johncurcio /
ETL4LODPlus
| Total Complexity | 42 |
| Total Lines | 323 |
| Duplicated Lines | 32.82 % |
| Changes | 1 | ||
| Bugs | 0 | Features | 1 |
Duplicate code is one of the most pungent code smells. A rule that is often used is to re-structure code once it is duplicated in three or more places.
Common duplication problems, and corresponding solutions are:
Complex classes like br.ufrj.ppgi.greco.kettle.GraphSemanticLevelMarkerStep often do a lot of different things. To break such a class down, we need to identify a cohesive component within that class. A common approach to find such a component is to look for fields/methods that share the same prefixes, or suffixes.
Once you have determined the fields that belong together, you can apply the Extract Class refactoring. If the component makes sense as a sub-class, Extract Subclass is also a candidate, and is often faster.
| 1 | package br.ufrj.ppgi.greco.kettle; |
||
| 49 | public class GraphSemanticLevelMarkerStep extends BaseStep implements StepInterface { |
||
| 50 | |||
| 51 | static Integer assessedValueLevel; |
||
| 52 | |||
| 53 | public GraphSemanticLevelMarkerStep(StepMeta stepMeta, StepDataInterface stepDataInterface, int copyNr, |
||
| 54 | TransMeta transMeta, Trans trans) { |
||
| 55 | super(stepMeta, stepDataInterface, copyNr, transMeta, trans); |
||
| 56 | } |
||
| 57 | |||
| 58 | @Override |
||
| 59 | public boolean init(StepMetaInterface smi, StepDataInterface sdi) { |
||
| 60 | if (super.init(smi, sdi)) { |
||
| 61 | return true; |
||
| 62 | } else |
||
| 63 | return false; |
||
| 64 | } |
||
| 65 | |||
| 66 | @Override |
||
| 67 | public void dispose(StepMetaInterface smi, StepDataInterface sdi) { |
||
| 68 | super.dispose(smi, sdi); |
||
| 69 | } |
||
| 70 | |||
| 71 | /** |
||
| 72 | * Metodo chamado para cada linha que entra no step |
||
| 73 | */ |
||
| 74 | View Code Duplication | public boolean processRow(StepMetaInterface smi, StepDataInterface sdi) throws KettleException { |
|
|
Duplication
introduced
by
|
|||
| 75 | GraphSemanticLevelMarkerStepMeta meta = (GraphSemanticLevelMarkerStepMeta) smi; |
||
| 76 | GraphSemanticLevelMarkerStepData data = (GraphSemanticLevelMarkerStepData) sdi; |
||
| 77 | |||
| 78 | String rulesFileName = meta.getRulesFilename(); |
||
| 79 | String LOVFileName = meta.getBrowseFilename(); |
||
| 80 | |||
| 81 | // Obtem linha do fluxo de entrada e termina caso nao haja mais entrada |
||
| 82 | Object[] row = getRow(); |
||
| 83 | if (row == null) { // Nao ha mais linhas de dados |
||
| 84 | setOutputDone(); |
||
| 85 | return false; |
||
| 86 | } |
||
| 87 | |||
| 88 | // Executa apenas uma vez. Variavel first definida na superclasse com |
||
| 89 | // valor true |
||
| 90 | if (first) { |
||
| 91 | first = false; |
||
| 92 | |||
| 93 | // Obtem todas as colunas ateh o step anterior. |
||
| 94 | // Chamar apenas apos chamar getRow() |
||
| 95 | data.outputRowMeta = new RowMeta(); |
||
| 96 | |||
| 97 | // Adiciona os metadados do step atual |
||
| 98 | meta.getFields(data.outputRowMeta, getStepname(), null, null, this); |
||
| 99 | } |
||
| 100 | |||
| 101 | // Logica do step |
||
| 102 | // Leitura de campos Input |
||
| 103 | RowMetaInterface rowMeta = getInputRowMeta(); |
||
| 104 | int indexGraph = rowMeta.indexOfValue(meta.getInputGraph()); |
||
| 105 | Object graph = (indexGraph >= 0) ? row[indexGraph] : null; |
||
| 106 | |||
| 107 | // Set output row |
||
| 108 | Method[] methods = graph.getClass().getMethods(); |
||
| 109 | boolean hasListStatements = false; |
||
| 110 | for (Method method : methods) { |
||
| 111 | if (method.getName().equals("listStatements")) { |
||
| 112 | hasListStatements = true; |
||
| 113 | break; |
||
| 114 | } |
||
| 115 | } |
||
| 116 | |||
| 117 | if (hasListStatements) { |
||
| 118 | tripleWriter(graph, null, data, rulesFileName, LOVFileName); |
||
| 119 | } |
||
| 120 | |||
| 121 | return true; |
||
| 122 | } |
||
| 123 | |||
| 124 | private int tripleWriter(Object model, Object[] row, GraphSemanticLevelMarkerStepData data, String rulesFileName, |
||
| 125 | String LOVFileName) throws KettleStepException { |
||
| 126 | int numPutRows = 0; |
||
| 127 | |||
| 128 | try { |
||
| 129 | // Recreates the graph sent by the previous step |
||
| 130 | Model inputRecreatedGraph = recreateGraph(model); |
||
| 131 | |||
| 132 | // Identify inputGraph Semantic Level |
||
| 133 | Statement stamp = markGraphSemanticLevel(inputRecreatedGraph, rulesFileName, LOVFileName); |
||
| 134 | |||
| 135 | // Creates output with the semantic level stamp |
||
| 136 | Object[] outputRow = row; |
||
| 137 | int i = 0; |
||
| 138 | outputRow = RowDataUtil.addValueData(outputRow, i++, stamp.getSubject().toString()); |
||
| 139 | outputRow = RowDataUtil.addValueData(outputRow, i++, stamp.getPredicate().toString()); |
||
| 140 | outputRow = RowDataUtil.addValueData(outputRow, i++, stamp.getObject().toString()); |
||
| 141 | |||
| 142 | // Joga tripla no fluxo |
||
| 143 | putRow(data.outputRowMeta, outputRow); |
||
| 144 | |||
| 145 | numPutRows++; |
||
| 146 | |||
| 147 | } |
||
| 148 | |||
| 149 | catch (SecurityException e) { |
||
| 150 | // TODO Auto-generated catch block |
||
| 151 | e.printStackTrace(); |
||
| 152 | } |
||
| 153 | |||
| 154 | return numPutRows; |
||
| 155 | } |
||
| 156 | |||
| 157 | private Model recreateGraph(Object model) { |
||
| 158 | |||
| 159 | // Recreates a Model from a Object |
||
| 160 | Object it = null; |
||
| 161 | Model inputModel = ModelFactory.createDefaultModel(); |
||
| 162 | |||
| 163 | try { |
||
| 164 | it = model.getClass().getMethod("listStatements").invoke(model); |
||
| 165 | while ((Boolean) it.getClass().getMethod("hasNext").invoke(it)) { |
||
| 166 | Object stmt = it.getClass().getMethod("next").invoke(it); |
||
| 167 | |||
| 168 | String subject = stmt.getClass().getMethod("getSubject").invoke(stmt).toString(); |
||
| 169 | String predicate = stmt.getClass().getMethod("getPredicate").invoke(stmt).toString(); |
||
| 170 | String object = stmt.getClass().getMethod("getObject").invoke(stmt).toString(); |
||
| 171 | |||
| 172 | Resource r = ResourceFactory.createResource(subject); |
||
| 173 | Property p = ResourceFactory.createProperty(predicate); |
||
| 174 | inputModel.add(r, p, object); |
||
| 175 | |||
| 176 | } |
||
| 177 | |||
| 178 | } catch (IllegalAccessException e) { |
||
| 179 | // TODO Auto-generated catch block |
||
| 180 | e.printStackTrace(); |
||
| 181 | } catch (IllegalArgumentException e) { |
||
| 182 | // TODO Auto-generated catch block |
||
| 183 | e.printStackTrace(); |
||
| 184 | } catch (InvocationTargetException e) { |
||
| 185 | // TODO Auto-generated catch block |
||
| 186 | e.printStackTrace(); |
||
| 187 | } catch (NoSuchMethodException e) { |
||
| 188 | // TODO Auto-generated catch block |
||
| 189 | e.printStackTrace(); |
||
| 190 | } catch (SecurityException e) { |
||
| 191 | // TODO Auto-generated catch block |
||
| 192 | e.printStackTrace(); |
||
| 193 | } |
||
| 194 | |||
| 195 | return inputModel; |
||
| 196 | } |
||
| 197 | |||
| 198 | private Statement markGraphSemanticLevel(Model inputModel, String rulesFileName, String LOVFileName) { |
||
| 199 | // Variables initializations |
||
| 200 | ResIterator resourceSet = inputModel.listSubjects(); |
||
| 201 | Model innerModel = ModelFactory.createDefaultModel(); |
||
| 202 | Resource r = resourceSet.nextResource(); |
||
| 203 | Property p = ResourceFactory.createProperty("sstamp:hassemanticlevel"); |
||
| 204 | |||
| 205 | // Tive que criar um model para trabalhar com um Resource |
||
| 206 | Statement outputGraphSemanticLevel = innerModel.createStatement(r, p, "sstamp:notMarked"); |
||
| 207 | |||
| 208 | // Identify the levels of each statement on the inputGraph |
||
| 209 | StmtIterator statementSet = inputModel.listStatements(); |
||
| 210 | Integer valueLevel = 0; |
||
| 211 | |||
| 212 | while (statementSet.hasNext()) { |
||
| 213 | Statement s = statementSet.nextStatement(); |
||
| 214 | |||
| 215 | String semanticLevel = assessSemanticLevel(s, rulesFileName, LOVFileName); |
||
| 216 | |||
| 217 | if (valueLevel < assessedValueLevel) { |
||
| 218 | outputGraphSemanticLevel = innerModel.createStatement(r, p, semanticLevel); |
||
| 219 | valueLevel = assessedValueLevel; |
||
| 220 | |||
| 221 | } |
||
| 222 | |||
| 223 | } |
||
| 224 | return outputGraphSemanticLevel; |
||
| 225 | } |
||
| 226 | |||
| 227 | private static String assessSemanticLevel(Statement s, String rulesFileName, String LOVFileName) { |
||
| 228 | Interpreter i = new Interpreter(); |
||
| 229 | |||
| 230 | String assessedDescriptionLevel = "NotIdentified"; |
||
| 231 | |||
| 232 | // Is Literal? |
||
| 233 | // if (s.getLiteral() != null) |
||
| 234 | // assessedLevel="laiid:low"; |
||
| 235 | |||
| 236 | String inputSubject = s.getSubject().toString(); |
||
| 237 | String inputPredicate = s.getPredicate().toString(); |
||
| 238 | String inputObject = s.getObject().toString(); |
||
| 239 | |||
| 240 | try { |
||
| 241 | // Open LOV xml file |
||
| 242 | DocumentBuilderFactory docBuilderFactory2 = DocumentBuilderFactory.newInstance(); |
||
| 243 | DocumentBuilder docBuilder2 = docBuilderFactory2.newDocumentBuilder(); |
||
| 244 | Document doc2 = docBuilder2.parse(new File(LOVFileName.toString())); |
||
| 245 | |||
| 246 | // Open SemanticLevelFrameWork files |
||
| 247 | DocumentBuilderFactory docBuilderFactory = DocumentBuilderFactory.newInstance(); |
||
| 248 | DocumentBuilder docBuilder = docBuilderFactory.newDocumentBuilder(); |
||
| 249 | Document doc = docBuilder.parse(new File(rulesFileName)); |
||
| 250 | NodeList listOfFrames = doc.getElementsByTagName("Frame"); |
||
| 251 | int totalFrames = listOfFrames.getLength(); |
||
| 252 | |||
| 253 | // valida variaveis para o interpreter |
||
| 254 | i.set("inputSubject", inputSubject); |
||
| 255 | i.set("inputPredicate", inputPredicate); |
||
| 256 | i.set("inputObject", inputObject); |
||
| 257 | |||
| 258 | // Get Prefix |
||
| 259 | String prefixo = ""; |
||
| 260 | if (inputPredicate.contains(":")) { |
||
| 261 | int index = inputPredicate.indexOf(":"); |
||
| 262 | prefixo = inputPredicate.substring(0, index); |
||
| 263 | } |
||
| 264 | |||
| 265 | i.set("isVocabulary", isVocabulary(prefixo, doc2)); |
||
| 266 | i.set("isOntology", isOntology(prefixo, doc2)); |
||
| 267 | i.set("s", s); |
||
| 268 | |||
| 269 | // busca as regras |
||
| 270 | for (int k = 0; k < totalFrames; k++) { |
||
| 271 | Node ruleFrameNode = listOfFrames.item(k); |
||
| 272 | if (ruleFrameNode.getNodeType() == Node.ELEMENT_NODE) { |
||
| 273 | Element ruleFrameElement = (Element) ruleFrameNode; |
||
| 274 | NodeList ruleList = ruleFrameElement.getElementsByTagName("Rule"); |
||
| 275 | Element ruleElement = (Element) ruleList.item(0); |
||
| 276 | NodeList textRList = ruleElement.getChildNodes(); |
||
| 277 | NodeList levelValueList = ruleFrameElement.getElementsByTagName("LevelValue"); |
||
| 278 | Element levelValueElement = (Element) levelValueList.item(0); |
||
| 279 | NodeList levelDescriptionList = ruleFrameElement.getElementsByTagName("LevelDescription"); |
||
| 280 | Element levelDescriptionElement = (Element) levelDescriptionList.item(0); |
||
| 281 | |||
| 282 | NodeList textLValueList = levelValueElement.getChildNodes(); |
||
| 283 | NodeList textLDescriptionList = levelDescriptionElement.getChildNodes(); |
||
| 284 | |||
| 285 | // Rule Evaluation |
||
| 286 | if ((Boolean) i.eval(textRList.item(0).getNodeValue().trim())) { |
||
| 287 | assessedDescriptionLevel = textLDescriptionList.item(0).getNodeValue().trim(); |
||
| 288 | assessedValueLevel = Integer.valueOf(textLValueList.item(0).getNodeValue()); |
||
| 289 | |||
| 290 | // TODO avaliar sair do for |
||
| 291 | } |
||
| 292 | } |
||
| 293 | } |
||
| 294 | |||
| 295 | } catch (ParserConfigurationException e) { |
||
| 296 | // TODO Auto-generated catch block |
||
| 297 | e.printStackTrace(); |
||
| 298 | } catch (SAXException e) { |
||
| 299 | // TODO Auto-generated catch block |
||
| 300 | e.printStackTrace(); |
||
| 301 | } catch (IOException e) { |
||
| 302 | // TODO Auto-generated catch block |
||
| 303 | e.printStackTrace(); |
||
| 304 | } catch (EvalError e) { |
||
| 305 | // TODO Auto-generated catch block |
||
| 306 | e.printStackTrace(); |
||
| 307 | } |
||
| 308 | |||
| 309 | return assessedDescriptionLevel; |
||
| 310 | } |
||
| 311 | |||
| 312 | View Code Duplication | private static boolean isOntology(String prefix, Document doc) { |
|
| 313 | NodeList listOfResults = doc.getElementsByTagName("result"); |
||
| 314 | int totalResults = listOfResults.getLength(); |
||
| 315 | for (int k = 0; k < totalResults; k++) { |
||
| 316 | Node resultNode = listOfResults.item(k); |
||
| 317 | Element resultElement = (Element) resultNode; |
||
| 318 | NodeList bindingList = resultElement.getElementsByTagName("binding"); |
||
| 319 | Element bindingPrefixElement = (Element) bindingList.item(0); |
||
| 320 | NodeList literalList = bindingPrefixElement.getElementsByTagName("literal"); |
||
| 321 | Element literalElement = (Element) literalList.item(0); |
||
| 322 | NodeList textLiList = literalElement.getChildNodes(); |
||
| 323 | if (prefix.equals(textLiList.item(0).getNodeValue().trim())) { |
||
| 324 | Element bindingDescElement = (Element) bindingList.item(3); |
||
| 325 | NodeList literalDescList = bindingDescElement.getElementsByTagName("literal"); |
||
| 326 | Element literalDescElement = (Element) literalDescList.item(0); |
||
| 327 | NodeList textDescList = literalDescElement.getChildNodes(); |
||
| 328 | Element bindingTitleElement = (Element) bindingList.item(2); |
||
| 329 | NodeList literalTitleList = bindingTitleElement.getElementsByTagName("literal"); |
||
| 330 | Element literalTitleElement = (Element) literalTitleList.item(0); |
||
| 331 | NodeList textTitleList = literalTitleElement.getChildNodes(); |
||
| 332 | if (textDescList.item(0).getNodeValue().trim().toLowerCase().contains("ontology")) { |
||
| 333 | return true; |
||
| 334 | } |
||
| 335 | if (textTitleList.item(0).getNodeValue().trim().toLowerCase().contains("ontology")) { |
||
| 336 | return true; |
||
| 337 | } |
||
| 338 | } |
||
| 339 | } |
||
| 340 | return false; |
||
| 341 | } |
||
| 342 | |||
| 343 | View Code Duplication | private static boolean isVocabulary(String prefix, Document doc) { |
|
| 344 | NodeList listOfResults = doc.getElementsByTagName("result"); |
||
| 345 | int totalResults = listOfResults.getLength(); |
||
| 346 | for (int k = 0; k < totalResults; k++) { |
||
| 347 | Node resultNode = listOfResults.item(k); |
||
| 348 | Element resultElement = (Element) resultNode; |
||
| 349 | NodeList bindingList = resultElement.getElementsByTagName("binding"); |
||
| 350 | Element bindingPrefixElement = (Element) bindingList.item(0); |
||
| 351 | NodeList literalList = bindingPrefixElement.getElementsByTagName("literal"); |
||
| 352 | Element literalElement = (Element) literalList.item(0); |
||
| 353 | NodeList textLiList = literalElement.getChildNodes(); |
||
| 354 | if (prefix.equals(textLiList.item(0).getNodeValue().trim())) { |
||
| 355 | Element bindingDescElement = (Element) bindingList.item(3); |
||
| 356 | NodeList literalDescList = bindingDescElement.getElementsByTagName("literal"); |
||
| 357 | Element literalDescElement = (Element) literalDescList.item(0); |
||
| 358 | NodeList textDescList = literalDescElement.getChildNodes(); |
||
| 359 | Element bindingTitleElement = (Element) bindingList.item(2); |
||
| 360 | NodeList literalTitleList = bindingTitleElement.getElementsByTagName("literal"); |
||
| 361 | Element literalTitleElement = (Element) literalTitleList.item(0); |
||
| 362 | NodeList textTitleList = literalTitleElement.getChildNodes(); |
||
| 363 | if (textDescList.item(0).getNodeValue().trim().toLowerCase().contains("vocabulary")) { |
||
| 364 | return true; |
||
| 365 | } |
||
| 366 | if (textTitleList.item(0).getNodeValue().trim().toLowerCase().contains("vocabulary")) { |
||
| 367 | return true; |
||
| 368 | } |
||
| 369 | } |
||
| 370 | } |
||
| 371 | return false; |
||
| 372 | } |
||
| 375 |
Loading history...