Examples with PrintWriter java.io.PrintWriter used on opensource projects

Example 66 with PrintWriter

use of java.io.PrintWriter in project CoreNLP by stanfordnlp.

the class PairwiseModel method writeModel.

public void writeModel(String outputPath) throws Exception {
    File outDir = new File(outputPath);
    if (!outDir.exists()) {
        outDir.mkdir();
    }
    try (PrintWriter writer = new PrintWriter(outputPath + "config", "UTF-8")) {
        writer.print(str);
    }
    try (PrintWriter writer = new PrintWriter(outputPath + "/weights", "UTF-8")) {
        classifier.printWeightVector(writer);
    }
    classifier.writeWeights(outputPath + "/model.ser");
}

Example 67 with PrintWriter

use of java.io.PrintWriter in project CoreNLP by stanfordnlp.

the class PairwiseModelTrainer method test.

public static void test(PairwiseModel model, String predictionsName, boolean anaphoricityModel) throws Exception {
    Redwood.log("scoref-train", "Reading compression...");
    Compressor<String> compressor = IOUtils.readObjectFromFile(StatisticalCorefTrainer.compressorFile);
    Redwood.log("scoref-train", "Reading test data...");
    List<DocumentExamples> testDocuments = IOUtils.readObjectFromFile(StatisticalCorefTrainer.extractedFeaturesFile);
    Redwood.log("scoref-train", "Building test set...");
    List<Pair<Example, Map<Integer, CompressedFeatureVector>>> allExamples = anaphoricityModel ? getAnaphoricityExamples(testDocuments) : getExamples(testDocuments);
    Redwood.log("scoref-train", "Testing...");
    PrintWriter writer = new PrintWriter(model.getDefaultOutputPath() + predictionsName);
    Map<Integer, Counter<Pair<Integer, Integer>>> scores = new HashMap<>();
    writeScores(allExamples, compressor, model, writer, scores);
    if (model instanceof MaxMarginMentionRanker) {
        writer.close();
        writer = new PrintWriter(model.getDefaultOutputPath() + predictionsName + "_anaphoricity");
        testDocuments = IOUtils.readObjectFromFile(StatisticalCorefTrainer.extractedFeaturesFile);
        allExamples = getAnaphoricityExamples(testDocuments);
        writeScores(allExamples, compressor, model, writer, scores);
    }
    IOUtils.writeObjectToFile(scores, model.getDefaultOutputPath() + predictionsName + ".ser");
    writer.close();
}

Example 68 with PrintWriter

use of java.io.PrintWriter in project CoreNLP by stanfordnlp.

the class SieveCoreferenceSystem method runAndScoreCoref.

public static double runAndScoreCoref(SieveCoreferenceSystem corefSystem, MentionExtractor mentionExtractor, Properties props, String timeStamp) throws Exception {
    // prepare conll output
    PrintWriter writerGold = null;
    PrintWriter writerPredicted = null;
    PrintWriter writerPredictedCoref = null;
    String conllOutputMentionGoldFile = null;
    String conllOutputMentionPredictedFile = null;
    String conllOutputMentionCorefPredictedFile = null;
    String conllMentionEvalFile = null;
    String conllMentionEvalErrFile = null;
    String conllMentionCorefEvalFile = null;
    String conllMentionCorefEvalErrFile = null;
    if (Constants.PRINT_CONLL_OUTPUT || corefSystem.replicateCoNLL) {
        String conllOutput = props.getProperty(Constants.CONLL_OUTPUT_PROP, "conlloutput");
        conllOutputMentionGoldFile = conllOutput + "-" + timeStamp + ".gold.txt";
        conllOutputMentionPredictedFile = conllOutput + "-" + timeStamp + ".predicted.txt";
        conllOutputMentionCorefPredictedFile = conllOutput + "-" + timeStamp + ".coref.predicted.txt";
        conllMentionEvalFile = conllOutput + "-" + timeStamp + ".eval.txt";
        conllMentionEvalErrFile = conllOutput + "-" + timeStamp + ".eval.err.txt";
        conllMentionCorefEvalFile = conllOutput + "-" + timeStamp + ".coref.eval.txt";
        conllMentionCorefEvalErrFile = conllOutput + "-" + timeStamp + ".coref.eval.err.txt";
        logger.info("CONLL MENTION GOLD FILE: " + conllOutputMentionGoldFile);
        logger.info("CONLL MENTION PREDICTED FILE: " + conllOutputMentionPredictedFile);
        logger.info("CONLL MENTION EVAL FILE: " + conllMentionEvalFile);
        if (!Constants.SKIP_COREF) {
            logger.info("CONLL MENTION PREDICTED WITH COREF FILE: " + conllOutputMentionCorefPredictedFile);
            logger.info("CONLL MENTION WITH COREF EVAL FILE: " + conllMentionCorefEvalFile);
        }
        writerGold = new PrintWriter(new FileOutputStream(conllOutputMentionGoldFile));
        writerPredicted = new PrintWriter(new FileOutputStream(conllOutputMentionPredictedFile));
        writerPredictedCoref = new PrintWriter(new FileOutputStream(conllOutputMentionCorefPredictedFile));
    }
    mentionExtractor.resetDocs();
    if (corefSystem.doScore()) {
        corefSystem.initScorers();
    }
    while (true) {
        Document document = mentionExtractor.nextDoc();
        if (document == null)
            break;
        if (!props.containsKey(Constants.MUC_PROP)) {
            printRawDoc(document, true);
            printRawDoc(document, false);
        }
        printDiscourseStructure(document);
        if (corefSystem.doScore()) {
            document.extractGoldCorefClusters();
        }
        if (Constants.PRINT_CONLL_OUTPUT || corefSystem.replicateCoNLL) {
            // Not doing coref - print conll output here
            printConllOutput(document, writerGold, true);
            printConllOutput(document, writerPredicted, false);
        }
        // run mention detection only
        if (Constants.SKIP_COREF) {
            continue;
        }
        // Do Coreference Resolution
        corefSystem.coref(document);
        if (corefSystem.doScore()) {
            //Identifying possible coreferring mentions in the corpus along with any recall/precision errors with gold corpus
            corefSystem.printTopK(logger, document, corefSystem.semantics);
            logger.fine("pairwise score for this doc: ");
            corefSystem.scoreSingleDoc.get(corefSystem.sieves.length - 1).printF1(logger);
            logger.fine("accumulated score: ");
            corefSystem.printF1(true);
            logger.fine("\n");
        }
        if (Constants.PRINT_CONLL_OUTPUT || corefSystem.replicateCoNLL) {
            printConllOutput(document, writerPredictedCoref, false, true);
        }
    }
    double finalScore = 0;
    if (Constants.PRINT_CONLL_OUTPUT || corefSystem.replicateCoNLL) {
        writerGold.close();
        writerPredicted.close();
        writerPredictedCoref.close();
        //if(props.containsKey(Constants.CONLL_SCORER)) {
        if (corefSystem.conllMentionEvalScript != null) {
            //        runConllEval(corefSystem.conllMentionEvalScript, conllOutputMentionGoldFile, conllOutputMentionPredictedFile, conllMentionEvalFile, conllMentionEvalErrFile);
            String summary = getConllEvalSummary(corefSystem.conllMentionEvalScript, conllOutputMentionGoldFile, conllOutputMentionPredictedFile);
            logger.info("\nCONLL EVAL SUMMARY (Before COREF)");
            printScoreSummary(summary, logger, false);
            if (!Constants.SKIP_COREF) {
                //          runConllEval(corefSystem.conllMentionEvalScript, conllOutputMentionGoldFile, conllOutputMentionCorefPredictedFile, conllMentionCorefEvalFile, conllMentionCorefEvalErrFile);
                summary = getConllEvalSummary(corefSystem.conllMentionEvalScript, conllOutputMentionGoldFile, conllOutputMentionCorefPredictedFile);
                logger.info("\nCONLL EVAL SUMMARY (After COREF)");
                printScoreSummary(summary, logger, true);
                printFinalConllScore(summary);
                if (corefSystem.optimizeConllScore) {
                    finalScore = getFinalConllScore(summary, corefSystem.optimizeMetricType, corefSystem.optimizeSubScoreType.toString());
                }
            }
        }
    }
    if (!corefSystem.optimizeConllScore && corefSystem.doScore()) {
        finalScore = corefSystem.getFinalScore(corefSystem.optimizeMetricType, corefSystem.optimizeSubScoreType);
    }
    String scoresFile = props.getProperty(Constants.SCORE_FILE_PROP);
    if (scoresFile != null) {
        PrintWriter pw = IOUtils.getPrintWriter(scoresFile);
        pw.println((new DecimalFormat("#.##")).format(finalScore));
        pw.close();
    }
    if (corefSystem.optimizeSieves) {
        logger.info("Final reported score for sieve optimization " + corefSystem.optimizeScoreType + " : " + finalScore);
    }
    return finalScore;
}

Example 69 with PrintWriter

use of java.io.PrintWriter in project CoreNLP by stanfordnlp.

the class CRFFeatureExporter method printFeatures.

/**
   * Output features that have already been converted into features
   *  (using documentToDataAndLabels) in format suitable for CRFSuite.
   * Format is with one line per token using the following format
   * label feat1 feat2 ...
   *  (where each space is actually a tab)
   * Each document is separated by an empty line.
   *
   * @param exportFile file to export the features to
   * @param docsData array of document features
   * @param labels correct labels indexed by document, and position within document
   */
public void printFeatures(String exportFile, int[][][][] docsData, int[][] labels) {
    try {
        PrintWriter pw = IOUtils.getPrintWriter(exportFile);
        for (int i = 0; i < docsData.length; i++) {
            for (int j = 0; j < docsData[i].length; j++) {
                StringBuilder sb = new StringBuilder();
                int label = labels[i][j];
                sb.append(classifier.classIndex.get(label));
                for (int k = 0; k < docsData[i][j].length; k++) {
                    for (int m = 0; m < docsData[i][j][k].length; m++) {
                        String feat = classifier.featureIndex.get(docsData[i][j][k][m]);
                        feat = ubPrefixFeatureString(feat);
                        sb.append(delimiter);
                        sb.append(feat);
                    }
                }
                pw.println(sb.toString());
            }
            pw.println();
        }
        pw.close();
    } catch (IOException ex) {
        throw new RuntimeException(ex);
    }
}

Example 70 with PrintWriter

use of java.io.PrintWriter in project CoreNLP by stanfordnlp.

the class CRFFeatureExporter method printFeatures.

/**
   * Output features from a collection of documents to a file
   * Format is with one line per token using the following format
   * word label feat1 feat2 ...
   *  (where each space is actually a tab)
   * Each document is separated by an empty line
   * This format is suitable for modified crfsgd.
   *
   * @param exportFile file to export the features to
   * @param documents input collection of documents
   */
public void printFeatures(String exportFile, Collection<List<IN>> documents) {
    try {
        PrintWriter pw = IOUtils.getPrintWriter(exportFile);
        for (List<IN> doc : documents) {
            String str = getFeatureString(doc);
            pw.println(str);
        }
        pw.close();
    } catch (IOException ex) {
        throw new RuntimeException(ex);
    }
}