[950] | 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.io.BufferedReader; |
---|
| 18 | import java.io.File; |
---|
| 19 | import java.io.FileInputStream; |
---|
| 20 | import java.io.FileNotFoundException; |
---|
| 21 | import java.io.IOException; |
---|
| 22 | import java.io.InputStreamReader; |
---|
| 23 | import java.net.MalformedURLException; |
---|
| 24 | import java.net.URL; |
---|
| 25 | import java.util.ArrayList; |
---|
| 26 | import java.util.Collection; |
---|
| 27 | import java.util.List; |
---|
| 28 | import java.util.logging.Level; |
---|
| 29 | import java.util.regex.Matcher; |
---|
| 30 | import java.util.regex.Pattern; |
---|
| 31 | |
---|
| 32 | import de.ugoe.cs.autoquest.eventcore.Event; |
---|
| 33 | import de.ugoe.cs.autoquest.eventcore.gui.IInteraction; |
---|
| 34 | import de.ugoe.cs.autoquest.eventcore.gui.KeyboardFocusChange; |
---|
| 35 | import de.ugoe.cs.autoquest.eventcore.gui.MouseButtonInteraction; |
---|
| 36 | import de.ugoe.cs.autoquest.eventcore.gui.MouseClick; |
---|
| 37 | import de.ugoe.cs.autoquest.eventcore.gui.Scroll; |
---|
| 38 | import de.ugoe.cs.autoquest.eventcore.gui.TextInput; |
---|
| 39 | import de.ugoe.cs.autoquest.eventcore.guimodel.GUIElementFactory; |
---|
| 40 | import de.ugoe.cs.autoquest.eventcore.guimodel.GUIModel; |
---|
| 41 | import de.ugoe.cs.autoquest.eventcore.guimodel.GUIModelException; |
---|
| 42 | import de.ugoe.cs.autoquest.eventcore.guimodel.IGUIElement; |
---|
| 43 | import de.ugoe.cs.autoquest.eventcore.guimodel.IGUIElementSpec; |
---|
| 44 | import de.ugoe.cs.autoquest.eventcore.guimodel.ITextArea; |
---|
| 45 | import de.ugoe.cs.autoquest.eventcore.guimodel.ITextField; |
---|
| 46 | import de.ugoe.cs.autoquest.plugin.html.eventcore.HTMLEvent; |
---|
| 47 | import de.ugoe.cs.autoquest.plugin.html.guimodel.HTMLGUIElementSpec; |
---|
| 48 | import de.ugoe.cs.util.FileTools; |
---|
| 49 | import de.ugoe.cs.util.console.Console; |
---|
| 50 | |
---|
| 51 | /** |
---|
| 52 | * <p> |
---|
| 53 | * TODO comment |
---|
| 54 | * </p> |
---|
| 55 | * |
---|
| 56 | * @author Patrick Harms |
---|
| 57 | */ |
---|
| 58 | public class HTMLLogParser { |
---|
| 59 | |
---|
| 60 | /** |
---|
| 61 | * <p> |
---|
| 62 | * Name and path of the robot filter. |
---|
| 63 | * </p> |
---|
| 64 | */ |
---|
| 65 | private static final String ROBOTFILTERFILE = "data/robots/robotfilter.txt"; |
---|
| 66 | |
---|
| 67 | /** |
---|
| 68 | * <p> |
---|
| 69 | * Field that contains a regular expression that matches all robots |
---|
| 70 | * contained in {@link #ROBOTFILTERFILE}. |
---|
| 71 | * </p> |
---|
| 72 | */ |
---|
| 73 | private String robotRegex = null; |
---|
| 74 | |
---|
| 75 | /** |
---|
| 76 | * |
---|
| 77 | */ |
---|
| 78 | private Pattern htmlElementPattern = Pattern.compile("(\\w+)(\\[(\\d+)\\]|\\(id=([\\w-]+)\\))"); |
---|
| 79 | |
---|
| 80 | /** |
---|
| 81 | * |
---|
| 82 | */ |
---|
| 83 | private List<List<Event>> sequences; |
---|
| 84 | |
---|
| 85 | /** |
---|
| 86 | * |
---|
| 87 | */ |
---|
| 88 | private GUIModel guiModel; |
---|
| 89 | |
---|
| 90 | /** |
---|
| 91 | * <p> |
---|
| 92 | * TODO: comment |
---|
| 93 | * </p> |
---|
| 94 | * |
---|
| 95 | * @param source |
---|
| 96 | */ |
---|
| 97 | public void parseFile(String source) throws IllegalArgumentException { |
---|
| 98 | if (source == null) { |
---|
| 99 | throw new IllegalArgumentException("source must not be null"); |
---|
| 100 | } |
---|
| 101 | |
---|
| 102 | parseFile(new File(source)); |
---|
| 103 | } |
---|
| 104 | |
---|
| 105 | /** |
---|
| 106 | * <p> |
---|
| 107 | * TODO: comment |
---|
| 108 | * </p> |
---|
| 109 | * |
---|
| 110 | * @param source |
---|
| 111 | */ |
---|
| 112 | public void parseFile(File source) throws IllegalArgumentException { |
---|
| 113 | if (source == null) { |
---|
| 114 | throw new IllegalArgumentException("source must not be null"); |
---|
| 115 | } |
---|
| 116 | else if (!source.exists()) { |
---|
| 117 | throw new IllegalArgumentException("source file does not exist"); |
---|
| 118 | } |
---|
| 119 | else if (!source.isFile()) { |
---|
| 120 | throw new IllegalArgumentException("source is not a file"); |
---|
| 121 | } |
---|
| 122 | |
---|
| 123 | guiModel = new GUIModel(); |
---|
| 124 | |
---|
| 125 | BufferedReader reader = null; |
---|
| 126 | |
---|
| 127 | try { |
---|
| 128 | reader = |
---|
| 129 | new BufferedReader(new InputStreamReader(new FileInputStream(source), "UTF-8")); |
---|
| 130 | |
---|
| 131 | loadRobotRegex(); |
---|
| 132 | |
---|
| 133 | List<Event> sequence = new ArrayList<Event>(); |
---|
| 134 | |
---|
| 135 | int lineCounter = 0; |
---|
| 136 | String line = reader.readLine(); |
---|
| 137 | while ((line != null) && (!"".equals(line))) { |
---|
| 138 | lineCounter++; |
---|
| 139 | String[] values = line.substring(1, line.length() - 1).split("\" \""); |
---|
| 140 | |
---|
| 141 | String clientId = values[0]; |
---|
| 142 | long timestamp = Long.parseLong(values[1]); |
---|
| 143 | String title = values[2]; |
---|
| 144 | String uriString = values[3]; |
---|
| 145 | String agent = values[4]; |
---|
| 146 | String eventName = values[5]; |
---|
| 147 | String htmlElementPath = values[6]; |
---|
| 148 | |
---|
| 149 | List<String> eventParameters = null; |
---|
| 150 | if (values.length > 7) { |
---|
| 151 | eventParameters = new ArrayList<String>(); |
---|
| 152 | |
---|
| 153 | for (int i = 7; i < values.length; i++) { |
---|
| 154 | eventParameters.add(values[i]); |
---|
| 155 | } |
---|
| 156 | } |
---|
| 157 | |
---|
| 158 | if (isRobot(agent)) { |
---|
| 159 | // do not handle sessions of robots |
---|
| 160 | Console.println("ignoring robot session: " + agent); |
---|
| 161 | break; |
---|
| 162 | } |
---|
| 163 | |
---|
| 164 | |
---|
| 165 | try { |
---|
| 166 | URL url = new URL(uriString); |
---|
| 167 | IGUIElement guiElement = getGUIElement(url, title, htmlElementPath); |
---|
| 168 | IInteraction interaction = |
---|
| 169 | getInteraction(eventName, guiElement, eventParameters); |
---|
| 170 | |
---|
| 171 | if (interaction != null) { |
---|
| 172 | Event event = |
---|
| 173 | createEvent(clientId, interaction, guiElement, timestamp, agent); |
---|
| 174 | sequence.add(event); |
---|
| 175 | } |
---|
| 176 | } |
---|
| 177 | catch (MalformedURLException e) { |
---|
| 178 | Console.traceln(Level.FINE, "Ignored line " + lineCounter + ": " + |
---|
| 179 | e.getMessage()); |
---|
| 180 | } |
---|
| 181 | |
---|
| 182 | line = reader.readLine(); |
---|
| 183 | } |
---|
| 184 | |
---|
| 185 | Console.traceln(Level.INFO, "read user sequence with " + sequence.size() + |
---|
| 186 | " events from " + source); |
---|
| 187 | |
---|
| 188 | sequences = new ArrayList<List<Event>>(); |
---|
| 189 | sequences.add(sequence); |
---|
| 190 | } |
---|
| 191 | catch (Exception e) { |
---|
| 192 | Console.printerrln("could not parse file " + source + ": " + e); |
---|
| 193 | e.printStackTrace(); |
---|
| 194 | } |
---|
| 195 | } |
---|
| 196 | |
---|
| 197 | /** |
---|
| 198 | * <p> |
---|
| 199 | * TODO: comment |
---|
| 200 | * </p> |
---|
| 201 | * |
---|
| 202 | * @return |
---|
| 203 | */ |
---|
| 204 | public Collection<List<Event>> getSequences() { |
---|
| 205 | return sequences; |
---|
| 206 | } |
---|
| 207 | |
---|
| 208 | /** |
---|
| 209 | * <p> |
---|
| 210 | * TODO: comment |
---|
| 211 | * </p> |
---|
| 212 | * |
---|
| 213 | * @return |
---|
| 214 | */ |
---|
| 215 | public GUIModel getGuiModel() { |
---|
| 216 | return guiModel; |
---|
| 217 | } |
---|
| 218 | |
---|
| 219 | /** |
---|
| 220 | * <p> |
---|
| 221 | * Reads {@link #ROBOTFILTERFILE} and creates a regular expression that |
---|
| 222 | * matches all the robots defined in the file. The regular expression is |
---|
| 223 | * stored in the field {@link #robotRegex}. |
---|
| 224 | * </p> |
---|
| 225 | * |
---|
| 226 | * @throws IOException |
---|
| 227 | * thrown if there is a problem reading the robot filter |
---|
| 228 | * @throws FileNotFoundException |
---|
| 229 | * thrown if the robot filter is not found |
---|
| 230 | */ |
---|
| 231 | private void loadRobotRegex() throws IOException, FileNotFoundException { |
---|
| 232 | String[] lines = FileTools.getLinesFromFile(ROBOTFILTERFILE); |
---|
| 233 | StringBuilder regex = new StringBuilder(); |
---|
| 234 | for (int i = 0; i < lines.length; i++) { |
---|
| 235 | regex.append("(.*" + lines[i] + ".*)"); |
---|
| 236 | if (i != lines.length - 1) { |
---|
| 237 | regex.append('|'); |
---|
| 238 | } |
---|
| 239 | } |
---|
| 240 | robotRegex = regex.toString(); |
---|
| 241 | } |
---|
| 242 | |
---|
| 243 | /** |
---|
| 244 | * <p> |
---|
| 245 | * TODO: comment |
---|
| 246 | * </p> |
---|
| 247 | * |
---|
| 248 | * @param htmlElementPath |
---|
| 249 | * @return |
---|
| 250 | * @throws GUIModelException |
---|
| 251 | * @throws |
---|
| 252 | */ |
---|
| 253 | private IGUIElement getGUIElement(URL url, String title, String htmlElementPath) |
---|
| 254 | throws GUIModelException |
---|
| 255 | { |
---|
| 256 | String[] pathElements = htmlElementPath.split("/"); |
---|
| 257 | List<IGUIElementSpec> guiElementPath = new ArrayList<IGUIElementSpec>(); |
---|
| 258 | |
---|
| 259 | for (String pathElement : pathElements) { |
---|
| 260 | if ((pathElement != null) && (!"".equals(pathElement))) { |
---|
| 261 | Matcher matcher = htmlElementPattern.matcher(pathElement); |
---|
| 262 | if (!matcher.matches()) { |
---|
| 263 | throw new IllegalArgumentException("could not parse HTML element " + pathElement); |
---|
| 264 | } |
---|
| 265 | |
---|
| 266 | String type = matcher.group(1); |
---|
| 267 | String indexStr = matcher.group(3); |
---|
| 268 | String id = matcher.group(4); |
---|
| 269 | |
---|
| 270 | int index = -1; |
---|
| 271 | |
---|
| 272 | if ((indexStr != null) && (!"".equals(indexStr))) { |
---|
| 273 | index = Integer.parseInt(indexStr); |
---|
| 274 | } |
---|
| 275 | |
---|
| 276 | guiElementPath.add(new HTMLGUIElementSpec(url, title, type, id, index)); |
---|
| 277 | } |
---|
| 278 | } |
---|
| 279 | |
---|
| 280 | return guiModel.integratePath(guiElementPath, GUIElementFactory.getInstance()); |
---|
| 281 | } |
---|
| 282 | |
---|
| 283 | /** |
---|
| 284 | * <p> |
---|
| 285 | * TODO: comment |
---|
| 286 | * </p> |
---|
| 287 | * |
---|
| 288 | * @param eventName |
---|
| 289 | * @param guiElement |
---|
| 290 | * @param eventParameters |
---|
| 291 | * @return |
---|
| 292 | */ |
---|
| 293 | private IInteraction getInteraction(String eventName, |
---|
| 294 | IGUIElement guiElement, |
---|
| 295 | List<String> eventParameters) |
---|
| 296 | { |
---|
| 297 | IInteraction result = null; |
---|
| 298 | |
---|
| 299 | if ("onclick".equals(eventName)) { |
---|
| 300 | int[] coordinates = |
---|
| 301 | getCoordinateParameter(eventName, eventParameters, 0, "click coordinates"); |
---|
| 302 | result = new MouseClick |
---|
| 303 | (MouseButtonInteraction.Button.LEFT, coordinates[0], coordinates[1]); |
---|
| 304 | } |
---|
| 305 | else if ("onfocus".equals(eventName)) { |
---|
| 306 | result = new KeyboardFocusChange(); |
---|
| 307 | } |
---|
| 308 | else if ("onscroll".equals(eventName)) { |
---|
| 309 | int[] coordinates = |
---|
| 310 | getCoordinateParameter(eventName, eventParameters, 0, "click coordinates"); |
---|
| 311 | result = new Scroll(coordinates[0], coordinates[1]); |
---|
| 312 | } |
---|
| 313 | else if ("onchange".equals(eventName)) { |
---|
| 314 | String value = getStringParameter(eventName, eventParameters, 0, "selected value"); |
---|
| 315 | |
---|
| 316 | if ((guiElement instanceof ITextArea) || (guiElement instanceof ITextField)) { |
---|
| 317 | result = new TextInput(value, null); |
---|
| 318 | } |
---|
| 319 | else { |
---|
| 320 | throw new IllegalArgumentException("can not handle onchange events on GUI " + |
---|
| 321 | "elements of type " + guiElement.getClass()); |
---|
| 322 | } |
---|
| 323 | } |
---|
| 324 | else if ("onunload".equals(eventName) || "onbeforeunload".equals(eventName) || |
---|
| 325 | "onpagehide".equals(eventName)) |
---|
| 326 | { |
---|
| 327 | Console.traceln(Level.FINE, "Ignored event name \"onunload\""); |
---|
| 328 | } |
---|
| 329 | else { |
---|
| 330 | throw new IllegalArgumentException("unknown event name: \"" + eventName + "\""); |
---|
| 331 | } |
---|
| 332 | |
---|
| 333 | return result; |
---|
| 334 | } |
---|
| 335 | |
---|
| 336 | /** |
---|
| 337 | * <p> |
---|
| 338 | * TODO: comment |
---|
| 339 | * </p> |
---|
| 340 | * |
---|
| 341 | * @param clientId |
---|
| 342 | * @param interaction |
---|
| 343 | * @param guiElement |
---|
| 344 | * @param timestamp |
---|
| 345 | * @param agent |
---|
| 346 | * @return |
---|
| 347 | */ |
---|
| 348 | private Event createEvent(String clientId, |
---|
| 349 | IInteraction interaction, |
---|
| 350 | IGUIElement guiElement, |
---|
| 351 | long timestamp, |
---|
| 352 | String agent) |
---|
| 353 | { |
---|
| 354 | return new HTMLEvent(clientId, interaction, guiElement, timestamp, agent); |
---|
| 355 | } |
---|
| 356 | |
---|
| 357 | /** |
---|
| 358 | * <p> |
---|
| 359 | * TODO: comment |
---|
| 360 | * </p> |
---|
| 361 | * |
---|
| 362 | * @param eventName |
---|
| 363 | * @param eventParameters |
---|
| 364 | * @param i |
---|
| 365 | * @param string |
---|
| 366 | * @return |
---|
| 367 | */ |
---|
| 368 | private String getStringParameter(String eventName, |
---|
| 369 | List<String> eventParameters, |
---|
| 370 | int parameterIndex, |
---|
| 371 | String parameterDesc) |
---|
| 372 | throws IllegalArgumentException |
---|
| 373 | { |
---|
| 374 | String value = |
---|
| 375 | eventParameters.size() > parameterIndex ? eventParameters.get(parameterIndex) : null; |
---|
| 376 | |
---|
| 377 | if (value == null) { |
---|
| 378 | throw new IllegalArgumentException |
---|
| 379 | (eventName + " event does not provide the " + parameterDesc); |
---|
| 380 | } |
---|
| 381 | |
---|
| 382 | return value; |
---|
| 383 | } |
---|
| 384 | |
---|
| 385 | /** |
---|
| 386 | * <p> |
---|
| 387 | * TODO: comment |
---|
| 388 | * </p> |
---|
| 389 | * |
---|
| 390 | * @param eventName |
---|
| 391 | * @param eventParameters |
---|
| 392 | * @return |
---|
| 393 | */ |
---|
| 394 | private int[] getCoordinateParameter(String eventName, |
---|
| 395 | List<String> eventParameters, |
---|
| 396 | int parameterIndex, |
---|
| 397 | String parameterDesc) |
---|
| 398 | { |
---|
| 399 | String value = |
---|
| 400 | getStringParameter(eventName, eventParameters, parameterIndex, parameterDesc); |
---|
| 401 | |
---|
| 402 | String[] intStrs = value.split(","); |
---|
| 403 | |
---|
| 404 | if ((intStrs == null) || (intStrs.length != 2)) { |
---|
| 405 | throw new IllegalArgumentException("the " + parameterDesc + " of an " + eventName + |
---|
| 406 | " event does not provide two correct coordinates"); |
---|
| 407 | } |
---|
| 408 | |
---|
| 409 | try { |
---|
| 410 | return new int[] { Integer.parseInt(intStrs[0]), Integer.parseInt(intStrs[1]) }; |
---|
| 411 | } |
---|
| 412 | catch (NumberFormatException e) { |
---|
| 413 | throw new IllegalArgumentException("the coordinates provided as " + parameterDesc + |
---|
| 414 | " of an " + eventName + " event are no numbers"); |
---|
| 415 | } |
---|
| 416 | } |
---|
| 417 | |
---|
| 418 | /** |
---|
| 419 | * <p> |
---|
| 420 | * Checks whether an agent is a robot. |
---|
| 421 | * </p> |
---|
| 422 | * |
---|
| 423 | * @param agent |
---|
| 424 | * agent that is checked |
---|
| 425 | * @return true, if the agent is a robot; false otherwise |
---|
| 426 | */ |
---|
| 427 | private boolean isRobot(String agent) { |
---|
| 428 | return agent.matches(robotRegex); |
---|
| 429 | } |
---|
| 430 | |
---|
| 431 | } |
---|