Examples with DerivedField org.dmg.pmml.DerivedFieldDocument.DerivedField used on opensource projects

Example 26 with DerivedField

use of org.dmg.pmml.DerivedFieldDocument.DerivedField in project knime-core by knime.

the class PMMLStringConversionTranslator method createDerivedFields.

private DerivedField[] createDerivedFields() {
    DATATYPE.Enum dataType = PMMLDataDictionaryTranslator.getPMMLDataType(m_parseType);
    OPTYPE.Enum optype = PMMLDataDictionaryTranslator.getOptype(m_parseType);
    int num = m_includeCols.size();
    DerivedField[] derivedFields = new DerivedField[num];
    for (int i = 0; i < num; i++) {
        DerivedField df = DerivedField.Factory.newInstance();
        String name = m_includeCols.get(i);
        df.setDisplayName(name);
        /* The field name must be retrieved before creating a new derived
             * name for this derived field as the map only contains the
             * current mapping. */
        String fieldName = m_mapper.getDerivedFieldName(name);
        df.setName(m_mapper.createDerivedFieldName(name));
        df.setDataType(dataType);
        df.setOptype(optype);
        FieldRef fieldRef = df.addNewFieldRef();
        fieldRef.setField(fieldName);
        derivedFields[i] = df;
    }
    return derivedFields;
}

Example 27 with DerivedField

use of org.dmg.pmml.DerivedFieldDocument.DerivedField in project knime-core by knime.

the class PMMLStringConversionTranslator method initializeFrom.

/**
 * {@inheritDoc}
 */
@Override
@SuppressWarnings("unchecked")
public List<Integer> initializeFrom(final DerivedField[] derivedFields) {
    if (derivedFields == null) {
        return Collections.EMPTY_LIST;
    }
    int num = derivedFields.length;
    List<Integer> consumed = new ArrayList<Integer>(num);
    for (int i = 0; i < derivedFields.length; i++) {
        DerivedField df = derivedFields[i];
        /**
         * This field contains the name of the column in KNIME that
         * corresponds to the derived field in PMML. This is necessary if
         * derived fields are defined on other derived fields and the
         * columns in KNIME are replaced with the preprocessed values.
         * In this case KNIME has to know the original names (e.g. A) while
         * PMML references to A*, A** etc.
         */
        String displayName = df.getDisplayName();
        if (!df.isSetFieldRef()) {
            // only reading field references
            continue;
        }
        DataType dataType = PMMLDataDictionaryTranslator.getKNIMEDataType(df.getDataType());
        if (dataType.isCompatible(IntValue.class)) {
            m_parseType = IntCell.TYPE;
        } else if (dataType.isCompatible(DoubleValue.class)) {
            m_parseType = DoubleCell.TYPE;
        } else if (dataType == StringCell.TYPE) {
            m_parseType = StringCell.TYPE;
        } else {
            // only processing int, double and string conversions
            continue;
        }
        FieldRef fieldRef = df.getFieldRef();
        if (displayName != null) {
            m_includeCols.add(displayName);
        } else {
            m_includeCols.add(m_mapper.getColumnName(fieldRef.getField()));
        }
        consumed.add(i);
    }
    return consumed;
}

Example 28 with DerivedField

use of org.dmg.pmml.DerivedFieldDocument.DerivedField in project knime-core by knime.

the class CategoryToNumberApplyNodeModel method getMapValues.

/**
 * @param model the PMML model
 * @return the field in the first FieldColumnPair of the MapValues mapped
 * to the MapValues Model
 */
private Map<String, DerivedField> getMapValues(final PMMLPortObject model) {
    Map<String, DerivedField> mapValues = new LinkedHashMap<String, DerivedField>();
    DerivedField[] derivedFields = model.getDerivedFields();
    for (DerivedField derivedField : derivedFields) {
        MapValues map = derivedField.getMapValues();
        if (null != map) {
            // This is the field name the mapValues is based on
            String name = map.getFieldColumnPairArray()[0].getField();
            mapValues.put(name, derivedField);
        }
    }
    return mapValues;
}

Example 29 with DerivedField

use of org.dmg.pmml.DerivedFieldDocument.DerivedField in project knime-core by knime.

the class NumberToCategoryApplyNodeModel method createRearranger.

/**
 * Creates a rearranger that processes the derived fields with MapValues
 * in the given model.
 */
private ColumnRearranger createRearranger(final DataTableSpec spec, final PMMLPortObject model) {
    // Retrieve columns with string data in the spec
    Set<String> intCols = new LinkedHashSet<String>();
    Set<String> otherCols = new LinkedHashSet<String>();
    for (DataColumnSpec colSpec : spec) {
        if (colSpec.getType().isCompatible(IntValue.class)) {
            intCols.add(colSpec.getName());
        } else {
            otherCols.add(colSpec.getName());
        }
    }
    if (intCols.isEmpty()) {
        if (null == model) {
            // during configure
            setWarningMessage("No columns to process.");
        } else {
            // during execute
            setWarningMessage("No columns to process, returning input.");
        }
    }
    // The map values in the model if present
    Map<String, DerivedField> mapValues = null != model ? getMapValues(model) : Collections.<String, DerivedField>emptyMap();
    // Create rearranger
    ColumnRearranger rearranger = new ColumnRearranger(spec);
    for (String col : mapValues.keySet()) {
        DerivedField derivedField = mapValues.get(col);
        MapValues map = derivedField.getMapValues();
        // this PMML MapValues model is found but has wrong type.
        if (null != model && otherCols.contains(col)) {
            String outColumn = null == derivedField.getDisplayName() || derivedField.getDisplayName().trim().isEmpty() ? derivedField.getName() : derivedField.getDisplayName();
            LOGGER.warn("Cannot create column \"" + outColumn + "\" since the input column \"" + col + "\" is not of type IntValue.");
            continue;
        }
        // this PMML MapValues model.
        if (null != model && !intCols.contains(col)) {
            String outColumn = derivedField.getMapValues().getFieldColumnPairList().get(0).getField();
            LOGGER.warn("Cannot create column \"" + outColumn + "\" since the column \"" + col + "\" is not in the input.");
            continue;
        }
        NumberToCategoryApplyCellFactory factory = new NumberToCategoryApplyCellFactory(spec, col, m_settings, map);
        if (m_settings.getAppendColumns()) {
            rearranger.append(factory);
        } else {
            rearranger.replace(factory, col);
        }
    }
    return rearranger;
}

Example 30 with DerivedField

use of org.dmg.pmml.DerivedFieldDocument.DerivedField in project knime-core by knime.

the class PMMLNeuralNetworkTranslator method addOutputLayer.

/**
 * Writes the PMML output layer of the MLP.
 *
 * @param nnModel
 *            the neural network model.
 * @param mlp
 *            the underlying {@link MultiLayerPerceptron}.
 * @param spec
 *            the port object spec
 */
protected void addOutputLayer(final NeuralNetwork nnModel, final MultiLayerPerceptron mlp, final PMMLPortObjectSpec spec) {
    int lastlayer = mlp.getNrLayers() - 1;
    String targetCol = spec.getTargetFields().iterator().next();
    Layer outputlayer = mlp.getLayer(lastlayer);
    Perceptron[] outputperceptrons = outputlayer.getPerceptrons();
    HashMap<DataCell, Integer> outputmap = mlp.getClassMapping();
    NeuralOutputs neuralOuts = nnModel.addNewNeuralOutputs();
    neuralOuts.setNumberOfOutputs(BigInteger.valueOf(outputperceptrons.length));
    for (int i = 0; i < outputperceptrons.length; i++) {
        NeuralOutput neuralOutput = neuralOuts.addNewNeuralOutput();
        neuralOutput.setOutputNeuron(lastlayer + "," + i);
        // search corresponding output value
        String colname = "";
        for (Entry<DataCell, Integer> e : outputmap.entrySet()) {
            if (e.getValue().equals(i)) {
                colname = ((StringValue) e.getKey()).getStringValue();
            }
        }
        DerivedField df = neuralOutput.addNewDerivedField();
        df.setOptype(OPTYPE.CATEGORICAL);
        df.setDataType(DATATYPE.STRING);
        if (mlp.getMode() == MultiLayerPerceptron.CLASSIFICATION_MODE) {
            df.setOptype(OPTYPE.CATEGORICAL);
            df.setDataType(DATATYPE.STRING);
        } else if (mlp.getMode() == MultiLayerPerceptron.REGRESSION_MODE) {
            df.setOptype(OPTYPE.CONTINUOUS);
            df.setDataType(DATATYPE.DOUBLE);
        }
        if (mlp.getMode() == MultiLayerPerceptron.CLASSIFICATION_MODE) {
            NormDiscrete normDiscrete = df.addNewNormDiscrete();
            normDiscrete.setField(targetCol);
            normDiscrete.setValue(colname);
        } else if (mlp.getMode() == MultiLayerPerceptron.REGRESSION_MODE) {
            FieldRef fieldRef = df.addNewFieldRef();
            fieldRef.setField(targetCol);
        }
    }
}