Example usage for edu.stanford.nlp.io IOUtils getPrintWriter

Introduction

In this page you can find the example usage for edu.stanford.nlp.io IOUtils getPrintWriter.

Prototype

public static PrintWriter getPrintWriter(String filename) throws IOException 

Source Link

Usage

From source file:ilcc.ccgparser.nnparser.IncNNParser.java

public void writeModelFile(String modelFile) {
    try {//from  www.  j  a  v  a 2 s. c o m
        double[][] W1 = classifier.getW1();
        double[] b1 = classifier.getb1();
        double[][] W2 = classifier.getW2();
        double[][] E = classifier.getE();

        Writer output = IOUtils.getPrintWriter(modelFile);

        HashMap<String, ArrayList<CCGJRuleInfo>> uRules = srparser.treebankRules.getUnaryRules();
        HashMap<String, ArrayList<CCGJRuleInfo>> bRules = srparser.treebankRules.getBinaryRules();
        HashMap<String, ArrayList<CCGJRuleInfo>> rRules = srparser.treebankRules.getRevealRules();

        output.write("dict=" + knownWords.size() + "\n");
        output.write("pos=" + knownPos.size() + "\n");
        output.write("ccg cats=" + knownCCGCats.size() + "\n");
        output.write("embeddingSize=" + E[0].length + "\n");
        output.write("hiddenSize=" + b1.length + "\n");
        output.write("numTokens=" + (W1[0].length / E[0].length) + "\n");
        output.write("preComputed=" + preComputed.size() + "\n");
        output.write("classes=" + actsMap.size() + "\n");
        output.write("UnaryRules=" + uRules.size() + "\n");
        output.write("BinaryRules=" + bRules.size() + "\n");
        output.write("RevealRules=" + rRules.size() + "\n");
        int index = 0;

        // Classes
        for (ArcJAction act : actsList)
            output.write(act.toString() + "\n");

        // Unary and Binary Rules
        for (String key : uRules.keySet()) {
            ArrayList<CCGJRuleInfo> list = uRules.get(key);
            output.write(key);
            for (CCGJRuleInfo info : list) {
                output.write("  " + info.toString());
            }
            output.write("\n");
        }
        for (String key : bRules.keySet()) {
            ArrayList<CCGJRuleInfo> list = bRules.get(key);
            output.write(key);
            for (CCGJRuleInfo info : list) {
                output.write("  " + info.toString());
            }
            output.write("\n");
        }
        for (String key : rRules.keySet()) {
            ArrayList<CCGJRuleInfo> list = rRules.get(key);
            output.write(key);
            for (CCGJRuleInfo info : list) {
                output.write("  " + info.toString());
            }
            output.write("\n");
        }

        // First write word / POS / label embeddings
        for (String word : knownWords) {
            output.write(word);
            for (int k = 0; k < E[index].length; ++k)
                output.write(" " + E[index][k]);
            output.write("\n");
            index = index + 1;
        }
        for (String pos : knownPos) {
            output.write(pos);
            for (int k = 0; k < E[index].length; ++k)
                output.write(" " + E[index][k]);
            output.write("\n");
            index = index + 1;
        }
        for (String label : knownCCGCats) {
            output.write(label);
            for (int k = 0; k < E[index].length; ++k)
                output.write(" " + E[index][k]);
            output.write("\n");
            index = index + 1;
        }

        // Now write classifier weights
        for (int j = 0; j < W1[0].length; ++j)
            for (int i = 0; i < W1.length; ++i) {
                output.write("" + W1[i][j]);
                if (i == W1.length - 1)
                    output.write("\n");
                else
                    output.write(" ");
            }
        for (int i = 0; i < b1.length; ++i) {
            output.write("" + b1[i]);
            if (i == b1.length - 1)
                output.write("\n");
            else
                output.write(" ");
        }
        for (int j = 0; j < W2[0].length; ++j)
            for (int i = 0; i < W2.length; ++i) {
                output.write("" + W2[i][j]);
                if (i == W2.length - 1)
                    output.write("\n");
                else
                    output.write(" ");
            }

        // Finish with pre-computation info
        for (int i = 0; i < preComputed.size(); ++i) {
            output.write("" + preComputed.get(i));
            if ((i + 1) % 100 == 0 || i == preComputed.size() - 1)
                output.write("\n");
            else
                output.write(" ");
        }

        output.close();
    } catch (IOException e) {
        throw new RuntimeIOException(e);
    }
}

---

From source file:knu.univ.lingvo.coref.SieveCoreferenceSystem.java

License:Open Source License

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);
        }//  ww w .ja v  a  2s .c o  m
        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();
    }

    //
    // Parse one document at a time, and do single-doc coreference resolution in each.
    //
    // In one iteration, orderedMentionsBySentence contains a list of all
    // mentions in one document. Each mention has properties (annotations):
    // its surface form (Word), NER Tag, POS Tag, Index, etc.
    //

    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;
        }

        corefSystem.coref(document); // Do Coreference Resolution

        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);
                finalScore = printFinalConllScore(summary);
                if (false || corefSystem.optimizeConllScore) {
                    //finalScore = getFinalConllScore(summary, corefSystem.optimizeMetricType, corefSystem.optimizeSubScoreType.toString());
                }
            }
        }
    }

    if (true || !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;
}

---

From source file:knu.univ.lingvo.coref.SieveCoreferenceSystem.java

License:Open Source License

/**
 * Run and score coref distributed/*ww w.j a v a 2 s. co m*/
 */
public static void runAndScoreCorefDist(String runDistCmd, Properties props, String propsFile)
        throws Exception {
    PrintWriter pw = IOUtils.getPrintWriter(propsFile);
    props.store(pw, null);
    pw.close();
    /* Run coref job in a distributed manner, score is written to file */
    List<String> cmd = new ArrayList<String>();
    cmd.addAll(Arrays.asList(runDistCmd.split("\\s+")));
    cmd.add("-props");
    cmd.add(propsFile);
    ProcessBuilder pb = new ProcessBuilder(cmd);
    // Copy environment variables over
    Map<String, String> curEnv = System.getenv();
    Map<String, String> pbEnv = pb.environment();
    pbEnv.putAll(curEnv);

    logger.info("Running distributed coref:" + StringUtils.join(pb.command(), " "));
    StringWriter outSos = new StringWriter();
    StringWriter errSos = new StringWriter();
    PrintWriter out = new PrintWriter(new BufferedWriter(outSos));
    PrintWriter err = new PrintWriter(new BufferedWriter(errSos));
    SystemUtils.run(pb, out, err);
    out.close();
    err.close();
    String outStr = outSos.toString();
    String errStr = errSos.toString();
    logger.info("Finished distributed coref: " + runDistCmd + ", props=" + propsFile);
    logger.info("Output: " + outStr);
    if (errStr.length() > 0) {
        logger.info("Error: " + errStr);
    }
}

---