Search in sources :

Example 16 with UsageException

use of com.oracle.labs.mlrg.olcut.config.UsageException in project tribuo by oracle.

the class OCIModelCLI method main.

/**
 * Entry point for the OCIModelCLI.
 * @param args The CLI arguments.
 * @throws IOException If the model or dataset failed to load.
 * @throws ClassNotFoundException If the model or dataset had an unknown class.
 */
public static void main(String[] args) throws IOException, ClassNotFoundException {
    OCIModelOptions o = new OCIModelOptions();
    ConfigurationManager cm;
    try {
        cm = new ConfigurationManager(args, o, false);
    } catch (UsageException e) {
        logger.info(e.getMessage());
        return;
    }
    switch(o.mode) {
        case CREATE_AND_DEPLOY:
            createModelAndDeploy(o);
            break;
        case DEPLOY:
            deploy(o);
            break;
        case SCORE:
            modelScoring(o);
            break;
    }
}
Also used : UsageException(com.oracle.labs.mlrg.olcut.config.UsageException) ConfigurationManager(com.oracle.labs.mlrg.olcut.config.ConfigurationManager)

Example 17 with UsageException

use of com.oracle.labs.mlrg.olcut.config.UsageException in project tribuo by oracle.

the class TrainTest method main.

/**
 * Runs a TrainTest CLI.
 * @param args the command line arguments
 * @throws IOException if there is any error reading the examples.
 */
public static void main(String[] args) throws IOException {
    // 
    // Use the labs format logging.
    LabsLogFormatter.setAllLogFormatters();
    KMeansOptions o = new KMeansOptions();
    ConfigurationManager cm;
    try {
        cm = new ConfigurationManager(args, o);
    } catch (UsageException e) {
        logger.info(e.getMessage());
        return;
    }
    if (o.general.trainingPath == null) {
        logger.info(cm.usage());
        return;
    }
    ClusteringFactory factory = new ClusteringFactory();
    Pair<Dataset<ClusterID>, Dataset<ClusterID>> data = o.general.load(factory);
    Dataset<ClusterID> train = data.getA();
    // public KMeansTrainer(int centroids, int iterations, DistanceType distType, int numThreads, int seed)
    KMeansTrainer trainer = new KMeansTrainer(o.centroids, o.iterations, o.distType, o.initialisation, o.numThreads, o.general.seed);
    Model<ClusterID> model = trainer.train(train);
    logger.info("Finished training model");
    ClusteringEvaluation evaluation = factory.getEvaluator().evaluate(model, train);
    logger.info("Finished evaluating model");
    System.out.println("Normalized MI = " + evaluation.normalizedMI());
    System.out.println("Adjusted MI = " + evaluation.adjustedMI());
    if (o.general.outputPath != null) {
        o.general.saveModel(model);
    }
}
Also used : UsageException(com.oracle.labs.mlrg.olcut.config.UsageException) ClusterID(org.tribuo.clustering.ClusterID) Dataset(org.tribuo.Dataset) ClusteringFactory(org.tribuo.clustering.ClusteringFactory) ConfigurationManager(com.oracle.labs.mlrg.olcut.config.ConfigurationManager) ClusteringEvaluation(org.tribuo.clustering.evaluation.ClusteringEvaluation)

Example 18 with UsageException

use of com.oracle.labs.mlrg.olcut.config.UsageException in project tribuo by oracle.

the class ConfigurableTrainTest method main.

/**
 * @param args the command line arguments
 * @param <T> The {@link Output} subclass.
 */
@SuppressWarnings("unchecked")
public static <T extends Output<T>> void main(String[] args) {
    // 
    // Use the labs format logging.
    LabsLogFormatter.setAllLogFormatters();
    ConfigurableTrainTestOptions o = new ConfigurableTrainTestOptions();
    ConfigurationManager cm;
    try {
        cm = new ConfigurationManager(args, o);
    } catch (UsageException e) {
        logger.info(e.getMessage());
        return;
    }
    if (o.general.trainingPath == null || o.general.testingPath == null || o.outputFactory == null) {
        logger.info(cm.usage());
        System.exit(1);
    }
    Pair<Dataset<T>, Dataset<T>> data = null;
    try {
        data = o.general.load((OutputFactory<T>) o.outputFactory);
    } catch (IOException e) {
        logger.log(Level.SEVERE, "Failed to load data", e);
        System.exit(1);
    }
    Dataset<T> train = data.getA();
    Dataset<T> test = data.getB();
    if (o.trainer == null) {
        logger.warning("No trainer supplied");
        logger.info(cm.usage());
        System.exit(1);
    }
    if (o.transformationMap != null) {
        o.trainer = new TransformTrainer<>(o.trainer, o.transformationMap);
    }
    logger.info("Trainer is " + o.trainer.getProvenance().toString());
    logger.info("Outputs are " + train.getOutputInfo().toReadableString());
    logger.info("Number of features: " + train.getFeatureMap().size());
    final long trainStart = System.currentTimeMillis();
    Model<T> model = ((Trainer<T>) o.trainer).train(train);
    final long trainStop = System.currentTimeMillis();
    logger.info("Finished training classifier " + Util.formatDuration(trainStart, trainStop));
    Evaluator<T, ? extends Evaluation<T>> evaluator = train.getOutputFactory().getEvaluator();
    final long testStart = System.currentTimeMillis();
    Evaluation<T> evaluation = evaluator.evaluate(model, test);
    final long testStop = System.currentTimeMillis();
    logger.info("Finished evaluating model " + Util.formatDuration(testStart, testStop));
    System.out.println(evaluation.toString());
    if (o.general.outputPath != null) {
        try {
            o.general.saveModel(model);
        } catch (IOException e) {
            logger.log(Level.SEVERE, "Error writing model", e);
        }
    }
    if (o.crossValidation) {
        if (o.numFolds > 1) {
            logger.info("Running " + o.numFolds + " fold cross-validation");
            CrossValidation<T, ? extends Evaluation<T>> cv = new CrossValidation<>((Trainer<T>) o.trainer, train, evaluator, o.numFolds, o.general.seed);
            List<? extends Pair<? extends Evaluation<T>, Model<T>>> evaluations = cv.evaluate();
            List<Evaluation<T>> evals = evaluations.stream().map(Pair::getA).collect(Collectors.toList());
            // Summarize across everything
            Map<MetricID<T>, DescriptiveStats> summary = EvaluationAggregator.summarize(evals);
            List<MetricID<T>> keys = new ArrayList<>(summary.keySet()).stream().sorted(Comparator.comparing(Pair::getB)).collect(Collectors.toList());
            System.out.println("Summary across the folds:");
            for (MetricID<T> key : keys) {
                DescriptiveStats stats = summary.get(key);
                System.out.printf("%-10s  %.5f (%.5f)%n", key, stats.getMean(), stats.getStandardDeviation());
            }
        } else {
            logger.warning("The number of cross-validation folds must be greater than 1, found " + o.numFolds);
        }
    }
}
Also used : UsageException(com.oracle.labs.mlrg.olcut.config.UsageException) TransformTrainer(org.tribuo.transform.TransformTrainer) Trainer(org.tribuo.Trainer) MetricID(org.tribuo.evaluation.metrics.MetricID) DescriptiveStats(org.tribuo.evaluation.DescriptiveStats) ConfigurationManager(com.oracle.labs.mlrg.olcut.config.ConfigurationManager) Pair(com.oracle.labs.mlrg.olcut.util.Pair) Evaluation(org.tribuo.evaluation.Evaluation) Dataset(org.tribuo.Dataset) IOException(java.io.IOException) Model(org.tribuo.Model) CrossValidation(org.tribuo.evaluation.CrossValidation) OutputFactory(org.tribuo.OutputFactory)

Example 19 with UsageException

use of com.oracle.labs.mlrg.olcut.config.UsageException in project tribuo by oracle.

the class SQLToCSV method main.

/**
 * Reads an SQL query from the standard input and writes the results of the
 * query to the standard output.
 *
 * @param args Single arg is the JDBC connection string.
 */
public static void main(String[] args) {
    LabsLogFormatter.setAllLogFormatters();
    SQLToCSVOptions opts = new SQLToCSVOptions();
    ConfigurationManager cm;
    try {
        cm = new ConfigurationManager(args, opts);
    } catch (UsageException e) {
        logger.info(e.getUsage());
        System.exit(1);
    }
    if (opts.dbConfig == null) {
        if (opts.connString == null) {
            logger.log(Level.SEVERE, "Must specify connection string with -n");
            System.exit(1);
        }
        if (opts.username != null || opts.password != null) {
            if (opts.username == null || opts.password == null) {
                logger.log(Level.SEVERE, "Must specify both of user and password with -u, -p if one is specified!");
                System.exit(1);
            }
        }
    } else if (opts.username != null || opts.password != null || opts.connString != null) {
        logger.warning("dbConfig provided but username/password/connstring also provided. Options from -u, -p, -n being ignored");
    }
    String query;
    try (BufferedReader br = opts.inputPath != null ? Files.newBufferedReader(opts.inputPath) : new BufferedReader(new InputStreamReader(System.in, StandardCharsets.UTF_8))) {
        StringBuilder qsb = new StringBuilder();
        String l;
        while ((l = br.readLine()) != null) {
            qsb.append(l);
            qsb.append("\n");
        }
        query = qsb.toString().trim();
    } catch (IOException ex) {
        logger.log(Level.SEVERE, "Error reading query: " + ex);
        System.exit(1);
        return;
    }
    if (query.isEmpty()) {
        logger.log(Level.SEVERE, "Query is empty string");
        System.exit(1);
    }
    Connection conn = null;
    try {
        if (opts.dbConfig != null) {
            conn = opts.dbConfig.getConnection();
        } else if (opts.username != null) {
            conn = DriverManager.getConnection(opts.connString, opts.username, opts.password);
        } else {
            conn = DriverManager.getConnection(opts.connString);
        }
    } catch (SQLException ex) {
        logger.log(Level.SEVERE, "Can't connect to database: " + opts.connString, ex);
        System.exit(1);
    }
    try (Statement stmt = conn.createStatement()) {
        stmt.setFetchSize(1000);
        stmt.setFetchDirection(ResultSet.FETCH_FORWARD);
        ResultSet results;
        try {
            results = stmt.executeQuery(query);
        } catch (SQLException ex) {
            logger.log(Level.SEVERE, "Error running query", ex);
            try {
                conn.close();
            } catch (SQLException ex1) {
                logger.log(Level.SEVERE, "Failed to close connection", ex1);
            }
            return;
        }
        try (ICSVWriter writer = new CSVParserWriter(opts.outputPath != null ? Files.newBufferedWriter(opts.outputPath) : new BufferedWriter(new OutputStreamWriter(System.out, StandardCharsets.UTF_8), 1024 * 1024), new RFC4180Parser(), "\n")) {
            writer.writeAll(results, true);
        } catch (IOException ex) {
            logger.log(Level.SEVERE, "Error writing CSV", ex);
            System.exit(1);
        } catch (SQLException ex) {
            logger.log(Level.SEVERE, "Error retrieving results", ex);
            System.exit(1);
        }
    } catch (SQLException ex) {
        logger.log(Level.SEVERE, "Couldn't create statement", ex);
        try {
            conn.close();
        } catch (SQLException ex1) {
            logger.log(Level.SEVERE, "Failed to close connection", ex1);
        }
        System.exit(1);
        return;
    }
    try {
        conn.close();
    } catch (SQLException ex1) {
        logger.log(Level.SEVERE, "Failed to close connection", ex1);
    }
}
Also used : UsageException(com.oracle.labs.mlrg.olcut.config.UsageException) InputStreamReader(java.io.InputStreamReader) RFC4180Parser(com.opencsv.RFC4180Parser) SQLException(java.sql.SQLException) Statement(java.sql.Statement) Connection(java.sql.Connection) IOException(java.io.IOException) BufferedWriter(java.io.BufferedWriter) ICSVWriter(com.opencsv.ICSVWriter) BufferedReader(java.io.BufferedReader) ResultSet(java.sql.ResultSet) CSVParserWriter(com.opencsv.CSVParserWriter) OutputStreamWriter(java.io.OutputStreamWriter) ConfigurationManager(com.oracle.labs.mlrg.olcut.config.ConfigurationManager)

Example 20 with UsageException

use of com.oracle.labs.mlrg.olcut.config.UsageException in project tribuo by oracle.

the class PreprocessAndSerialize method main.

/**
 * Run the PreprocessAndSerialize CLI.
 * @param args The CLI args.
 */
public static void main(String[] args) {
    LabsLogFormatter.setAllLogFormatters();
    PreprocessAndSerializeOptions opts = new PreprocessAndSerializeOptions();
    ConfigurationManager cm;
    try {
        cm = new ConfigurationManager(args, opts);
    } catch (UsageException e) {
        logger.info(e.getUsage());
        System.exit(1);
    }
    logger.info("Reading datasource into dataset");
    MutableDataset<?> dataset = new MutableDataset<>(opts.dataSource);
    logger.info("Finished reading dataset");
    if (opts.output.endsWith("gz")) {
        logger.info("Writing zipped dataset");
    }
    try (ObjectOutputStream os = IOUtil.getObjectOutputStream(opts.output.toString(), opts.output.endsWith("gz"))) {
        os.writeObject(dataset);
    } catch (IOException e) {
        logger.log(Level.SEVERE, "Error writing serialized dataset", e);
        System.exit(1);
    }
}
Also used : UsageException(com.oracle.labs.mlrg.olcut.config.UsageException) IOException(java.io.IOException) ObjectOutputStream(java.io.ObjectOutputStream) ConfigurationManager(com.oracle.labs.mlrg.olcut.config.ConfigurationManager) MutableDataset(org.tribuo.MutableDataset)

Aggregations

ConfigurationManager (com.oracle.labs.mlrg.olcut.config.ConfigurationManager)32 UsageException (com.oracle.labs.mlrg.olcut.config.UsageException)32 Label (org.tribuo.classification.Label)15 Dataset (org.tribuo.Dataset)14 IOException (java.io.IOException)8 ObjectOutputStream (java.io.ObjectOutputStream)7 RegressionFactory (org.tribuo.regression.RegressionFactory)7 Regressor (org.tribuo.regression.Regressor)7 RegressionEvaluation (org.tribuo.regression.evaluation.RegressionEvaluation)7 FileOutputStream (java.io.FileOutputStream)6 MutableDataset (org.tribuo.MutableDataset)4 LabelEvaluation (org.tribuo.classification.evaluation.LabelEvaluation)4 LabelEvaluator (org.tribuo.classification.evaluation.LabelEvaluator)4 BufferedWriter (java.io.BufferedWriter)3 ObjectInputStream (java.io.ObjectInputStream)3 OutputStreamWriter (java.io.OutputStreamWriter)3 Model (org.tribuo.Model)3 LabelFactory (org.tribuo.classification.LabelFactory)3 ArgumentException (com.oracle.labs.mlrg.olcut.config.ArgumentException)2 BufferedInputStream (java.io.BufferedInputStream)2