List of usage examples for edu.stanford.nlp.util StringUtils printToFile
public static void printToFile(String filename, String message, boolean append)
From source file:opennlp.tools.parse_thicket.kernel_interface.MultiSentenceExtendedForestSearchResultsProcessorSetFormer.java
License:Apache License
private List<HitBase> formTreeForestDataSet(List<HitBase> hits, String query, boolean isPositive) { List<HitBase> newHitList = new ArrayList<HitBase>(), newHitListReRanked = new ArrayList<HitBase>(); // form the training set from original documents. Since search results // are ranked, we set the first half as positive set, // and the second half as negative set. // after re-classification, being re-ranked, the search results might // end up in a different set List<String[]> treeBankBuffer = new ArrayList<String[]>(); int count = 0; for (HitBase hit : hits) { count++;/*w w w . ja v a 2 s. c o m*/ // if orig content has been already set in HIT object, ok; otherwise // set it String searchResultText = hit.getPageContent(); if (searchResultText == null) { try { HitBase hitWithFullSents = extractor.formTextFromOriginalPageGivenSnippet(hit); for (String paragraph : hitWithFullSents.getOriginalSentences()) { List<String[]> res = formTreeKernelStructure(paragraph, count, hits, isPositive); for (String[] rl : res) { StringUtils.printToFile(new File(path + trainingFileName), rl[0] + " \n", true); } // treeBankBuffer.addAll(res); } } catch (Exception e) { e.printStackTrace(); } } newHitList.add(hit); } // write the lits of samples to a file ProfileReaderWriter.appendReport(treeBankBuffer, path + trainingFileName, ' '); return newHitList; }