|
244 | // determine the specification parts |
---|
245 | if (spec.startsWith("/")) { |
---|
246 | throw new IllegalArgumentException("can not handle absolute specifications"); |
---|
247 | } |
---|
248 | else if (spec.endsWith("/")) { |
---|
249 | throw new IllegalArgumentException("specifications may not end with a /"); |
---|
250 | } |
---|
251 | |
---|
252 | String[] tagSpecs = spec.split("/"); |
---|
253 | |
---|
254 | if (tagMatchesTagSpec(tagName, index, id, parent, tagSpecs)) { |
---|
255 | return true; |
---|
256 | } |
---|
257 | } |
---|
258 | } |
---|
259 | |
---|
260 | return false; |
---|
261 | } |
---|
262 | |
---|
263 | /** |
---|
264 | * <p> |
---|
265 | * TODO: comment |
---|
266 | * </p> |
---|
267 | * |
---|
268 | * @param tagName |
---|
269 | * @param parent |
---|
270 | * @param spec |
---|
271 | * @return |
---|
272 | */ |
---|
273 | private boolean tagMatchesTagSpec(String tagName, |
---|
274 | int index, |
---|
275 | String id, |
---|
276 | HTMLGUIElement parent, |
---|
277 | String[] tagSpecs) |
---|
278 | { |
---|
279 | |
---|
280 | if (tagSpecs.length > 0) { |
---|
281 | Matcher matcher = htmlElementSpecPattern.matcher(tagSpecs[tagSpecs.length - 1]); |
---|
282 | |
---|
283 | if (!matcher.matches()) { |
---|
284 | throw new IllegalArgumentException |
---|
285 | ("illegal tag specification " + tagSpecs[tagSpecs.length - 1]); |
---|
286 | } |
---|
287 | |
---|
288 | if (!tagName.equals(matcher.group(1))) { |
---|
289 | return false; |
---|
290 | } |
---|
291 | |
---|
292 | String idCondition = matcher.group(4); |
---|
293 | |
---|
294 | if (idCondition != null) { |
---|
295 | if (!idCondition.equals(id)) { |
---|
296 | return false; |
---|
297 | } |
---|
298 | } |
---|
299 | |
---|
300 | String indexCondition = matcher.group(3); |
---|
301 | |
---|
302 | if (indexCondition != null) { |
---|
303 | try { |
---|
304 | if (index != Integer.parseInt(indexCondition)) { |
---|
305 | return false; |
---|
306 | } |
---|
307 | } |
---|
308 | catch (NumberFormatException e) { |
---|
309 | throw new IllegalArgumentException |
---|
310 | ("illegal tag index specification " + indexCondition, e); |
---|
311 | } |
---|
312 | } |
---|
313 | |
---|
314 | if (tagSpecs.length > 1) { |
---|
315 | if (parent instanceof HTMLPageElement) { |
---|
316 | return tagMatchesTagSpec(((HTMLPageElement) parent).getTagName(), |
---|
317 | ((HTMLPageElement) parent).getIndex(), |
---|
318 | ((HTMLPageElement) parent).getHtmlId(), |
---|
319 | (HTMLGUIElement) parent.getParent(), |
---|
320 | Arrays.copyOfRange(tagSpecs, 0, tagSpecs.length - 1)); |
---|
321 | } |
---|
322 | else { |
---|
323 | throw new IllegalArgumentException |
---|
324 | ("specification matches documents or servers. This is not supported yet."); |
---|
325 | } |
---|
326 | } |
---|
327 | else { |
---|
328 | return true; |
---|
329 | } |
---|
330 | } |
---|
331 | else { |
---|
332 | return true; |
---|
333 | } |
---|
334 | } |
---|
335 | |
---|
336 | /* (non-Javadoc) |
---|
337 | * @see de.ugoe.cs.autoquest.plugin.html.AbstractDefaultLogParser#handleEvent(String, Map) |
---|
338 | */ |
---|
339 | @Override |
---|
340 | protected boolean handleEvent(String type, Map<String, String> parameters) throws SAXException { |
---|
341 | String targetId = parameters.get("target"); |
---|
342 | |
---|
343 | if (targetId == null) { |
---|
344 | if (parseParams.size() != 0) { |
---|
345 | throw new SAXException |
---|
346 | ("old log file versions can not be parsed with parse parameters"); |
---|
347 | } |
---|
348 | |
---|
349 | String targetDocument = parameters.get("targetDocument"); |
---|
350 | String targetDOMPath = parameters.get("targetDOMPath"); |
---|
351 | |
---|
352 | if ((targetDocument == null) || (targetDOMPath == null)) { |
---|
353 | throw new SAXException("event has no target defined"); |
---|
354 | } |
---|
355 | |
---|
356 | targetId = determineTargetId(targetDocument, targetDOMPath); |
---|
357 | |
---|
358 | if (targetId == null) { |
---|
359 | // the target id can not be determined yet |
---|
360 | return false; |
---|
361 | } |
---|
362 | } |
---|
363 | |
---|
364 | IGUIElement target = super.getGUIElementTree().find(targetId); |
---|
365 | |
---|
366 | if (target == null) { |
---|
367 | // event not processible yet |
---|
368 | return false; |
---|
369 | } |
---|
370 | |
---|
371 | IEventType eventType = |
---|
372 | HTMLEventTypeFactory.getInstance().getEventType(type, parameters, target); |
---|
373 | |
---|
374 | if (eventType != null) { |
---|
375 | Event event = new Event(eventType, target); |
---|
376 | |
---|
377 | String timestampStr = parameters.get("timestamp"); |
---|
378 | |
---|
379 | if (timestampStr != null) { |
---|
380 | event.setTimestamp(Long.parseLong(timestampStr)); |
---|
381 | } |
---|
382 | |
---|
383 | ((HTMLGUIElement) event.getTarget()).markUsed(); |
---|
384 | |
---|
385 | super.addToSequence(event); |
---|
386 | } |
---|
387 | // else ignore unknown event type |
---|
388 | |
---|
389 | return true; |
---|
390 | } |
---|
391 | |
---|
392 | /** |
---|
393 | * <p> |
---|
394 | * used to determine the id of a target denoted by an event. This is only required for older |
---|
395 | * document formats. The new formats use concrete ids. |
---|
396 | * </p> |
---|
397 | */ |
---|
398 | private String determineTargetId(String targetDocument, String targetDOMPath) |
---|
399 | throws SAXException |
---|
400 | { |
---|
401 | IGUIElement document = super.getGUIElementTree().find(targetDocument); |
---|
402 | |
---|
403 | if (document == null) { |
---|
404 | return null; |
---|
405 | } |
---|
406 | |
---|
407 | if (!(document.getSpecification() instanceof HTMLDocumentSpec)) { |
---|
408 | throw new SAXException("an id that should refer to an HTML document refers to" + |
---|
409 | "something else"); |
---|
410 | } |
---|
411 | |
---|
412 | GUIModel model = super.getGUIElementTree().getGUIModel(); |
---|
413 | IGUIElement child = document; |
---|
414 | String[] pathElements = targetDOMPath.split("/"); |
---|
415 | int pathIndex = 0; |
---|
416 | |
---|
417 | HTMLPageElementSpec compareSpec; |
---|
418 | String tagName; |
---|
419 | int index; |
---|
420 | String htmlId; |
---|
421 | |
---|
422 | while ((pathIndex < pathElements.length) && (child != null)) { |
---|
423 | if ((pathElements[pathIndex] != null) && (!"".equals(pathElements[pathIndex]))) { |
---|
424 | Matcher matcher = htmlElementPattern.matcher(pathElements[pathIndex]); |
---|
425 | if (!matcher.matches()) { |
---|
426 | throw new SAXException |
---|
427 | ("could not parse target DOM path element " + pathElements[pathIndex]); |
---|
428 | } |
---|
429 | |
---|
430 | tagName = matcher.group(1); |
---|
431 | String indexStr = matcher.group(3); |
---|
432 | htmlId = matcher.group(4); |
---|
433 | |
---|
434 | index = -1; |
---|
435 | if ((indexStr != null) && (!"".equals(indexStr))) { |
---|
436 | index = Integer.parseInt(indexStr); |
---|
437 | } |
---|
438 | |
---|
439 | compareSpec = new HTMLPageElementSpec |
---|
440 | ((HTMLDocumentSpec) document.getSpecification(), tagName, htmlId, index); |
---|
441 | |
---|
442 | List<IGUIElement> children = model.getChildren(child); |
---|
443 | child = null; |
---|
444 | |
---|
445 | for (IGUIElement candidate : children) { |
---|
446 | if (compareSpec.getSimilarity(candidate.getSpecification())) { |
---|
447 | child = candidate; |
---|
448 | break; |
---|
449 | } |
---|
450 | } |
---|
451 | } |
---|
452 | |
---|
453 | pathIndex++; |
---|
454 | } |
---|
455 | |
---|
456 | if (child != null) { |
---|
457 | return super.getGUIElementTree().find(child); |
---|
458 | } |
---|
459 | else { |
---|
460 | return null; |
---|
461 | } |
---|
462 | } |
---|
463 | |
---|
464 | /** |
---|
465 | * <p> |
---|
466 | * checks if tags with the provided name must be handled in the GUI model. As an example, |
---|
467 | * it is not necessary to handle "head" tags and anything included in them. |
---|
468 | * </p> |
---|
469 | * |
---|
470 | * @param tagName the tag name to check |
---|
471 | * |
---|
472 | * @return true, if the tag must be considered, false else |
---|
473 | */ |
---|
474 | private boolean tagNameMustBeConsidered(String tagName) { |
---|
475 | if (!tagName.startsWith("input_")) { |
---|
476 | for (int i = 0; i < tagName.length(); i++) { |
---|
477 | // all known HTML tags are either letters or digits, but nothing else. Any GUI model |
---|
478 | // containing something different is proprietary and, therefore, ignored. |
---|
479 | if (!Character.isLetterOrDigit(tagName.charAt(i))) { |
---|
480 | return false; |
---|
481 | } |
---|
482 | } |
---|
483 | } |
---|
484 | |
---|
485 | return |
---|
486 | !"head".equals(tagName) && !"title".equals(tagName) && !"script".equals(tagName) && |
---|
487 | !"style".equals(tagName) && !"link".equals(tagName) && !"meta".equals(tagName) && |
---|
488 | !"iframe".equals(tagName) && !"input_hidden".equals(tagName) && |
---|
489 | !"option".equals(tagName) && !"tt".equals(tagName) && !"br".equals(tagName) && |
---|
490 | !"colgroup".equals(tagName) && !"col".equals(tagName) && !"hr".equals(tagName) && |
---|
491 | !"param".equals(tagName) && !"sfmsg".equals(tagName); |
---|
492 | |
---|
493 | } |
---|
494 | |
---|
495 | } |
---|